Qualcomm

I walked in and the first thing I saw was the bottom of the big crane boom arm with the weights And I was like, why are there Olympic weights here? And then I was like, oh because we've got a professional Boom arm camera. This is amazing. All right, let's do it You Welcome to season 11 episode six of acquired the podcast about great technology companies and the stories and playbooks behind them I'm Ben Gilbert, and I'm the co-founder and managing director of Seattle based Pioneer Square Labs and our venture fund PSL Ventures And I'm David Rosenthal and I am an angel investor based in San Francisco. And we are your hosts There's an incredible property of the universe where Electromagnetic signals can be broadcast and travel through space at the speed of light to be received at a different point in the universe Now a tiny fraction of these frequencies are detectable by humans as visible light Some other frequencies can be dangerous like x-rays or gamma rays But there's a part of the spectrum that is not detectable to humans and it's not harmful at modest doses That can be used to transmit invisible messages all around us all the time without any of us having any idea It's like magic. Yeah These frequencies have been used for over a century to broadcast TV and radio shows Presidential messages and important news updates in the last 50 years Humans have gotten tremendously clever at proposing some parts of the RF spectrum to be used for cell phones But the story of how we got from transmitting small messages on a single frequency To having billions of humans concurrently sending megabytes or gigabytes of data every minute Has been an incredible journey of invention and entrepreneurship the company most responsible for the mind bending system of how it all works today Is Qualcomm and today we will dive into their entire history and strategy unpacking their products Which to the outside observer is really best described as a layered series of magic tricks and spoiler alert for listeners This is an incredible story I had no idea before we dove into the research like this one is up there with like Nvidia TSMC there is so much stuff you can't make up in this story. It's incredible largest fabulous chip company in the world Indeed The other thing we should say listeners This was super fun to do this episode live in person in Lisbon Our huge thank you to the Solana foundation for hosting us at Solana breakpoint Many longtime listeners will know Austin federa from the slack He was kind enough to invite us and and really fun to do it there Especially given Solana's tie to Qualcomm with Anatoli having worked there for over 10 years indeed All right listeners now is a great time to tell you about a longtime friend of the show Vanta AI has scrambled the whole security picture. It used to be that you proved that you were secure once a year on audit or a static pdf Then everyone would not and you're done But in an AI first world that doesn't hold up anymore. Yep your risk surface changes every week now Vendor turns on an AI feature or someone writes in a new model without telling IT And your posture is different than it was last week let alone at your last audit Vanta's own research found that around 70% of companies have this Shadow AI running with no security review at all right, and that's where Vanta comes in They're the leading agentic trust platform meaning they've built the thing that closes the gap And the way that they close that gap is Vanta agent think of it as a GRC engineer that's governance risk and compliance Except that it's software and it doesn't sleep It finds the issues drafts the fixes and cuts the time that you'd spend on vendor assessments in half in half Which is exactly why more than 16,000 companies today run on Vanta Companies like ramp cursor and snowflake all stay audit ready and catch the risks that crop up between audits across every vendor Every AI tool the whole environment And that's the real value Trust has to be continuous now, which is why Vanta automates your security your compliance and the work to earn and prove trust We're huge fans of Vanta over here and literally hundreds of acquired listeners have become Vanta customers at their companies over the years So you can get a thousand dollars off Vanta at Vanta.com/acquired That's v-a-n-t-a.com/acquired for a thousand dollars off and just tell them that Ben and David sent you After this episode come talk about it with us. There are 13,000 other smart kind people in the slack acquired.fm/slack Without further ado on to our live show That's a lot of breakpoint and listeners know that this is not investment advice David and I may have investments in the companies we discuss in the show is for information and entertainment purposes only Well one small bit of a do before we dive into the story is we owe a big Thank you to Dave Mach the author of the incredible book the Qualcomm equation Which is not well known but is the definitive history of Qualcomm and ranks right up there with among the best Business books that business histories that we've used as a source on acquired throughout the whole history of the show It's awesome and the books not even really published under like a real publisher It's published under an industry association. There's no audiobook. There's no kindle you have to read the physical book Yeah, it's you it's amazing. I literally the other day texted Ben a photo that I noticed on the back cover and Ben of course it's seen it too Uh of one of the blurbs. I'm gonna I'm gonna read it here now says Dave Mach helps uncover the single most important business story Single most important business story that has yet to be told how Qualcomm came to rule the wireless industry Think of it as a recipe book of for one of the most innovative and leveraged business models of all time Whose words does that sound like Ben? That sounds like a deep business model thinker and someone who uh who truly appreciates capitalism and its finest And is willing to go find the rare gems the rare diamonds and the rough That is written and said by none other than Bill Gurley A benchmark capital for this almost unknown book. I bet it's gonna be a lot more known after this episode. Yep Well Dave starts the book And it's such an apt place to start with a quote by Edwin land Who I was not familiar with until recently when David center on the founders podcast familiarized us with Edwin Edwin was the founder of Polaroid and Steve Jobs's hero Uh, and he had this quote that Dave starts this book with true creativity is characterized by a succession of acts Each dependent on the one before and suggesting the one after So with act one of the Qualcomm story We start in Austria here in Europe in the mid 1930s in the pre-World War II era as Hitler and Mussolini and the Nazis were rising to power And we'll start is this the first time we've been able to say here in Europe unacquired. Yeah, it is the first time. It is the first time Uh, and we start you might think if you know anything about Qualcomm history I think in mid-30s they're like oh, I didn't know erwin Jacobs co-founder and CEO of Qualcomm was born in Europe He was not he was born in new bedford massachusetts. We start with somebody very different We start with one of the most famous Film actresses Hollywood film actresses of all time a woman named heady Lamar And side note the fact that we're starting with heady Lamar on the story of how modern telecommunications came to be is so cool I remember we reached out to the nzs capital folks and said hey, you know, do you have any great resources on on Qualcomm? And they sent back this excerpt of you should go read up on heady Lamar I was like are they trolling me right now? Yeah, you cannot make this stuff up. This is like why we do the show So heady Was an incredible she was like just an incredible human being She was world famous Incredibly talented actress incredibly beautiful. She would later be build like the way mgm She was one of the mgm starlets marketed her was as the most beautiful woman in the world She was also a genius So she starred in samson and delilah ecstasy zig-free girl many many more But what most people at the time even up into her death did not know and certainly her husband at the time In austria in the mid 1930s did not know Was that she had incredible powers of observation and was way more intelligent than Anybody else around her. So this said husband. That's quite the character His name was Friedrich Mandel and he was not a good dude He was a Nazi arms dealer which made him very rich at the time Just probably how he met he and they became married Heady though Probably unknown to Friedrich and certainly unknown to his business associates including Hitler and Mussolini Heady was a jewish And so Friedrich would bring his beautiful, you know, film actress world renowned film actress bride to his business meetings You know with the Nazi military powers and Heady was listening in to everything that was going on and as the situation deteriorated. In 1937, she disguised herself as one of her maids and escaped to Paris, and then from Paris, made it to the U.S., went to Hollywood and lived in Hollywood for most of the rest of her life. When she came to the U.S., though, she knew like an incredible amount of inside information about the Nazi war effort. And she was incredibly motivated because obviously she was from a Jewish family, she hated the Nazis, hated her former husband, and wanted to contribute. And specifically, she knew that the Nazis were working on and using to great effect a radio jamming technique for radio guided torpedoes that would be dropped from airplanes to attack Nazi submarines. It's also pretty amazing at this point in history that we had as humans the capability to radio guide the torpedo. And the torpedo, you know, gets propelled and you could guide it using radio frequencies deciding which way to turn the rudder. I did not know that technology existed in the 30s. This is crazy. The digital computer doesn't exist yet. The concept of digital doesn't exist yet, because we're going to get to that in a minute. This is all being done essentially with FM radios. And so, heady wants to contribute to the Allied war effort. And when you say with FM radios, therefore, pretty easy to jam. If you know that someone's broadcasting on jamming 92.3, and you start another signal on 92.3, you disrupt their signal and they're not able to hit their target with the weapon. Totally. So, heady teams up with her new Hollywood neighbor, a composer, a music composer named George Anfield. Bear with us here. I promise this is getting to Qualcomm. Who is a film music composer? And they, with her ideas and his musical prowess, they develop a concept that they patent and they get issued a confidential patent that says confidential for decades in the US military. By the way, this, I believe, did not become declassified until 1981. That's how long it was buried inside the US military. It was issued in 1942. So four decades that this history was completely unknown. They develop a novel technique to defeat RF frequency jamming by using frequency hopping. And what they describe becomes the origin of something called spread spectrum technology. So, if you're familiar at all with like the wireless world or Qualcomm, or I think you're here, spread spectrum and you're like, oh, that sounds familiar. Spread spectrum technology. This is the first, like, description of it in a technical document and a patent. By these two, like, incredibly unlikely people. And what it basically means is any way that you're going to transmit a single message across a variety of spectrums. So rather than just on, I'm going to keep saying, and J.M. and 92.3 to ground it in radio. But instead of just broadcasting on one frequency, they came up with this idea to hop. So change frequencies during different points in the message to evade anyone trying to jam the signal and move to a different frequency. And the reason she teamed up with a music composer for this is that the way you make this happen is you have incredibly precise time syncing on, in this case, the two ends, but in, you know, wireless use case, all endpoints of the communication channel incredibly precise syncing so that all endpoints know when to hop frequencies. And you're hopping frequencies like dozens or hundreds of times a second. And this can defeat jamming. This is great for cryptography. This is great for sending coded messages. It turns out this was not on anybody's radar pun intended at the time. It turns out that this is also the most efficient way to use radio bandwidth. But let's put a pin in that for now. And first, let's go back to this specific use case of we want to transmit from a plane to a torpedo and we want to be hopping around to different frequencies. And we want to change that at incredibly precise time. So the transmitter knows to change the frequency and the receiver knows to start receiving the message on a new frequency at very specific points in time. The concept of digital hasn't been invented. So how are we doing this David? What's the technology used to synchronize a schedule of frequency hops between a torpedo and an airplane? So here's where if this were a Hollywood movie, like one of Eddie's films, this single handedly would have like defeated the Nazis and all that. Unfortunately, the reality is there was no digital computing at the time. It wasn't possible. The US military tried very hard during World War II to make this happen. The whole Allied military. They couldn't make it work because like think about what you're trying to do here. And that vacuum tubes and analog computing was what was happening at the time. You would literally need to put like any act on a torpedo and drop it from the sky to make this happen. That was not feasible. It's worth sharing how their prototype worked though. So the way that they prototype this heady in the you know, 19 early 1940s is they took two player piano scrolls that had the same basically song and they mapped each note to a new frequency and they put the same player piano in the receiver, the same scroll and the receiver that they did on the transmitter and they pressed play on the player piano song at the same time. So it would know exactly where to hop around. So there were 88 frequency hops in their technical description of the pattern? Because they're 88 keys on a piano. So I guess literally you wouldn't be dropping any act from this guy. You'd be dropping a piano from the start. Yes, like a cartoon. Totally. Okay. So that is the origin that you can't make this up origin of spread spectrum technology. That's act one. Act two, we stay in World War Two around the same time, but a few years later, there is a young PhD grad, PhD grad from the Massachusetts Institute of Technology, the August Massachusetts Institute of Technology, who was working on code breaking for the Allies very famously at Bell Labs and at the Institute for Advanced Study in Princeton, New Jersey, where he intersects with luminaries like Albert Einstein, John Von Neumann, Alan Turing. We're not talking about any of those three folks, but by process of elimination, you can probably figure out who we are talking about. We're talking about Claude Shannon, literally the father of information theory, one of the fathers of computer science and the inventor of the concept of digital of the bit of information, like digital did not exist before Claude. So during the war, all of this effort culminates in what he publishes after the war, his masterwork on a mathematical theory of communication, which defines a bit, the new field of information theory, ushers in the digital era for the world, and combined with the other folks who we mentioned, Einstein, Turing, Von Neumann, and Bell Labs work on transistors during the war. These things come together to create the modern era of humans and the digital computer. Yep. So we've described like the Hollywood part, we're described here in Act 2, Claude Shannon, you know, birth of computing, all that. And it's worth maybe sharing a little bit about information theory if can I take a second? Of course. All right, so I had heard people reference information theory or communications theory dozens of times over the years and every time I'd open up the Wikipedia page, I'd see a bunch of complicated math equations and you quickly want to get to like, okay, but what is this? Why does everyone keep describing it as so important? And I think there's a pretty key concept that was an aha moment for me, which is all communication must happen through a medium. There's no communication that happens through nothing. You need some way to send signal from a transmitter to a receiver. And the method by which you communicate, the way you send signal, is governed by that medium. And so what I mean by that in particular is, let's use the analogy of a conversation. Well, if you're in a super loud room, then your message needs to be very loud, and it needs to sort of not be very noisy. It needs to be a super clear, super loud message because there's a lot of noise in the room. Whereas if you're in a really quiet room, then you can have kind of a message with a bunch of noise. Imagine someone talking, but there's a bunch of static. Well, that's okay if the medium itself, the room that you're communicating in, doesn't have a lot of noise itself. So there's this relationship between how noisy a message can be and how noisy the medium is that you're communicating in. And I think this is this very interesting aha moment where what he basically deduces is there is a theoretical limit to the amount of signal that you can pump through any given medium based on how noisy the medium is and based on the level of entropy or randomness in the message that you're trying to describe. So when I say entropy, let's say David, you're expecting me. You think there's a 99% chance that I'm coming to deliver the message to you. I just had breakfast. Well, if it's in a really loud noisy room and you know, there's I'm sick and I'm coughing and I tell you I just had breakfast because you were expecting it, it's fine if it's in a really garbage medium. But if you have no idea what I'm about to tell you and it could be everything from like, hey, you're fired to I just had breakfast and you have no idea. Like we need to have that in a pretty pristine environment. with really nice volume or gain on the signal. So that's sort of the high level concept of information theory, and more specifically of Shannon Hartley theorem describing the relationship between signal and medium. - Yeah, super, super cool stuff. So where this all comes together, in act three of our story here, which is gonna be a little longer, 'cause we're getting to Qualcomm as part of this, one Irwin Mark Jacobs. An American born in 1933, as we mentioned, in Scrappy New Bedford, Massachusetts, which used to be, I believe, the wealthiest town in America during the wailing era as we discussed during Standard Oil or Berkshire. I think it's Berkshire, actually. - It was Berkshire because 45 years before Irwin Jacobs was born in New Bedford, the Hathaway manufacturing company was started in New Bedford. - New Bedford. - New Bedford. - Right, before it merged with Berkshire, and before, of course, even by 1933, New Bedford was not the New Bedford wailing era, shall we say. So Irwin is a pretty amazing American story. So he grew up in like a very middle class family in this super scrappy area of the country. His dad worked a bunch of jobs and ended up running a local restaurant called the Boston Beef Market. Irwin was highly gifted in math and sciences as a kid going through school. He wanted to study math and science, and probably would have wanted to study engineering if he knew it existed in college. But his high school guidance counselor famously told him that there's no future for math and science in New Bedford. And frankly, his high school guidance counselor was probably right. So Irwin, though, had very good grades growing up. And the guidance counselor encouraged him to go to the world famous Cornell School of Hotel Management so that he could learn the hospitality management business and come back and work in the family business at the Boston Beef Market. - Which he did. - The school of hotel management. - This engineering genius, this like American pioneer of the wireless communication industry, that is what he went to college for. - And he would later credit the year and a half that he spent in the hotel management school at Cornell before transferring to electrical engineering. He would credit that year and a half with really helping him start first link of it, his first company, and then Qualcomm get out of academia and become an entrepreneur, because he actually learned about like business accounting in the real world applications and found that he kind of loved that too. Amazing. So after a year and a half at Cornell in the hotel management school, he learns about engineering and was like, "Oh, you can make money with math and science." This is actually like in demand, maybe not in New Bedford, but like in the rest of America. And so he goes to the D in at Cornell. He tells the story. He's like, "Hello, sir." You know, I saw more at Cornell. I would like to transfer from hotel management to electrical engineering. And the Dean's like, "Oh, you mean electrical engineering to hotel management, right?" He's like, "No, no, no, no." Hotel management to electrical engineering. - No, I want to do the harder one. I want to do the hard stuff. After the Dean picked himself up off the floor, he allowed it, perhaps with a degree of suspicion. Which he need not have because Erwin is another genius in this string of geniuses. He would graduate, go on to a PhD at MIT, which he would do in three years finishing his PhD in 1959, studying under none other than Claude Shannon, himself who after the war returned to MIT as a professor. - It's pretty interesting because so many of these stories that we tell, there's an immense element of genius. No question Erwin Jacobs and Jensen at Nvidia and Steve Jobs geniuses. And also. - There were like ten people in the world who knew this stuff at the time, and they were among them. - Yeah, it's the most incredible, right place, right time in history too, because without studying under Claude Shannon, the father of information theory, it's extremely unlikely that Erwin Jacobs becomes the Erwin Jacobs he went on to be. - Totally. And then without what's going to come later in Heddy Lamar, that he would start, welcome, amazing. So Erwin is so young, Erwin is so talented that after he finishes his PhD in three years, mere like five years removed from being a hotel management major at Cornell and Shannon and MIT, asked him to stay on as a professor at MIT immediately, which he does. He spends five years teaching at MIT during which he teaches the first course for students on digital communications. In the world, I believe, applying Shannon's theories to disseminate amongst practical engineers being trained at MIT, he and a fellow faculty member write the first textbook on digital communications that is still in use today. You can still, it is the Bible of digital communication theory. You can buy it on Amazon and written by Erwin distilled from the father himself of Claude Shannon. He spends five years teaching there. And then in 1964, he takes a sabbatical and heads out to California to do a sabbatical at JPL, at Jet Propulsion Labs, working on the US space program and communications with satellites in the US space program at the time, where he intersects, faithfully, with another recent MIT electrical engineering PhD grad, one Andrea or Andrew, as it was anglicized, Viterbi, a Jewish immigrant from Italy who got his PhD from MIT in 1957, who was working at JPL and they become fast friends. So fast friends, in fact, that when Erwin returns back to Boston to cold, snowy, bleak Boston near his upbringing in Massachusetts after his sabbatical, Erwin then gets a call shortly thereafter from one of his former professors at Cornell that a new engineering school in San Diego is being started, the new UC San Diego, and there's an opportunity for Jacobs to come out and start the electrical engineering department at UCSD. He says, "Well, I really enjoyed my time out there. I've got this great friend, Andy. Let's do it." I would make the exact same decision. So, he and his family, Erwin and his family, move out to UCSD. And while he's out there, he continues doing his contracting work with defense contractors and JPL and the US space program. And this is sort of one-off at this time. I mean, he's like doing it under his own name. He hasn't really started a company. It's just kind of Erwin doing contracting. Totally. He is like the first, you know, electrical engineering professor at UCSD. That's his full-time job, but because he's in such close proximity to everything going on at JPL and NASA and the like, he's doing that on kind of like one day a weekish. And one day, he and Andy and another professor from UCLA are up at NASA Ames in Mountain View doing consulting work up there. They're flying back and they're all kind of lamenting they're like, "This is super cool that we're doing this. We're making more money than academia. We're helping our country. We're participating in the space race." But it's kind of hard to like balance all this stuff that we're doing. And they're like, "Hey, what if the three of us band together and form a company, kind of a shell company, to just kind of manage this consulting work that we all get? We could probably get some, you know, efficiencies here, maybe hire an assistant, help us out, that kind of stuff." And they say, "Great. You know, we don't intend this to be a real company. We're not going to make any products or anything. This is just to manage our consulting. They sort of tongue in cheek decide to call it link a bit, like linking a bit. It's very like academic joke. So, who is this third partner in link a bit? He ends up not kind of jelling with the other two, leaves shortly thereafter. His name is Len Kleinrock. And I read that the first time when I was like, "I've heard that name before. I know that name." And I'm going to guess 99% of listeners haven't heard that name, but if you're you and me and all we do all day is study tech history and the history of the internet, that name should ring a bell. Well, you know, at first you read this history and you're like, "Man, Bummer for Len. He missed out on founding Qualcomm." Well, he actually ended up okay because instead of founding Qualcomm, he founded the internet. He literally was the, I think the founding engineer on the ARPANET project at DARPA. Many people were involved in the ARPANET project. Yes, but ARPANET, which was the precursor to DARPAnet, which is the precursor to the internet, Len and one of his grad students at the time at UCLA, like the next year, right after this is happening, at the same time, they sent the first message on ARPANET, like the first internet transmission ever from UCLA to Stanford. He's one of the core founding fathers of the internet. So he ended up doing okay. He probably didn't make as much money, but he will be remembered in history. Pretty amazing. So Andy and Erwin, they're mostly continuing to work on NASA and Navy Defense project, in San Diego because, of course, San Diego is a US Navy town and most of what they're doing is working on satellite communications. If you know anything about satellite communications, the bandwidth that you have available to you is very, very narrow and you need to be very, very efficient with your communications. And that's still true to this day. I mean, any company in the sort of emerging space economy, it's a totally different engineering problem than you're used to today. Because if you ship code up to your satellite and you find a bug, it's like very expensive and very slow to go get enough bandwidth and actually make sure you have the right time window to update the code on the satellite. So it still kind of works the way that computers worked 30, 40 years ago. Yep. And so there, you know, it wasn't them. Like, this was the military, there was this, they got exposed to this trolling around to find the most, best, most efficient ways to use this narrow bandwidth channel that they had. And what ends up getting used, but this old patented spread spectrum technology from the World War II era invented by Heddy Lamar and George Anthiel. And the, the timing is perfect because the time of Lincoln bit is this sort of early 80s where that really 70s. Oh, Lincoln, it's early 70s. Yeah, like 15 years of Lincoln bit before. Oh, yeah, oh, yeah, there's a long. Lincoln bit is involved. Oh, we, you may not know. I've got some good surprises for it. So they start doing more and more of this. Erwin's exercising the like hotel management sort of side of his brain as he's doing this. He finds that he really enjoys it. They start bringing on other professors, other grad students in to link a bit to kind of build the sort of like army of the greatest, you know, information theory and wireless signal minds in the country. All for defense contracting. Almost all for, I don't think they were doing any commercial work at this point. I think it was all NASA and defense. Yep. And almost all satellite work. And so they start building the company that eventually in 1971, there's so much going on. Erwin decides he's going to take a sabbatical from UCSD and spend a year just organizing the company. He ends up never going back to UCSD ever because during that year, they get the idea. I believe it was during this year. Maybe they'd start to have inklings of it before that it's really nice. They've got all this technical talent. They're consulting on these projects that defense contractors mostly are the prime bitters for. They're like, wait a minute. Those guys are making all the money. We're doing all the differentiated engineering work here. What if we started bidding on some contracts ourselves? We would probably make a lot more money as like a kind of product contract focused services company ourselves rather than just as a sub consultant on these projects. And that lesson persists to this day too. If you can pull off being the prime contractor to the government on a big contract, the economics are much better than if you get subcontracted by one of the primes. And if you can be a prime, the primes back then, primes being prime defense contractors, they're still the primes today. That is a gravy train that like, yeah, Raytheon, Lockheed, Boeing, all these companies. So of course they start doing this, but like there's a reason the primes then are the primes now. Link a bit is not going to be a prime then or ever. So they need to, if they're going to do this, they need to move into the commercial sphere. So this is, this is like one of these. It's so good. It's like history was made for acquired. Do you know what the first like contract project that link a bit did was if you knew you would just be like, I don't, I don't smiling so wide right now. So they hear about, remember their expertise is in satellite communications. They hear about a regional retailer. No, did they do Wal-Mart satellite network? Yeah, they did. What? Yeah, they hear about this eccentric founder of this small mid-western regional retailer that for some reason wants to beam himself talking every day to all of, you know, from HQ to all of the local stores of this, a local outlets of this retailer. Link of its first project is doing the satellite communications system for Wal-Mart. That's wild. Listeners, for anyone who didn't listen to our Wal-Mart episode, Wal-Mart was for a very long time the most innovative retailer on the planet. I mean, until Amazon, basically. And one of the illustrations of this is in the late 70s and then continuing into the early eighties when they actually lit it up, they invest, they invested tens of millions of dollars into building a private satellite relay because the bandwidth available on the internet was insufficient for them at the time. Oh, they was doing it. It was just the argument. It was the line-rock, do I? Phone lines. The public WAN effectively, or pre-cursor to WAN, was insufficient to, you know, send the store data that they had actually been collecting and want to tabulate their results on a daily or weekly basis. But also this like. Yes, Sam wanted to broadcast out the, you know, the Saturday. Oh, it's so great. Wait, there's more Wal-Mart to come a little later in the episode. Stay tuned. Literally. So, God, you just crack yourself up. This is fun. We've probably got this from the actual episode. We get, occasionally we get these reviews for acquired or like comments that like one host is like really normal and the other a host is just like, laughs at his own jokes, crazy person. And I'm like, well, you know, at least they remember me. We are who we are. Nothing's changed and it's seven years in. We're not, yeah. I promise you it's not an act. Ask my wife. Okay. So, the next thing that they get into is, because they're in video, they're in satellite, they're in video now with Wal-Mart and they're doing these two-way communications, they build the video scrambling system for pay TV on cable systems. So it used to be before the link-a-bit solution for multiple access cable systems. If you were like were even mildly technical or could like play around with like a allen wrench, you could get HBO or any of the early pay TV channels for free. Yeah, the catchphrase there is security by obscurity. Yeah, exactly. They were just trying, you know, find one clever thing that consumers weren't likely to figure out by unscrewing their box and, you know, moving one wire or something. So, Jacobs and Viterbi and all the brain trust at link-a-bit, they solve that problem. And the HBO uses them and then all the other big pay TV channels. I think that's the inspiration behind the HBO opener. Yes, this scrambled. Because it's like descrambling and now bringing you this. So, that's Irwin and Andy right there. So, in 1980, they do this for the whole decade of the 70s. In 1980, the link-a-bit, the company gets acquired by a East Coast radio technology company called Maccom. I think it's how it was pronounced. It used to be actually Maccom and then it, you know, this is like weird 80s branding stuff. They changed the brand to M/A-COM, microwave communications, I think. Anyway, they sell the business for 25 million bucks in 1980, which like-- Nice early win. Not bad for some former academics, 25 million bucks, in 1980 dollars. And they had a lot of people at this point. I think there was like over a thousand employees. It grew within, it was on its way there, but it didn't grow over the next five years within Maccom to that big. So, I don't think it was, it grew to 1,500 people eventually. Like, this is a big freaking business. You know, Matt, the things we're talking about, like a lot of other retailers started using, you know, satellite networks. A lot of other cable TV, you know, channels wanted to use these to, like, and there were other products that they were building. Like, this is a huge-- Like, basically, they made a big mistake selling the company. You know, they hadn't listened to acquired. They didn't know all the lessons. They wouldn't have had Qualcomm if they didn't sell the company. Well, that's true. They made absolutely the right decision in selling the link-a-bit then. So, they stay with Maccom for five years. And then there's a leadership change at Maccom, and like, this is an East Coast technology company. So, they all leave in 1985. And they sit around for a couple months. And, you know, they're like, like, we made more money than we ever, like, dreamed we would. We got to be part of so many cool things. But, like, we're still young. And, like, the wireless communications industry is kind of just getting started. And this is 1985. So, the cellular telephone industry exists at this point. It had just started. We had the, you know, how we were on 5G now. And everybody remembers, the iPhone 3G, that second phone, and the Edge network that the first iPhone launched with was 2G. It was a little advancement on 2G. This was 1G. This was 1G, which was analog. No digital yet in Cellular. It was analog cellular. And cellular had just been an innovation. I mean, this notion that rather than, you know, communicating over long distances, we were actually going to put cell towers so that you only needed to communicate with your local tower and that could be relayed. And you had this sort of cellularification of all the geography that you needed to cover. That was new. And it's funny how today we don't even think about what the word cellular means, but that was the most recent innovation at the time. - Yeah, it's great. So, you know, Erwin and Andy, like they're, they are first-rate academics, you know, as hopefully we've told the story here, like among the most brilliant minds in the world, but they're also like, especially Erwin, like incredible business people, market analysts, like they're very aware, like the products they developed and link a bit, they're aware that this market is coming. And the reason they're so aware, like technically, it exists now cellular. It's all car phones at this point in time, because the way it works, it was essentially, it was just like the torpedoes back at the day. It was essentially a FM radio broadcaster that you would wire up into your super high power. Super high power. You needed like a lot of freaking power. - You had to put it in a car for what you're talking about, and because you couldn't, like, there was not a battery available to. - You needed a running internal combustion engine to make this thing work. - Yes, on the end points, on the end points. And bandwidth was super limited, and like these systems were thousands and thousands of dollars in early '80s dollars. And despite all that, the consumer demand for car phones was incing, like this was just like, you know, there were wait lists years long for consumers to get car phones installed, and the fledgling carriers at the time, like they only had so much bandwidth they could fit, because literally, there's no efficient use of channels. It's just like the torpedoes back in the day. Like, they couldn't keep up with all the demand. I remember when my parents who were lawyers, like, they had car phones in the '80s. Did your parents know? - No, my great uncle had one. But it is interesting thinking about, you know, when you're listening on an FM radio, you have 99.1, and then you click up on the dial and it says 99.3, and then you click up on it, it says 99.5, and you can't even have 0.2, 0.4, 0.6, because that's too close, there would be interference. So you start thinking about, and this isn't exactly right, I'm gonna oversimplify this a little bit, but you start thinking about, well, geez, how many slots are there to communicate in this analog way with a cell tower near me? What can a cell tower handle? 100 phones, 200 phones, 500 phones? Either way, it's not gonna stop. - Not like much more than 100. - Yeah. Yeah, I mean, you think about how many radio stations there are, it's not much more than that. So, you know, if the link about folks are in an 80, they see this, they know, and they're like, oh, this industry isn't, it's infancy, we see this amazing demand, we are literally the best, we know there's a better way to do this. We know you can do this digitally, we know you can do it way better, we know how to do it the best, so they found a new company in July of 1985, with seven total Andy Irwin and five other of the best link-a-bit engineers. They meet at Irwin's house, and they just started to start this new company, and they name it Qualcomm. - Quality Communications. - Which is short for quality communications, which I had no freaking idea when we did the research, but then you're like, oh, quality communications. And then when you know all this history, it makes sense. Like, they are the highest quality, you know, but they know how to do quality communications. This is a communications company, and they can provide quality to the nobody else. - There's so many companies named this way too, these things become these household brands, and then it's like, you don't even think about what the original meaning was. - Totally, totally. Because the industry was still so early, and you think for a minute about what is involved in building out a cellular telephone network, there is enormous capex, like, laying cable, we've talked a little bit about the cable industry history not acquired, like, that required enormous capex. Like, this is like literally putting towers in the ground, putting base stations on them, building these thousand dollar mobile phones. It requires a lot of money to participate in this. - It's money, and it's a bunch of competencies, because not only when, are you thinking about the real estate for the tower and putting in the tower, and putting the base stations on the tower, well then you need to figure out, well, how are those towers? What's the protocol? What's the technical method that it's communicating with phones, and making sure that the phones have all the correct hardware, and it's not just antennas, it's very specialized chips. And so then you're like, okay, well, do we need to then make phones, and do we need to build a consumer brand, and do we need to market to consumers? Do we need to be our own carrier? Do we sell to carriers? There's a way to sort of like bite and try and eat the whole elephant here, or you could say, okay, we're just gonna try and be one small part of this, 'cause we have an idea for how to make this better, but if you're just doing one small part of it, and inventing the means by which the technical method that the phones communicate with the towers, there's a bunch of stakeholders that you gotta get on board with your thing, carriers, the government in terms of licensing spectrum, phone manufacturers, chip makers, base station makers, so there's this really interesting crux that they're at at this point of the company, where they're saying, we know we can do this better, we have a specific idea about how to make this better, which we'll get to in a second, but there's some really trying to figure out how much of the elephant to try to eat themselves. - And this story, hopefully this first 45 minutes of the episode was interesting. Fun telling this crazy World War II Hollywood history of all the technical aspect that comes to this. The business history of Qualcomm, just like Bill Gurley said on the blurb of this book, it is one of the most brilliant strategic executions of entering a market, period, like writ large effort, like this is on par within video. If not, honestly, more brilliant. - It seems more difficult, 'cause if you were to pitch me this idea a priori, as an investor, I would tell you immediately know, because I see 15 different needles, all of which you must thread perfectly, a story that's entirely path dependent, so you're not gonna get one thing until you get the previous thing, and that was a needle that you were threading, so the likelihood of success is unbelievably low. And yet here we are talking about Qualcomm. So they knew two things at the outset of founding. One, this is a massive opportunity that they eventually wanted to pursue, was bringing their expertise to bringing cellphone, terrestrial cell phone networks into the digital era, and building the dominant gorilla company in this soon to be massive industry. And two, they knew they couldn't do it yet. So they actually started in the same fashion that Linkabit did. They're like, okay, we're gonna bootstrap up by doing consulting work. So one of the first consulting projects they do is with Hughes, one of the defense primes, Hughes-like Howard Hughes, pretty awesome, on a proposal to the FCC for a mobile satellite network. They're like, all right, well, we'll learn about consumer and mobile telephony services, enter the market, we'll work on the satellite network. - And we're talking like Jurassic Park smartphones. - Yes. - That's exactly. Big Hawk and things, super expensive, but like when you really need it, it's nice that there exists a satellite phone network. - Yes, so while they're working on this, they're working on like, okay, how can we like, where are the experts at, you know, optimizing satellite communication channels for efficiency? They come up with an application of spread spectrum to use multiple access, multiple conversations access the same channels at the same time. - Yep. - That they call, they use a technique called CDMA, code, division, multiple access. - Which the first time you hear this phrase sounds like complete jargon, like meaningless, and then you stare at the Wikipedia article for a while to try and unpack each one. So we'll break it into parts, multiple access. Well, that's fairly straightforward. Rather than being broadcast, so like a TV network, we have multiple endpoints that all wanna communicate with each other using whatever the same communication medium is. So rather than using one single frequency to all try to pile on there at the same time, which of course wouldn't work in that analog world that we were talking about. I wanna call you on 92-3. You wanna call Bob on 92-3. My mom wants to call my dad on 92.3. You quickly get into a situation where like, everything's just colliding with each other. So multiple access on just single analog frequency doesn't work so you gotta divide up and say everybody gets their own frequency and that's sort of the way the world evolved. So you mentioned code division. Before we get to code division, can we talk about a different type of division? - Yes, we certainly can. - So before we get to the CD in CDMA, code division, let's, so we've got the multiple access part. Bunch of people trying to communicate using the same medium. - Well, the things that we were talking about before everybody gets their own frequency, that was called FDMA, frequency division, multiple access. So a pretty straightforward way that you might divide up the error waves in order to have multiple conversations. And the way the telecommunications industry works is, remember I opened the episode by saying it's basically a layered set of magic tricks. This is sort of the next iteration on top. And if you say, okay, rather than sending analog signals, what if we were sending digital signals? So if I'm talking to David, there's a lot of sort of pauses about half the conversation is actually empty error. And if two folks out in the audience are talking to each other, a lot of your time is actually empty error. So we don't both need the entire frequency all the time. And if we are communicating using a digital signal instead of an analog signal, then actually we can parcel up the information into digital packets. - And just rotate the time of when different packets are. - Right, so the very crude example is, if we're at a dinner party, I can have my conversation for 30 seconds in a room and then, you know, I pause and I stop talking, a different conversation can happen for 30 seconds. Of course, that's too crude and that's far too long. In a time division network, what you basically do is say, I get some digital packets for these milliseconds, then the next milliseconds, you get your digital packets, then the next few milliseconds, someone else gets their digital packets and we'll keep round robbing it between the 20 conversations that we're all having and when it gets reassembled on the other side by some other phone or something. - Thanks to transistors and digital technology, this can all happen fast enough that, like, you don't even notice. - Yeah, you're like, oh, the signal maybe sounds a little compressed. It's not as good as if we're talking to each other, actually, face to face. But there's no, like, weird blips or pauses in the conversation. And though we're all borrowing different time slots on the same frequency, it actually sounds pretty smooth to me. So that's the next iterative event. - In this case where Europe was way farther ahead than the US, Europe was basically ready to implement this time division multiple access digital standard in Europe for European cell phone technology and that was driven by Ericsson, the big European infrastructure provider. So I think, like, just to pause and reflect, big innovation going from, maybe, 20, 30, 50X, like, you get a lot more capacity by saying, instead of just one person gets a frequency at any given time, you now get a whole bunch of people who can use that frequency because the signal's digital, because it's time division. This is the movement from frequency division, multiple access, FDMA to time division, multiple access or TDMA. And it's actually, he said, 30, 50, maybe now that kind of is, but, like, back then, it was 3 to 5X. Really, I think the right analogy is, like, it is time sharing. Time sharing is what it is. And it's kind of like the old computing model of, like, time sharing on a teletype on a mainframe. That's what's going on here. - Yep. And so, over to Qualcomm, so they're thinking about doing this satellite communication thing. And remember, Erwin studied with Claude Shannon. So he's always thinking about what is the most efficient way to use all the way up to the theoretical limit of how much signal can be communicated in a given medium at a given time. And he's sort of looking at TDMA, and they're like, I think there's something even more efficient than this. And we need something more efficient than this for this satellite network. - And these guys were all around the beginning of the internet, and like, you know anything about how the internet works-- - And packets with networks generally. - Packet switching. It's not time sharing. - No, it is. Everybody compresses their data as much as they possibly can into a digital packet. They fire it off, and it bounces around a series of places until it hits the other side, gets decoded, and hopefully the protocol is written correctly where, as you're sort of opening your packets and sequencing them all in the right way, it seems perfect and how the message was originally intended to be when it was encoded in the first place. He said the magic word decoded, and that's what these guys figure out. They're like, duh, we'll just use code. And then everybody will send all the conversations all the same time, all across all the different channels. We'll maximally, efficiently use all the spectrum allocated, and we'll just depend a little code to the beginning of each digital conversation, and we'll get reassembled on the backend. So basically the same way the internet works. Yeah, so to break that down further, so you've got this really interesting situation now where all messages are encoded digitally, and I keep going back to this analogy that they use in the telecommunications industry of the dinner party. So rather than the sort of frequency, the FDA model of everybody's in their own room having their own conversation, that's not super efficient, or TDMA, which is you put five or ten people in a room, but they need to wait their turn to have their conversation. Well, what code division basically is, as the analogy goes, is, well, everybody can communicate in whatever room they want, they're all just communicating in their own language, and the person that they're communicating to understands that language. They can sort of listen and disregard noise that's coming in from people having it. It's like you're saying, if I'm expecting your message to be, I had breakfast this morning, then I don't care how much noise is in this system. You don't care how much space. I either know you said that or you didn't say that. You're like, I'm disregarding all the Spanish, and I'm just listening for English that sounds something sort of like describing someone's state of breakfast. And that's an oversimplification. If you really wanted to sort of dig into it, what you're basically doing is you run any given packet through literally an encoding. So maybe my encoding is 10010, so you detect, so you encode whatever the packet of information is, you run it through, sort of add it to 10010, and then you end up with this signal that you can sort of stack on top of other messages. So imagine a digital signal, like a digital wave, where all of our messages are layered on top of each other. So the top of the peaks of some of the wave are extra high, and the troughs are extra low for others. And when it all arrives all together on the other side, the other side knows how to decode all of our messages. But individually subtracts all of our messages, which are layered all on top of each other, off the very same digital signal until it basically has all of our messages spread apart. It disregards any of the ones that doesn't match the code that I'm looking for, that I'm listening for, and it says I just care about the message that came from Ben, which was 10010, or whatever, code I just made up. And that reassembles it. The schtick for CDMA, and what these guys do, just this brilliant, like they saw it, they had the background, they had the engineering, they like everything, right place, right time, and the business sense, they developed this, and they freaking patent it. In 1986, well before years before Qualcomm gets actually directly involved in the cellular industry at all, they patent the method and technique for code division multiple access, applied to terrestrial cellular networks in 1986, in US patent number 4,901,307, which is one of the most valuable patents in history. Yep. Unreal. Literally they played such a long game when they threaded needle after needle after needle, and that was just the first. And when you think about why that is so valuable, when you really distill down what the CDMA patent is, it was the very first time that you could say, well, rather than thinking about one specific frequency, just imagine you have all the frequencies available to you, and everybody can all the time broadcast their message on whatever the next available frequency is, and we have the technology to just figure it out on the other side. Oh, and by the way, you don't even need to do it with super high power, so it's good for battery life and that sort of thing, because sense is imploded. You don't need an internal combustion engine to power this thing, right? The other side knows what it's looking for, so this is the equivalent of there's a bunch of people whispering in a gigantic house to each other all in different languages. So it's this way more efficient way to use a given medium to have the absolute maximum amount of conversations or signal transmission in that medium. Okay. So, Qualcomm pounded 1985 patent issued 1986 or applied for it in 1986, which is worth remembering. So, it'll expire in 2006. That's right. That's right. Looking ahead. Force outing. Qualcomm doesn't enter the wireless industry until 1989. What happens in the interim, this is the next Walmart, oh, it's a good, it literally just can't make this stuff up. So they get approached to bid on another contract, fledgling Qualcomm does, from a company called OmniNet, which has this idea that they think the Qualcomm folks are going to be perfect to implement. They want to make a mobile satellite network specifically to connect commercial semi-trucks on the roads in America and network them up to the distribution centers for retailers and other companies who ship a lot of things in the US. This is right in their wheelhouse, Qualcomm and Irwin are like, great, we're going to bid on this contract. They win it, they start working with OmniNet and they make it work and one of the very first customers is of course Walmart, which implements their own proprietary fleet of trucks, building further their technical advantage over just about every other retailer in America. At this point, they've walked away from the satellite contract, right? They sort of like, the huge satellite thing that actually just never happened. So they developed this technology, they patent it, they were like, oh, but there's no money here because the contract. Yeah, the FCC was like, yeah, satellite, Jurassic Park phones, not going to be a thing. Right. So instead, they're focused on this OmniNet. So they focus on this and they also have like a lot of the business, you know, relationships already from the previous iteration of what they were doing, a link of it, including with Walmart, many of the other large companies and retailers. I believe it's Schneider trucking, it becomes one of actually the first customer I think for that. So they work on building that, it becomes pretty clear like, this is going to be the interim main product, Qualcomm and OmniNet merge in 1988. They raise $3.5 million in funding as part of that. They bring the product to market at the end of 1988, as often as possible. company tracks. People might have heard of it. It was part of Qualcomm for a long time before I believe and ended up getting spun out to private equity. And in 1989, in the first year of business for Omnitracks, they do $32 million in revenue in 1989, which is something like, it's like inflation adjusted $100 million. It's a lot of money. And there's a lot of demand for this product in the first year of the product launch. Year one. Now there's a lot of COGS, like this isn't SAS Rev. No, yeah, yeah, talking about. And there's particularly a lot of COGS because one of the things they learn from doing this and one of the reasons the companies merge, they first kind of like the link of it is, you know, they remember Walmart was their customer for the link of it satellite, you know, thing. Walmart is very happy to integrate an implement technology themselves. Most other customers are not. So they go around and they're like, you know, pitching this to tracking companies, retailers and the like. And most of them are being like, well, this is like cool, but we're not going to operate our own dispatch centers and messaging. We try to have a small and IT department as possible. Yeah. Why on earth are you asking us to do all this work and just handing us this pile of technology? Yeah. So Irwin is like, well, what if what if we just operate it for you? And we provide a whole full stack solution. We don't tell you the technology. We sell you a solution, which is like every enterprise company that you ever, you know a company has become enterprisey when they cross the chasm and their website no longer has like products pricing about and it changes to solutions. Yeah. Solutions. They make the business discovery of solutions. We all should say like, this is a tremendously dilutive financing event. This is Qualcomm saying we need money so badly to fund the development of Omni Tracks for this customer Omni net that the most attractive option for us is to sell half the equity in our company. So everyone gets diluted 50% by merging with the customer themselves in order to get just a few million dollars to continue funding this effort. It's a pretty different time than today where you go raise a seed round and you sell 5, 10, 20% of your business for too many seed rounds that are happening for 5% dilution these days. But they were. And so it's a very it's crazy to think the position that they were in where everyone was looking at Irwin and he was like, hey, I think this is literally the best path forward in order for us to get the few million dollars we need to get. I think some people were pretty bitter about this. I totally and you could imagine too. It's not like an idea like they had done a bunch of work already. This was going to happen. They were going to go to market. There were just a couple of years away from making a hundred million dollars and inflation adjusted dollars and yet they had to give up half the company. They literally were a couple of years away from making actual 100 million because the business doubles every year for like 5 years from a 32 million dollar base like. Wow. Pretty awesome. So now that this is in place they're like, all right, we have both a cash flow spigot that we can use. And now like a base of business that we can finance and like borrow against and raise equity against to pursue the real big idea in our original patent. Here's the other just brilliant thing. What happened originally was not in fact there were other people who knew about code division multiple access. Other folks could have been in a position to patent this and pursue it. But at the time nobody believed it could actually work because you needed such sophisticated processing power on both the endpoints on the base stations and the endpoints to actually make this work. It sounded completely freaking crazy. It needs to happen in real time. I mean people need to have conversations without a perceptible delay and you are cutting a comment. You're first doing the analog to digital encoding where you're taking their voice and you're actually turning it into a digital signal. You're cutting it up into a bunch of packets. You're encoding those packets with every user's unique code. You're sending it over the airwaves to your most local cell tower. That cell tower is relaying it across a variety of other cell towers to where the other person on the end of the conversation is having the call. And then the whole pipeline is happening in reverse on the handsets. On the handsets. And so this was the thing like main real could believe you could do this processing on the on the base stations on the infrastructure side. But like the idea that like at a car like something powered by an internal combustion engine like at a car or or having forbidden not a car like a mobile phone, like a Zach Moris phone that you know somebody would hold in their hand that you could do this on something like that was crazy in 1986. But the Qualcomm guys they know about Moore's law, which like most people didn't know about it at that time. And they're like yeah I'm pretty sure you give it one or two more you know turns of the crank on Moore's law here and like I think we could maybe do this. There are so many things that we've talked about in the last I mean I don't acquire generally but especially in the last year where their success came from correctly forecasting where Moore's law would be at the time that they ship their product. Yeah so no like at the time of shipping. No like it's not possible today. But when we're going to ship this which is still going to be several years in the future it will be possible then to be like so cool. And like the fact that it's like there were so few people that knew that then and like crazy. All right listeners now is a great time to thank our longtime friend of the show service now if you are running a large enterprise AI agents are likely spread across every team and deploying them is no longer the hard part. Yeah the hard part is knowing what permissions they have what employees are using them for or what decisions AI is making. AI security for an enterprise at scale is not a small concern like the risks are real. Exactly and the challenge with AI is governing it securing it measuring it and making sure that it actually delivers value that is why service now built the AI control tower. Yep AI control tower gives enterprises a single place to see manage govern and optimize AI across the entire business and it works with any AI not just theirs every device on your network every permission across every system every AI agent visible and secure in one place and service now can do this because they've spent more than 20 years building the operational backbone of the enterprise the workflows governance approval security controls and institutional knowledge that power how work actually gets done across it HR customer service finance and security service now already runs more than 100 billion workflows annually and trillions of transactions for more than 85 percent of the fortune 500 so when companies need a place to govern AI at enterprise scale they're building on a platform at the center of how their business already operates and in a future that isn't going to be one AI it's going to be thousands of AI agents working across every function of the company but the question is who's managing them all so if you're trying to turn AI ambition into real business outcomes and make it work safely securely at scale go check out servicenow.com/acquired and tell them that Ben and David sent you so in September of 1988 all these factors you know they've got the financing capability to take swing at this they see a path with Moore's law to it being technically feasible they've got the patent they're literally the only ones who can do this and then the market timing so in September 1988 the US Cellular Telecommunications Industry Association or CTIA as most people know it and then it's related entity the TIA the Telecommunications Industry Association they release performance requirements they spec for performance requirements for the planned upgrade of the US's cellular networks from the analog 1G networks to the new digital 2G networks and this is just the US one Europe has its own Europe's already well on its way GSM, Ericson, TDMA it's all happening here in Europe the Qualcomm folks of course they eagerly anticipate the release of the spec and they look at it and they're like oh my god this could not have been written better it's written for us it's a dream it's written for us they realize two things one of course TDMA is the the frontrunner and Ericson now that to like you know do the US too because they're successfully doing it in Europe and not only is it being done in Europe it makes sense to adopt in the US too because it's kind of nice to have a global standard and because it's quite believable like okay one big thing I have to believe is we're switching to digital I can believe that another big thing I have to believe is that you're able to use the same frequency for several conversations at once through cutting up you know different time windows okay I can believe that but gosh how much new stuff are you trying to invent all at the same time anything cooler than that feels like I got to take a leap of faith and show me it can work and Ericson's well on the way to like pilots proving showing it works this actually works they're big companies they've succeeded before they're the right vendors that everyone trusts so the spec that the CTIA publishes the Qualcomm guys they just must have just been like beaming year to year they realize that TDMA because of the capacity limits of TDMA it's not going to meet spec like you can have the best inflammation implementation of TDMA it's not going to allow for enough compression to actually meet the spec that the US wants to hit so here this is a I've been I've been like waiting to bring this thing up so at this point in history the US standards body has is correctly forecasting the incredible popularity of cell phones in the US so they're setting a really high bar for the amount of phones that need to be able to use this network and the reason that they have since changed their tune is in 1980 this is a fun bit of trivia AT&T who has been the incumbent for a hundred years on all things telecommunications commissioned McKinsey and company to predict the cell phone it all goes back to McKinsey always always to predict the cell phone usage in the United States in the year 2000. So it flashed for 20 years in the future. The consulting group argued that cellular telephony would be a niche market. Yes, of course. They forecasted 900,000 people would be subscribed to a cellular telephony network in the year 2000. I think I have 900,000 cellular connections personally. So as you know, that number was off by over 100x. There were 109 million people, not 900,000. 109 million subscribed in the year 2000. So it does make the point that in 1980, it was super not obvious. Like you had some of the smartest people in the world, both in domain depth at AT&T and just good business model thinkers at McKinsey, wildly misforcasting this. And to illustrate how big the miss was, AT&T eventually bought McCaw Cellular for 12.6 billion dollars to become AT&T wireless, which is the AT&T we actually all know today, and catch up in mobile telephony. So this like 2G spec that was written is right around the time that a lot of the people in the industry are starting to realize like, uh-oh, were we super wrong in what we all thought just a few years ago? The potential of this thing was? So that's like, you know, back to the original Edwin Land quote starting the episode of like creativity, like one act following another, you know, enabled by it suggested to suggesting the next like, this is the next like needle day thread, you know, domino that falls of TDMA didn't hit the spec. And they could kind of foresee this, you know, because they knew what the demand was and they knew TDMA wasn't going to be able to do it. So, uh, here's the next, and this is cool, like I didn't expect to get into kind of like geopolitics on this, but the one great that you know, the US has like a ton of bureaucracy and regulation like all of this being like, you know, case in point. But one incredible, I think this took five years to eventually. And like the standards bodies and like all like this is not the free market like by any means, but the one difference in the US process for all this versus the European process. And it was the difference that made all of the difference was the US government said the industry associations, you guys can set the specs and all that and that can be official. But it's not mandatory. So like in Europe, it was like mandatory, like the TDMA, which DSM was based on like mandatory. That's it. And plenty of other countries, you know, mandatory. And the US was like, this is the industry standard and like we recommend that any mobile carrier follows it. But if you want to do your own thing, like as long as it meets the performance spec, you can use whatever technology you want. And importantly, standards bodies are decoupled from government agencies. So the FCC allocates spectrum, but these standards bodies are literally just industry associations. And they need to exist because there's so much coordination between all the different manufacturers and carriers and companies involved that like, you need to have a standard, otherwise the innovation doesn't happen because no one knows what to build against and no one can sort of effectively collaborate enough. So once all this, the standard comes out, Qualcomm immediately like goes to Washington, like, oh, and they go to DC and they're like, hey, just to make sure, we just want to like be crystal clear, can you confirm to us that even though this other thing is the standard, if a given carrier, mobile operator, wanted to use something different, as long as it used a spec, like that's cool. That's not illegal, right? And they're like, yep, that's the case. They're like, okay, cool, thank you. We'll be back. And so that was like the next needle they thread. They're totally undaunted. They go, they're like, great. We can go pitch individual carriers on using CDMA as a technology. So they start a sales process. This is now the beginning of 1989. They start a road show. They go out pitching this new novel CDMA standard versus the TDMA industry standard. And this starts what is known. Literally, I tweeted this the other day, in the Wikipedia entry for all this, this is like canonically known as the Holy Wars of Wireless. There's so much telecom nerdery. And it really is Holy Wars. Because it's about belief. So many people were just like, I don't believe you that CDMA will work. And it was literally only the Qualcomm folks who thought it would work. And not just, you know, I'm reminded of the Don Valentine, like I knew the future based on, they didn't know the future per se. But based on all their experience, they were very, very confident that it would work and it would win despite the seemingly overwhelming odds. Because they knew a secret, which was that at the end of the day, as long as there was not government enforced standardized regulation, they knew that economics would win in the market. And there's so many benefits of CDMA versus TDMA. We've covered some of them. One of the other ones is that the voice quality is actually much better than TDMA. There's a whole litany of benefit. Security is much better. I mean, it was originally created for the government to beam stuff up and down to satellites. Another huge one is it literally, if you're operating a cell network and you can have more subscribers per unit of infrastructure is literally cheaper. So you're, you're going to, yeah, it's a lower cost technology. This is the thing. So there's one benefit that actually matters. All the others are like nice to have on a feature spec. There's one benefit that is going to allow them to be super sure they're going to win, which is that it is like an order of three to five X more efficient to operate. Unfortunately, they originally pitched 40X. That's the standard that everyone would spend. Oh, that was versus animal. I think there's three to five X more than TDMA. So that meant if you were a carrier and you went with this crazy CDMA thing and it actually worked, you could fit on a given set of spectrum that you are operating with. You could fit three to five X more subscribers, three to five X more monthly revenue on that same fixed cost base than your competitors who are using TDMA. And if you know anything about like, if we've learned anything unacquired about economics of industries and power in Hamilton, Helmer and all that, like, if you have a scale advantage like, or you have a power advantage of differential profit margins versus your competitors, you are going to run the table on your competitors in any given market. Yes. If a customer is worth more to me than they're worth to you and we can offer them the same value, they're, I'm going to win because you can just lower prices and get all the customers and make more profits along the way. And there's, we've only sort of scratched the surface on this episode of reasons to doubt that code division was the right technology. There were all these other crazy hoops they had to jump over. One of them is the the near-far interference problem. Oh, yeah. This is like, like, if you think about it. So like, let's keep the whispering analogy going. The code division idea is that we can all talk really quietly and use the smallest amount of power and the smallest amount of sort of gain in our signal to communicate with each other. So it's much more efficient than all these other high gain, high power, high volume signals that everyone else is trying to use. Well, if I'm using a really low gain signal and I'm far from my, the base station, from the cell tower, that's an issue because the people who are really close are going to sort of drown me out. Imagine we're all whispering, but I'm miles away. Well, you're going to hear the person whispering right next to you. So, you know, we're very early days in powerful chips powerful power management. And you've got Qualcomm pitching the industry that they're going to do this. And people are like, wait, but you have to turn down the gain on anybody really close to the towers and turn up the gain on anybody really far from the towers. And you have to know in real time and adjust in real time all of that. So you have to be good at power management chips. Also, how are you going to know how far away someone is from the tower? And they're like, well, we'll be able to just observe the signal that is coming back from the tower or perhaps do it on the tower, observe the signal coming from the phone itself. And we will in real time determine if it needs to go up or down. And this is blowing people's minds in the mid 80s. They're like, oh no, we got that. In real time, you're going to modify a signal based on what you're currently hearing from that signal. And then Qualcomm comes in way over the top and says, oh, also there's this new thing called GPS that is coming out. And we're going to start from the military basing the technology on GPS. So we know how far away someone is from the cell tower based on GPS, which doesn't really exist yet. Like there's always impossibilities with this system that theoretically is better, but we've never witnessed any of the building blocks that are going to go into it actually work in practice yet back to the magic thing, like just the technological magic that went into this at every stage of the way. They're like, yeah, we got this figured out. And they patent every single piece of this. Every single piece. The first patent we talked about is the most valuable, but there is all string of dozens, hundreds, thousands of other patents that come after this that are just incredibly valuable. So they start the red show pretty quickly in February of 1989. One of the largest carriers in the Southern California area, Pactel Wireless, is interesting because they get it, like the economic argument. Basically, they're like, all right, if this works, like, yeah, you got us. So they put up a million dollars to fund a prototype. They're like, okay, prove to us that this works, build a prototype. Qualcomm for the rest of the year works on this. November of 1989, they hosted demo. you know, with the Pactel money, but they invite the whole rest of the industry in San Diego. And there's a famously a little hiccup where like they're about to, you know, Irwin's giving like a big speech, introducing it, then they're gonna do the actual demo. They've got vans driving around the city and then like a base station back at Qualcomm HQ and they're gonna make it all work. He's giving the interest speech and one of the engineers is like, frantically waving at the back, they keep talking, keep talking. They had to reboot the GPS system. And so like he's, you know, he makes a little equip of like as a former professor, it was easy for me to keep talking. He's told the story like a million times. Anyway, there is something funny too about this original demo where they're not a consumer, hardware manufacturer yet. They've never built a phone. There are a bunch of academics and consultants and you know, the electrical engineers. And so for this demo, the cell phone that they build, basically looks like a mini fridge with like a handset hanging off of it. I mean, they build the most-- - There's a photo of it in the book, it's awesome. It's awesome. We'll come back to building handsets in a sec. So it works. And then like Pactel's like, great, we're in. And then somebody-- - Which Pactel, by the way, would eventually get rolled up into Verizon. I think they're basically a Verizon's West Coast operator at this point. - Some of the other industry folks who come, they're like, well, this is impressive, it works. But like San Diego's a pretty forgiving environment for cellular technology. Like this is a very like geographically easy city to operate in terms of wireless signals. Prove to us that this can work in like an urban jungle environment. And Qualcomm's like, okay, how about New York? (laughs) And they're like, well, we'll see you there. So in February of 1990, they do a successful demo in Manhattan, in New York City. On the back of that, they sign 9x, 9x Mobile, which is one of the largest New York carriers. And then in August, they sign Ameritech, which is one of the largest-- - Chicago, I think. - Chicago, yep. I think I got a big chunk of them in West. - West. - And then there's another really move. They start going international. So like here in the U.S., there's all this like forward momentum that's already happened with the 1G analog services. And you know, the TDM and all that. They're like, what if we go out to countries where it's just tabularly clean slate? And we pitch this as like the obvious best technology. And famously, South Korea. In fact, the like government mandated standards. The South Korean government is like, yep, this is clearly the best government mandated. All, you know, they were building up the first cell phone networks in South Korea that were going to be these digital, you know, next-gen networks. - Yep. - All CDMA, all Qualcomm. South Korea for a time was, I think, close to 40% of Qualcomm's revenues. - Wow. - 'Cause the whole country. Like, and it was one of the, you know, most advanced mobile countries, all the season Qualcomm. - There's lots of benefits to the free market and freedom. - There's also benefits to regulatory and government catchers. - Yes. - Coming in over the top with an edict is also beneficial. - Yes. - In December of 1991, on the back of all this, they go public. There is a poultry $68 million in their IPO. - Like a series B. - Yeah, totally. A 2021 series B. So, finally, in 1993, the US Industry Association, the CTA and the TIA, does actually adopt CDMA as a second standard officially. It's like, okay, now you have our blessing. It's like, well, it doesn't matter. We already got like half the industry signed up with us. Anyway, you know, thanks for nothing. At that point, Qualcomm does a secondary offering. There is another 150 million on the public markets. A couple years later, they do, maybe a year later, there is another 500 million on the public markets. So they're very well capitalized. And why are they raising all this money? Back to the omnitracks and like this, you know, solutions discovery of like enterprise, you know, the people that they're pitching as their core customers, the wireless carriers, they are sophisticated operators. But there's a whole ecosystem of technology providers to them. And they already, except in the case of South Korea, you know, they already have built out like towers, infrastructure, they can replace all that. And so, you know, it's a big ask, even with the economic advantage, it's a real big ask for a Pactel or you know, a 9X or any of these folks. Oh, if you're a Pactel, you're like, it sounds great to me that you are going to have this much better standard and this much better technology. Are you going to replace my towers? Are you going to replace my base stations? Are you going to replace all of my customers' handsets? Right, like all of our customers buy phones from phone manufacturers. So are those phone manufacturers signed up? Yeah, right. It quickly becomes a rat's nest of industry dependencies. Qualcomm, they're still relatively small. San Diego, you know, technology startup. They can't do all this stuff. So they do start signing some partnerships with both base station infrastructure providers and handset makers. They sign Nokia, big win, big European manufacturer as a partner. But they realize, you know, to do this whole solution, like specifically, there's kind of four parts to making a CDMA wireless network work. We've talked about all of them, just to enumerate them here. You need the core IP and technology that we've talked about. Qualcomm's got that for sure. You need the infrastructure, the CDMA, like base stations that go on the towers, you know, all that. Like the back ends, the switching all that. You need that infrastructure. It needs to be CDMA. The old stuff's not going to work with it. The CDMA stuff's not going to work with it. You need the handsets for consumers to work. Same deal. It's got to be CDMA. And then probably most importantly, in order to make those two sets of infrastructure work, you need the silicon, the semiconductors, that go into them. And so somebody's got to do all four of those things. All four of those things need to happen. Qualcomm's for sure got number one covered. The question is, who's going to do 2, 3, and 4? Well, I was like, you know, they science start signing partners, but they're like, you know, we really need to spur adoption. I think we kind of got to do everything ourselves. We need to offer the complete solution. Complete solution. And this is a major undertaking. This is why they raise all this money in the public markets. Which is quite interesting, because despite-- I mean, none of us are buying Qualcomm phones today. No, Qualcomm brand phones. Today, spoiler alert, Qualcomm today is the largest fabulous semiconductor company in the world. That crazy. Bigger than Nvidia. Bigger than Nvidia, and they don't make-- And bigger than Apple? Don't make infrastructure. Think bigger than Apple. Oh, yeah, yeah. In terms of numbers of orders they're placing with chip foundries, Qualcomm is the biggest. So how do you get from there to here? So they did need to run this really interesting playbook where, even though it wasn't going to be the thing that they necessarily did long term, in order to get their solution adopted, they had to do it in the most-- We'd strap it up. So they do another just brilliant move. They create two joint ventures. I believe both of them-- I know the hands I won, but I believe both were 51% owned by Qualcomm, 49% owned by the partner. On the infrastructure side, they partner with Northern Telecom, Nortel, to do a JV to manufacture CDMA basedation equipment. And then another wonderful, acquired, full circle moment. They call up our friends in Japan. They call up our friends in Japan who, at the time, their US manufacturing headquarters was based in San Diego-- That's convenient. California, very convenient. Our friends-- Sony. I guess Aki Amarita was running it at that point in time. The Sony Corporation to partner in a JV to make handsets. So I actually had a Qualcomm handset back in the day. Probably one of those little flip phones? Yeah, well, that was a lossy with Motorola. No, no, I had a brick phone, like a small brick, not as act more as brick, but a small brick. Because Qualcomm phone, that was made by the JV with Sony. It was a Sony phone with Qualcomm branding. But they're doing all this to be able to answer yes when a carrier's coming to them and saying, well, great. We'll be CDMA. But question mark, question mark, question mark. Qualcomm's like, yep, yep, and yep. We make all that stuff. So you should feel safe adopting us. IP infrastructure, handsets, silicon that goes into both. We got all of it. So we just talked about one, two, and three. We didn't talk about the silicon. And to be clear on the silicon, people know the Snapdragon brand today. This is not Snapdragon. This is not systems on a chip, CPUs. This is not a competitor to Apple's A15. This is literally the silicon to power the radios. And just that. It's to do the encoding, decoding, power management of literally just attenuating the airwaves to send CDMA encoded telephony back and forth. You're making it sound trivial. But this is actually-- this is the final-- I'm not making it sound trivial. Well, that's trivial. I mean, I couldn't do it. Yeah, right. You do it. This is the final just brilliant masterstroke in this long series of brilliant masterstrokes that Irwin and Qualcomm did at this time. I don't know any other chain of just brilliant strategic decisions one after the other. If this had been 10 years earlier, they would have had to do the same thing with silicon. They would have had to partner with Intel or AMD or somebody. TI, Texas Instruments, see this. One of the real men that had-- One of the real men that had fabs, of course, were referring to AMD founder, CEO. I think so. Jerry. Jerry, forgive us a little. So once said that real men have fabs and, of course, was proven desperately wrong over the-- They would have had to do the same thing they did with Sony and Nortel on the semiconductor side. And maybe they could have had some value capture from the Qualcomm IP, but they would have had to partner to make this stuff. But thanks to our acquired superhero, Morris Chang, Fabulous. semiconductors in 1989, 1990, 1991. Just starting to become a thing. Or just starting to become a thing. So they could design their own chips without having to actually have a foundry in-house to make them, and they could outsource that together. They could actually do all the important value-added work. Like it's totally, it's a rigging pentops and smiling curve in this industry, if you go from one to four of the IP, the two manufacturing, and then the semiconductors, all the value, all the differentiation in this industry is in the IP, and the semiconductors, and the manufacturing is a commodity. And Qualcomm would have been a great company if they had just captured the first. They captured the first and the last. They got all of the value, like all of the value, and we talked about on the Nvidia episodes. It was equally crazy, and future-seeing, to know that Fabulous was a thing that foundries were a thing, to be willing to work with foundries, and Qualcomm did it. It's like how many times is this company going to be in the right place at the right time? And just to, you know, the Silicon. And know it. Yeah, and right. And right. Seize it. And we're going to talk more about Silicon and Qualcomm as we go here. But, you know, just to paint the punchline here, today Qualcomm's total revenue is what close to 40 billion annually, I think, of which 85% is their semiconductor business. So without- 37 billion of their 44 billion of revenue is- So, but for this strategic decision, 85% of today's Qualcomm revenue would not exist. And they are the largest Fabulous semiconductor company in the world, bigger than Nvidia, who's number two. Crazy. Totally crazy. It makes sense. They started a couple of years before Nvidia. So, you know, compounding. It's a thing. It's right. So, they pulled this whole freaking thing off. It's just crazy. There's nothing more to say than. It's just one of the most impressive business stories I have ever heard. CDMA gets adopted as a major 2G standard for the next set of phones that come out. 57% market share in the U.S. in 2G, 100% market share in countries like South Korea. They end up getting, I should know this, I'm either 100% massive market share in China, which is adopting, you know, mobile selfie for the first time, and like this is so much. So, the first 1995 is the first year that these networks go live in the U.S. and internationally. Qualcomm does $383 million in revenue in 1995. In 1996, they do $814 million in revenue, oh my gosh, but here's the crazy thing. So, here's another like just wild, you can't make this stuff up. You would think Wall Street would love the stock, Wall Street bets would be going nuts for this stock, the equivalent at the time, not at all the case. The stock is basically flat, Wall Street kind of hates it because the manufacturing operations and the JVs require so much capital and they're tying up all the profits of the company. It gets, the stock gets punished basically all the way up until January of 1999. And a few interesting things happen, are you okay jumping to 99? Yeah, great, let's go on there anyway. So, a few interesting things happen in 99. One, Qualcomm starts to realize it's a pretty serious drag on our business to have this super capital intensive manufacturing operations. We're funneling all of this money that could be free cash flow for the business or could let us reinvest in new R&D into making phones and making base stations. We've got to do something about this. So, in March of '99, they sell their infrastructure business, the base stations to Ericsson, which was formerly one of their competitors. They're a big competitor. There was part of a licensing deal of all the lost, settlement deal of all the lawsuits that popped up between the two companies along the way. They're like, "Oh, great. We'll sell you our manufacturing." And this is basically them looking and saying, "I don't think we need that to bootstrap our strategy anymore. I think at this point, we've got enough momentum that we don't need to make our own base stations. We don't need to make our own cell phones." So, a thousand of the 9,500 Qualcomm employees become Ericsson employees, then they look over at their mobile phone business. Not one fun, not fun at the time, but fun now, a little footnote on that sale to Ericsson. The employees that got transferred as part of that were so freaking pissed that they lost their Qualcomm stock options. They got Ericsson. I don't think they even got equity in Ericsson at all. They actually found a class action lawsuit against Qualcomm to get their stock options back. I mean, over the next 18 months, the stock would basically be Tesla stock. That's this crazy moment that we're about to talk about. December 1999, Kiosera buys Qualcomm's mobile phone business, so they now officially just sell chips that they call QTC, the Qualcomm CDMA Technologies Group, and then they've got a second group, QTL, which is Qualcomm Technology Licensing. It's just, the business model is now set. They make silicon, they make licenses, they sell very high margin revenue. Licenses to their patent or chest, that's the business model for the future. They no longer have this drag on them. And they sell relatively high margin semiconductor designs because they don't fab any of the semis. And when they're selling these designs, they're not just saying, "Here's a chip. Give me five dollars for it." They're saying, "How much do you sell those phones for?" "Yeah, we'll take five percent of that." And you say, "What? What if I want to raise prices on my phones?" And Qualcomm says, "Yep, you'll still pay us five percent of that." And you're like, "What do you mean? I'll just go somewhere else." And they're like, "Where are you going to go?" We own all the patents. And by the way, in addition to paying us five percent of the phones, I think you should pay us to license these patents too. And all the customers go, "What?" And Qualcomm goes, "Where else are you going to go?" You make them sound so evil. I mean, they did invent it all, so they do have a right to monetize it. And Apple did, and the DOJ did, not the FTC, suit them for antitrust. Well, spoilers. We'll get to that. The punchline of all this, after the December of '99, offloading of the handset business to Kiyosera, which is actually a Japanese company. I also had Kiyosera phones growing up. Well, you bought all the good ones. I got all the good ones. Well, you were on a team you may know, right? I was on Singular, which was a GSM network, which got bought by AT&T wireless. It doesn't matter. It all becomes CDM, anyway. Eventually. As we will see, in the year 2000, after this sale, the height of the tech bubble, you know, this is like the benchmark emphasis. We're talking about eBay, e-boys, benchmarks making billions of dollars. Yahoo's going nuts, like it's the internet bubble. It's the tech bubble. And people are looking around. They're like, "What power is the internet? And what's going to power the next generation of the internet?" The single best performing stock for the entire year 2000 is Qualcomm. It appreciates the Qualcomm stock, appreciates 2,621% for the 360 days of the year 2000. I think it was a leap year. Yeah. It's unreal, 26.2X in the public markets in one year, the best performing stock of the craziest year until 2021, until last year in the stock markets. However, you would have had to know just the right moment to sell because it did not stay up there for very long. It would crash down over the next year such that it became only a 4X. From it's pre-1999 high. But if you bought it on the way up, you lost a lot. I'll take only a 4X on my 2021 investment all day long these days. Yeah. Pretty great. So, that's like the core, just crazy business story of Qualcomm to take it from there to today. The next generation of cell phone networks, 3G, which Ben and I probably vividly remember probably many folks listening. 3G 2, 3G, that's when there was a lot of debate, especially in the US, about GSM versus CDMA, and all thinking, naively, you would think at the time, oh, well, all the folks were going GSM, this bad for Qualcomm, GSM switched to CDMA anyway, so basically all 3G was CDMA. In Europe and in the US, just worldwide, they just ran the table. And the reason for that was 3G was all about data speeds. Broadband internet, data speeds. And CDMA was just like the vastly superior technology for-- Totally. You didn't have to encode anything from analog to digital. When you're talking into your phone, you got to encode the signal. But if you're downloading a website or you're sending an iMessage or you're sending a tweet, all that's digital information anyway, so it's already packets. It lends itself perfectly to CDMA's digital-required infrastructure. Then in 2005, Erwin retires as CEO-- I believe, and also as chairman of Qualcomm-- and interestingly, his son, one of his four sons, Paul Jacobs, takes over and becomes the company's CEO. Paul actually has a PhD in electrical engineering as well, spent his whole career at Qualcomm, rose [BLANK_AUDIO] The ranks becomes the CEO. - So an important thing, remember I put a pin in the idea that 20 years from 1985, when they filed that first patent, something else would happen. So Paul Jacobs becomes CEO also in 2005, Qualcomm buys Flarian technologies for $600 million. Now, Flarian did some interesting, like they had some interesting products, but they had a lot of patents that would become essential for 4G. So when we talk to some industry analysts about this, one view was, and I quote, "It was to refill the pot of missiles that Qualcomm promises not to fire at their customers if they pay additional money." So the key set of technologies here were OFDMA, which is, we're not gonna get into it, but it was sort of-- - That's why 4G becomes, it's sort of, 4G was based on OFDMA instead of CDMA, or thogonal frequency-- - Division multiplexing? - Yeah, we're not gonna dive into it, but it was more efficient than CDMA. CDMA, well, it was definitely the night and shining armor versus the previous set of technologies. It didn't quite hold up to the claims or the future proofing of sort of its evolution path that-- - Which makes, by this point in time, it's 20-year-old technology. - Totally, like. - So what we do see here now is after the Flarian acquisition, Qualcomm is able to continue their same exact business model because all of the patents that would be required for 4G and LTE and all that going forward, they own a lot of those too. - It's interesting, so the Paul Jacobs era of Qualcomm from 2005 to 2013, I think, '13 to '14. So somewhere about a decade. - Yep. - I think it's like very viewed at a very mixed light. His big strategic initiative was getting Qualcomm into IoT. IoT didn't really become a thing at least at that time and they said, I mean, maybe it's starting to work now, but yeah, it's starting to work now, but not in the time, everyone thought it did. And it was kind of like a lost era for Qualcomm, but when you look back on it, two things that actually were really great then one was that acquisition and getting, 'cause initially Qualcomm was fighting OFTM and trying to have CDMA still be the standard for 4G, eventually they did pivot and get into OFTM. So that was kind of an initial wrong move, but then I pivot and I save. But two, that's when they started building the Snapdragon and mobile systems on a chip and CPUs and taking on more of the processing on the early predecessors to smartphones. And that would just put them in such a good position for the modern smartphone era. - They sell the high-end Android chip today. I mean, the world is sort of standardized around. Apple makes the A-series chips for your iPhone and if you're buying a high-end Android phone, it's a Qualcomm, whatever. I don't know all the model numbers, but Series 8 Gen 1 or something is the Snapdragon. - Snapdragon, yeah. And they now brand everything Snapdragon. - They do, which makes teasing some of this apart very confusing, 'cause they've just slapped the Snapdragon label on so much that you're like, "Wait, but that's just an RF antenna." Qualcomm says Snapdragon and they're like, "Yeah, fake you out." Like that's the whole point of calling everything Snapdragon. - I mean, it gets to be fair, like the Silicon engineering and the chip design is so complete. Even for like, "Oh, just an RF antenna." Like, that is like a million times more complex than like any processor in a phone 10 years ago. So, it is truly differentiated work that they're doing. But that was, you know, obviously a huge win. And I, you know, to the point that I think today, Qualcomm makes on average about $20 for every smartphone sold in the world, including Apple iPhones. - Yes, so let's get into that. So, I've got the timeline from here. So, going to 2009, this is when like all the litigation really starts to happen. And people just flip from Qualcomm, we think really highly of you and you're a pioneer of technology and true inventors, which they are. They still spend a ton of the company's revenue and reinvested into R&D. But where they really start to be known by their customers and the media and the ecosystem as value capture pioneers. And so, they lose a loss-- - How you capture pioneers. That's a new way, that's another acquired teaching. Value capture pioneer. - Or what's the phrase that I use for Apple maximally extractive over their ecosystem? So, Qualcomm loses a lawsuit with Broadcom in 2009 has to pay $900 million in 2012. Paul Jacobs at the helm makes a really bad bet. Maybe it's a good bet, but bad outcome on a reflective display technology called Marisol. They spun up a $2 billion fab to make it-- - Well, they actually made a fab. - Yeah, there's ultimately zero customers for this next-- The promise was really-- - Real companies don't have fabs. - It was supposed to be like a screen that looks like a magazine page, but they were never really able to reproduce the image quality of the industry. - Right, I was working with the Wall Street Journal at this time and like, oh man. - That was the future. - 2013. - Turns out the iPad was the future. - Yes, Steve Maul and Comf comes in and becomes CEO where I suppose gets promoted to become CEO. Very technical leader. - He was COO before. - Well, COO before, but the problems, problems. They keep growing revenue. They keep doing well as a company, but the ecosystem issues for them and ecosystem reputation continues. So in 2015, they enter into not just an issue with other companies, but now with nations. So they have a licensing dispute with China. You have an activist investor who comes in that same year, John Apartmenters, to try to split up the licensing and the chip business. That activist investor is kind of saying, why do these need to be the same company? The licensing business is printing cash. - At this point in time, many semiconductor companies have split out the actual chip operations and the IP, like a lot of old semiconductor companies are basically just litigation companies at this point. - Yeah, so that's the Broadcom model. So it's interesting to say, okay, what is Broadcom at this point? Broadcom is actually a company called Avago where the CEO of that basically made a bet and said, I think the semiconductor industry is no longer experiencing growth. I think that industry should be harvesting profits 'cause I think it's predicated on Moore's law to accelerating, but basically saying, I don't think that this industry should be reinvesting as much in R&D anymore because it's a settled frontier. And what should be happening is we should be rolling up these companies, so Avago buys Broadcom, takes Broadcom's name, buys some other stuff like LSI Logic. - LSI Logic, oh, big Sequoia win. - Don Valentine's, one of his first various investments. And it's really, the Broadcom strategy is to roll up the semiconductor industry, squeeze them as much as possible. In fact, they're basically a private equity firm. Broadcom is borrowing lots and lots of debt to make the acquisitions that they're making and then squeezing them for profitability. So. - You've got my favorite piece of Broadcom history trivia that Avago, the sort of core of what Broadcom is, actually started its life as Hewlett Packard's chip division. What a sad state of affairs. - Yep. 2015, the company shakes off Janna Partners and doesn't split out the two businesses. I think that was the right call, and I'll tell you why in playbook. But we were talking about Broadcom. 2018, Broadcom comes in and tries to do a hostile takeover at a $117 billion valuation. And interestingly, it was financed by $106 billion of debt. So that company for the rest of its life, I mean, that would basically just be Qualcomm servicing the debt. So interestingly, the Trump administration got involved and said it would be a national security concern and block the deal. And while that may have been true for the reason that the Singapore-based Broadcom was sort of joined at the hip with Huawei. - They did a lot of business with Huawei. - This, I think, ends up being a big win for Qualcomm's lobbyists. I think they had great relationships with the US government and always have since the early days in being a government contractor. And a lot of people that we talked to viewed, or at least that I talked to, viewed this as Qualcomm being able to call in a favor and say, "This is a national security concern. Don't you think?" - We're calling it in the favor now. It's totally true. I mean, this deal was gonna go through and Qualcomm was gonna be everything you were just talking about with Broadcom, which would have been very, especially now. Like, we know about semiconductor. Like, this is one of the huge wins of the Trump administration for America was keeping Qualcomm an independent American company. Like, whether it was Qualcomm calling in a favor or just what, I think we can all look back in 2022 and be like, "This was an enormous win." - Yep. So, in 2017, going back one previous year, both the U.S. Federal Trade Commission and Apple sue Qualcomm for basically the same thing, saying that Qualcomm was using its market position as the dominant smartphone modem supplier to force manufacturers into paying excessive fees. And this is one that I wanna sort of dive in on. We spent a bunch of time advancing through the timeline to really get to this particular point, which I think is a great place to zoom in on Qualcomm's strategic position. today is this Apple lawsuit. So some background. Apple has always used either Samsung processors in the first iPhones until they switched to their own, but they still had to pay Qualcomm patent royalties for whatever RF stuff they were using. So whether, you know, let's treat the CPU as its completely own world, transitioning from Samsung to the A series processors, Apple totally has to buy stuff from Qualcomm. Maybe they could look somewhere else, but either way, they're paying Qualcomm, the licensing for it. Today, Apple does use Qualcomm cellular modems, which started in 2011, and there was just one year where they used Intel, where they did not use Qualcomm. We're going to talk about that. So the way that I essentially perceived this and why Apple eventually initiated the lawsuit is Qualcomm got greedy. They had patents on technologies that were part of standards that were set by industry consortiums all over the world, and they leveraged those patents in basically every way possible. And here's the economics as far as I could sort of sus it out. So they asked Apple for $7.50 per phone sold, which comes to about $2 billion a year, plus an additional 8 to 10 when they were going to raise prices later. And so you quickly get to a situation where that Qualcomm was sort of expecting Apple to pay $17, just two licensed patents, which is on top of the price that they were paying for those base band ships. So rack rate for a base band shipped, and base band ships are the same thing as sort of cellular modems, is $30 a chip. And it's not actually $30. It's more like 5% of whatever the average selling phone price is. Oh, guess what phones have a really high average selling price. High phones. And so if you think about $250 million a year, that is $7.5 billion a year that Apple would be paying Qualcomm, that would be 20% of the QCT revenue, 20% of all of the chip revenue that Qualcomm makes. And further, if you back out the 14 million a year from QCT, their chip segment, that doesn't come from the chips for handsets specifically, but rather there's some other stuff they're working on, automotive, IoT, and this new thing that they're calling the RF front-end radio's product line, which we'll also talk about, this is cool. Apple could make up up to one third of Qualcomm's handset chip revenue. Now analysts have estimated that Apple negotiated down from $30 to $10. Apple's general counsel during the lawsuit at the number $18 slip. So whether it's $10, $18, or $30 a pop, that is an enormous amount of revenue that Apple pays Qualcomm. Again, not for a Snapdragon, not for the CPU, not for the system on a chip, just for the RF cellular modem. Wild. So there's some other interesting things that came out in this lawsuit. Qualcomm asked Apple to speak out against Y-Max, which is a competing technology. They were like, we need to vocally speak out that our competitor is a bad piece of technology. They also stipulated that if Apple ever used a competing supplier, and keep in mind this deal is signed in the early days of the iPhone, if they ever used a competing supplier to Qualcomm, they would owe Qualcomm a billion dollars. So what Apple is basically doing is biting their time for there to be an actual credible competitor, and they had to wait all the way up until the 4G days, until they're looking at Intel, especially if we work with you and we work closely with you, we think you can be a credible competitor to Qualcomm right now. We think your cellular modem's business is close enough where our customers won't notice the difference, and we can tell Qualcomm that we're going to use you and try to get a little bit of leverage there. What Qualcomm interpret that as is, well, now you owe us a billion dollars, because look at our original deal we did. What this basically comes down to from a legal perspective is because Qualcomm owns patents that are a part of an industry standard, they have to charge a price that is fair, reasonable, and non-describentory, or friend is the industry terminology. And Apple is basically alleging, look, you're abusing the market because it's not fair, reasonable, and you're highly, highly unreasonable in the way that you're charging us this. So, around the time of the iPhone XS and XR, those phones actually did use Intel modems, but what was basically happening is the Intel modems were falling further and further behind Qualcomm. Apple was realizing, oh crap, we're going to miss 5G because there's no chance that Intel catches up and can actually develop a credible 5G chip. And so, they end up settling and sort of backing off their big lawsuit with Qualcomm. >>Well, in this way, we're going to escape our technical level of competency quickly if we haven't already. But 5G, it's pretty cool. This is where you were talking about patents. This all sounds so like icky, but the amount of engineering and IP and work that has to go into what we described originally back in the World War II. And it was so crazy complicated to make this stuff work back then. Now, it's just like a factor of a million more. The amount of processing, what Moore's Law has had to come up the curve to enable something like 5G is unreal. There's a dedicated processor in front now of the RF stack to do all the crazy multiplexing that is required for 5G bandwidth to work, right? Yes, so this RF front end, okay, so here's a fun little, so what is 5G? It actually is an open question. When 5G was first proposed, the proposal was to use the millimeter wave spectrum. This super high frequency part of the spectrum that for years people thought was basically impossible to work with because it was, it just requires incredibly sophisticated electronics to make it work. Not only that, but when you have really high frequency, and again, we're right on the edge of our competency here, but when you have really high frequency radios, they can't transmit through a lot of stuff. It doesn't handle concrete well. And so you end up needing a little base station on every street corner. Now, it can give you like 10 gig internet, like it's crazy, but it needs to be really close to you. And so as the telecoms were starting to build this out, of course, the initial review, they say we now have 5G. In fact, they even rebranded a bunch of LTE stuff to be 5G, so it's show up as 5G. You did this, right? All of a sudden, because I was on 18T at the time, used to say 4G and then all of a sudden it just said 5G on my phone. 5G? 5G? You're like really 5G. That's exactly the same stuff I was using before, but now you've rebranded it. So occasionally you'd walk by something that actually had a millimeter wave tower, and it would be like, oh my God, this is the fastest internet I've ever experienced. And then you'd like walk across the street. Oh, I remember like Neelye at the birds doing like, yes, Neelye is the world's expert. Yeah, yeah, yeah. Like on a specific street corner in like New York City or San Francisco getting like 5G's a 10 out of 10. And then you take one step to the rate and you're like, right, back on 4G. So here we are, 2022, five years after the initial hub hub about 5G started for consumers. And what is 5G? Well, the industry has decided to a lot two more areas of spectrum that are not millimeter wave and are easier to work with and are cheaper to build infrastructure for and are slower as 5G also. So now what that does to chip makers is it says if you're building a cellular modem in your phone, you have to have a really complex RF front end or what Qualcomm is calling their RFFE business. The RF front end basically needs to at any given point adjust in real time depending on what flavor of 5G. You're accessing so many different windows of spectrum. Yes, so far across the spectrum bands that like, yeah, there's all man, think about like back to the original Heidelbar and frequency hopping like it was all within one band. Now we're talking about like what they have a crazy number of bands. So Apple, going back to the Apple lawsuit, Apple sort of realizing we're screwed here if we don't have Qualcomm as our customer. So they settle with Qualcomm in this is in 2019. Apple says we will continue using Qualcomm's radios for now. I think they negotiated some discount to the exorbitant fees that they were having to pay Qualcomm. They Apple also paid four billion now switching over to the licensing side of the house to secure the patent licenses over the next six years. I think it was four and a half billion dollars for a six year deal. It's actually unclear who really wins here. I think Qualcomm wins in the short term because Apple's backup solution of Intel's modem fell entirely behind. But in the long term, I mean, what ended up happening is Apple actually bought that division away from Intel and they've been developing their own cellular modems in house. We know based on, I don't know if it was a slip of the tongue or an intentional thing, but we know from the most recent Qualcomm earnings call a week ago that the next version of the iPhone that comes out in November of 2023 will continue to use Qualcomm's chips. Even though Apple has been working on their own. So they're trying to do the PSMI on the modem? Yes. It's ludicrously hard to build the stuff that Qualcomm has built. So even next year's iPhone will have Qualcomm. Wow. Our front-ends amp. And I think they use our reference and cellular modems. But after that, Apple's definitely going to try and take this in-house. But Cristiano, the CEO of Qualcomm, sat on the most recent earnings call, after that, we do anticipate having almost $0 come from Apple in our chips business. So at least their foreshadowing to their shareholders, Qualcomm is that they think Apple is going to succeed at this. It's just going to take a couple of years. Well, this feels like the perfect time to talk about the other strategic chess move that Qualcomm made here. Yes, Newvia. Newvia. So this is another 2021 move. So Qualcomm bought this company called Newvia for $1.4 billion. What is Newvia? Well, Newvia was founded by former Apple Silicon people, including the chief architect of the A-Series chips. That seems like a good get. Yeah. And so-- Back to PASMI. Yes. One way to look at it is this is Qualcomm's ticket into the laptop CPU/system on a chip market. They already make SnapDragons for the high-end Android phones. And soon, they'll be able to make a competitor to Apple's M-Series chips for laptops and desktops and maybe even servers. And phones, too. I mean, iPads, phones, tablets, like-- So this is crazy. This is where it gets interesting. So SnapDragons, for anyone who listened to our ARM episode, you'll remember the difference between ARM makes an instruction set architecture that you can license or you can go big with them and just buy one of the actual ARM design chips off the shelf. Like buying a solution, you might say. Yes. SnapDragons use an off-the-shelf ARM design for their CPU. Apple just uses the ARM instruction set, but has done their own custom design to get the most performance. And that's why Apple Silicon is so far ahead of the competition. The Nuvia team can just do their own custom design of chips and actually be differentiated from stock ARM CPUs, just like Apple is doing. Unfortunately, everything cool about the SnapDragons chip doesn't actually include the CPU. The CPU is just a standard-ish-- Yeah, a standard-ish-y ARM. --the ARM design. So here's-- This is cool. So this is the path for SnapDragons to get on par with Apple Silicon. Yes, and for their CPUs to actually exactly. So but one caveat to this whole thing about maybe they'll do laptops, maybe they'll do servers. Qualcomm actually doesn't really want to do any of that. Qualcomm historically has failed every time they've tried to do servers or watches or smart home or displays. Every time they've strayed too far from their core competency, it hasn't been good. And probably what Qualcomm really wants is 20 bucks from Apple for every iPhone. I think that's a reasonable path forward. The CEO is pitching a much broader story than that to shareholders these days. So what Qualcomm actually wants is for the new via team to sort of like invest where they see the frontier going, where they see a much bigger tam, where Qualcomm sees a multi-hundred billion dollar opportunity, and that is IoT, automotive, and the RF front end. And so they sort of describe phone modems and phone systems on a chip as almost like a legacy business, and they're highlighting these other areas as sort of the growth business as the frontier. Interesting. But either way, new via seems to be the ticket, because if you can custom design chips using the ARM ISA, but be like the performance of Apple Silicon, I don't care what you're putting those in. That's a really good, powerful thing. Well, just, I mean, even like for technology, the technology industry writ large, to have just like with Android, you had a, you know, iPhone rivaling operating system available off the shelf for any kind of application that let a million flowers bloom to have the same thing for Apple Silicon. Like, that's pretty cool. All right, listeners. Now is a great time to talk about one of our favorite companies. Lagora, the agentic operating system for legal work. Yes. Most of you have dropped a contract into a chatbot by now. But as a lawyer, how far can that really take you? It's never seen your firm's precedence. It doesn't respect confidentiality. And you can't check where its answer came from. So Lagora's founders approached this the way that great founders do. Early on, the team moved into the office of a big law firm in Stockholm and learned four months. How the work actually gets done. A couple of thousands of enterprise customers and a 100,000 users later, they still work that way. They employ 100 legal engineers on staff. These are lawyers sitting inside of firms and legal departments to deploy Lagora onto live situations like the deal closing this week or the policy review due Friday. The product has gotten very powerful in a short period of time. A year ago, the goal was a good answer from a chatbot. With the Lagora agent, you hand it the work, and it comes back with a plan. Tasks like running diligence on a data room or reviewing a new vendor contract get completed inside the tool you use every day like Microsoft Word and against your own precedence and playbooks. Every finding is sourced, the lawyer directs the work and signs off before the system executes any steps. Every legal team can get access to the same AI models. So the models themselves do not give you an edge. What you can do with your own data, people, playbooks, that's the whole game. Today, over 100,000 legal professionals at more than 1,800 law firms and in-house legal teams in over 50 markets use Lagora. And when Lagora goes head-to-head with its top competitor in a pilot, it wins 70% of the time. A demo is one thing, but what matters most is an associate who reaches for it at two in the morning during a closing or whether a general counsel trusts what it says before the board meeting. So if your legal team wants to check it out, whether you're at a firm or in-house, go to lagora.com/acquired and be sure to tell them that Ben and David sent you. There are two other small things that happen that I think let's just sort of skip. I'll mention them briefly, but let's get into analysis. Paul Jacobs got kicked off the board of Qualcomm in 2018. He tried to take the company private through a buyout when there was all this sort of tumult about, is it going to be bought by Broadcom, all this stuff? And the board said, if you're going to try and make a hostile takeover in LBO, the company yourself, you can get right off the board. And so there are no members of the Jacobs family on the board of directors anymore. The other thing that happened in 2016 to 2018, Qualcomm tried to acquire NXP semi-conductors, but I think eventually China sort of just like drag their feet enough to kill that. I could tie it up in the whole Broadcom thing. Yes. But quick review of where they are today, and then we'll go into analysis. Qualcomm today has a $120 billion market cap, which two things, one, that's astonishing, that's impressive. They're technological pioneers, and they're amazing at value capture. Two, that is the same price that it was worth at the peak of the.com bubble. Wow, and just about the same amount of Broadcom offered to buy it for, right? Yep. Which is interesting. You know, by revenue, I think revenue, and probably also a number of chips there, the largest fabulous semiconductor company in the world, bigger than Nvidia, but a way lower market cap than Nvidia. Yep, I mean, here's my view on Qualcomm versus Nvidia. Do you bet on the intelligent connected edge, as the CEO, Cristiano Amon, would put it, or do you bet on AI? And they're both mega trends. AI has a far bigger potential, in my opinion, than the intelligent connected edge, which is wonderfully bust. Although I do really have a genuine appreciation after doing this episode for the amount of engineering that goes into wireless technological advances, which almost at a Moore's law, well, that's lower than Moore's law, like pace, but a steady drum beat have continued to improve. And so, I mean, now there's no difference between 5G and home broadband, and that's like understanding on the right street corner. I mean, they do 44 billion in revenue, chips make up, most of that at 37 billion licensing fees make up only 7 billion, but the licenses are a much higher margin business. It's a 69% margin. I think it's earnings before tax margin on licensing versus only 34% for the chips. So there's a super efficient business there in licensing. Revenues are growing 32%. Earnings are growing 47% year over year. This is an amazingly high growth rate company. Yeah, that's pretty awesome. They almost doubled their revenue over the last couple of years too, so that Cristiano's definitely coming up on high. - Cristiano does seem to be doing a good job. - Cristiano's the new CEO as of last year. I think he's been for, been in for about a year. So into analysis, what power do you think that qualcomm has? - Oh, well, patents. Is that a cornered resource? - Is that a cornered, I think that is a cornered resource. Oh, I think, yeah, Hamilton in seven powers, I think he does say patents are a cornered. I think they're in the canonical description of a cornered resource. That for sure, they have at least, maybe still do have network economies in the infrastructure side of the telecom industry and the handset side. Like, you have to, one locks in the other. - One locks in the other. Like, if you control the infrastructure standard, all the handsets will have to use that. If all the handsets use XYZ standard in the infrastructure after you had to be able to control both. I think there actually was a network. effect there. I also think there's scale economies. If you are a fabulous chip company, then it is worth all the R&D. The amount of R&D you're creating a Snapdragon designing and creating a Snapdragon and realized across a huge number of customers. So like it's really hard to start the next Qualcomm. If the frontier you want to compete on is making a better Snapdragon, that's not going to happen. I've got a fun one here. It's both fun to talk about because it always is, but I think actually as a I feel reasonably confident in. I think Qualcomm during the golden years that we told the history of had real process power. I think it was equivalent to the Pixar brain trust. Like that set of people working together under those set of circumstances were wholly unique in the industry and the world. And actually it's interesting. Like I read, you know, besides the Qualcomm equation book from Dave Mock, which is amazing, there's a ton of history out there about Qualcomm, especially in like local San Diego. Like like the lots of local San Diego publications and history books and stuff, especially because the Jacobs has given hundreds of millions of dollars to support the kids. We didn't talk about this. Everyone is. Erwin is one of the great philanthropists of the past century, like undoubtedly. But to UCSD, the UC system so many feel like so much of building infrastructure in San Diego comes from Qualcomm and the Jacobs family. So going and doing all the research all these local San Diego publications and, you know, historical documents, they all talk about the like wealth, the wellspring of startups and other technology companies that came out of Qualcomm. And indeed, there are like, you know, link a bit in Qualcomm. There are like 100 plus in the San Diego area that came out of Qualcomm. But you compare that to like the Silicon Valley, like what came out of Intel, what came out of Fairchild, what came out of the Trade RSC, there's not the same diaspora of success in Qualcomm. Like the plenty of success in, you know, Solana and Toli is part of the Qualcomm to diaspora. So it's not like there's none, but not at the same scale. And I think that actually de facto shows there was process power. Like it was that unique group of people in that unique situation. Oh, that's an interesting sort of like proof by example. Yeah. Huh. Deductive proof. Do you want to talk about the bear in bull case for the company? I have a few. Okay, go for it. All right. So here's the bear case. Qualcomm has very real competition from the low end that we didn't talk about. An example is media tech who not only makes the baseband modem chip, but also systems on a chip using the stock arm CPU designs. So media tech systems are way cheaper than Qualcomm. And I think they actually just surpassed Qualcomm in terms of number of units shipped. And so all the low and mid-end Android phones are using media tech. And so Qualcomm kind of needed to buy new via in order to differentiate the CPU and not just be using the stock arm design that media tech and everyone else is using on much cheaper chips. Historically they failed that everything that was not a phone that we talked about before. And now they're sort of saying the future is IoT and automotive. These things that are not phones. We'll see. They're just constantly in lawsuits. I mean, we didn't talk about this, but like China, South Korea, EU, Taiwan, all these companies, all these nations have sued. Somebody law firms must just be making a fortune off of industry. And the last one for the bear case for me is I really think that they finally poked the bearer talking about their customers enough to make them want to actually do something about it. The goal for Qualcomm should have been make as much money as you can without pissing people off too much. And I think over the last decade, they really upset Samsung, Apple, so many people that are starting to at least make their own radios or even consider systems on a chip. And so now that there's very viable alternatives for Silicon that people could either use in-house or competitors coming around at different angles, Qualcomm may lose their leverage to actually get a royalty out of each phone sold. Now licensing business is going to continue to be a juggernaut, smaller and revenue, but higher in margin. But, you know, that is the sort of bear case on the current Silicon business. Now, the bull case, like maybe the lawsuit thing is actually a bull case. They managed to keep making more and more money and have been reaffirmed over and over again in a bunch of jurisdictions that, you know, they settle their way out of these lawsuits or they, whatever, but they're able to keep making tons of money. The big bull case is you believe that this shift to automotive IoT and 5G RF front end is real. And so for those keeping track at home, everything I'm about to say is a part of the chip segment that does that $37 billion in revenue. Automotive does $2 billion in revenue. That's a very real business. The RF front end business that we were talking about, that does $4 billion a year in revenue. It's interesting. I mean, we rented a car here in Lisbon and for the family. And, of course, it has data built in, you know, 4G or 5G data, right? As does like just about every new car these days. Yep. The IoT segment is now doing over $7 billion a year. Qualcomm thinks overall this is a $100 billion opportunity. There's a bigger narrative that Cristiano's trying to uspouse around this intelligent connected edge that they call a $700 billion opportunity. That's getting a massive number. I know it reminds me a lot of the the Nvidia slide that talks about their trillion dollar tam. I mean, they're executing very well, but I think they're trying to sell a story in terms of addressable market that is hand wavy. Yeah. All right, playbook. So in the early days, this is a thing that we didn't talk about. We talked about the some of the ecosystem stuff, but there was this incredibly delicate dance of needing to be the best supplier to win deals, but also have other credible suppliers. No phone company was going to take a dependency on the CDMA technology when just one vendor existed. And so they need to evangelize and create their own competitors so that their customers could feel safe with this new technology. But of course, as long as they kept some things secret of how to eek out the absolute best performance from the innovations, they actually could still be the leader. So it was like figure out how to get a bunch of other people just good enough, which is fascinating. It's such an amazing case study in bootstrapping an industry. Yes. Yes. Similarly, they had a clever tactic in their IP strategy. So at Qualcomm, where I think they have something like 17,000 patents now, there's a decision every time there's a novel piece of technology about whether they should patent it or keep it a trade secret. And there's enough things patented so that you can't achieve any of these things, these magical things that we've been referring to all episode, these layers of magic without paying Qualcomm. But they don't patent everything because they want to keep an advantage for like consulting revenue or implementation fees or signing big deals where they say, not only do you get access to our patents, which may expire at some point, but if you work directly with us, you get access to the trade secrets. And you can pay us to basically generate services revenue for you to work with our engineers. I was thinking about this for playbook as we were going to. There's this really interesting dynamic to this industry that lends itself well to the IP and patent monetization scheme that Qualcomm has adopted, which is that the successive generations of wireless network Gs happen just fast enough that it's within the patent lifetime. So that like, you know, all that core CDMA patent, like all those patents are expired now, but it doesn't matter because we're so many generations beyond that like those patents are now worthless. So you get all the useful life during the protection period of the patent. And then when it's, you know, today, it's not like a generic drug where like, you know, Advil is still or Tylenol or whatever is still like, you know, useful. Right. That's a great point. It's also interesting that if you miss the window, like if Qualcomm had missed the window in the early 90s of evangelizing the technology for 2G, they made out of survive long enough to catch the next window 10 years later for 3G. So this is like one of the few industries where there's these super quantized time windows that exist when you can actually get in. Yep. Another one that I thought was pretty interesting, because I mentioned I think the business is actually makes sense together. The licensing business offers Qualcomm predictable, high margin revenue that they can basically use to fund R&D. So because they know they're going to keep getting that and because it's a big revenue stream, it lets them sort of take bets on new R&D. And when they do more R&D, that fuels the flywheel where they both get new products and they get more IP that they can continue putting into the licensing flywheel. So there is, I think there is a credible argument of why you want to keep them together. There's also a-- Qualcomm makes that argument explicitly. Totally. The not very credible argument is this thing's a cash cow and we want to keep our rich uncle around to make this a nice place to work. And you know, like they have several, I think, of nine airplanes. Like it's a relatively cushy company for what I understand. Well, San Diego is a very nice place. Yes. I do think the big picture is that the US government's patent system has granted Qualcomm a monopoly. And I think there's like, this is one of the few things we've covered on the show where The business exists because of the US's regulatory system. They've basically said, and then reaffirmed in a lot of these rulings, you are allowed to capture a ton of value from this. There's so many good debates about whether the patent system exists and serves its intended purpose of enabling people to spread the news about their innovation, so other people can add it, and the way we compensate you as we give you a 20-year exclusivity window, or whether something like this is an abuse of the system. But there's no way to argue that this is anything but a perfect execution of the game on the field. Yeah. It strikes me telling this whole story that, like, think about early stage venture capital company building and the like, you know, you said Ben, who would tell this story. If you were to give a venture capital as the Qualcomm pitch, and like, there's so many, there are like at least six or seven different hops where, you know, X and T, it looks like, well, and then a miracle happens, and then we succeed at this, and then another miracle happens, and then we succeed at that, and like usually, you know, my pattern matching as an investor in early stage companies is like, any time there's a single, and then a miracle happens, automatic pass, like, because if you're betting on a mirror, but sometimes, if you have a team, because this wasn't just like, and then a miracle happens, if you listen closely and like, really knew this team, they like, really knew, they had really high degree of confidence that all of these tight, you know, threading the needle moments were going to happen. Yep. And really doing a degree that just blows my mind. I've never heard anything like it. Yeah. And it just makes me think that like, some time, like to maybe just be a little more open to that, you know, that like, some time, like, some person off the walk, walked in off the street and said, like, gave you the Qualcomm pitch, for sure, it would not work. For sure. And the hardest thing about being a technology investor or someone participating in this ecosystem in any way is, it's a power law dynamic. This is a business of exceptions. And I've seen, and I'm sure you have to, so many counterfactuals too, where incredibly credible teams walk in off the streets with miracle, like, then a miracle happens and, yeah, it still doesn't work. Like, you know, but sometimes, but sometimes it, but sometimes, it never works, but sometimes it does. So that makes our industry fun. All right. So we're going to not do grading because we've decided to kill grading until we otherwise resurrect it. But I do think it's worth articulating a little bit of a takeaway. So my takeaway on Qualcomm is the last decade was basically the best decade for their business model and being in the right place at the right time to have an incredible business model around capitalizing on mobile. And in order for the next decade to be as successful, they need to be absolutely correct about their growth businesses around IoT, around automotive, and around whatever the intelligent connected edge ends up describing, because I think those are technologies that we don't quite know what they are yet. I think if they continue to try to run the same playbook in just the handset market that they have been, the best days are behind them because people have caught onto their games a little bit and are going to squeeze them from a bunch of different directions. Yep. Well, yes, totally agree. I think to paint the best version of the intelligent connected edge that I've heard Cristiano articulate is you sort of put plainly like, hey, we all agree that like the cloud is like I think, like we did the AWS episode, there's over $100 billion in like revenue backlog in the cloud. Yeah. We talked about it on the AWS episode like Snowball and snowmobile, like getting data to going from the cloud, so like still like one of the major pieces of lock in. And like you think about how data gets in and out of the cloud. Most of it's not by snowmobile, so most of it is wireless connected on the edge. And so if you think about it like that, you're like, okay, yeah, I can buy that this is a, you know, trillion dollar market. But how do you capture value in that and can they capture it in the same way that they have in the past? Like very much open questions. Now is a great time to talk about one of our favorite companies, Statsig. Yes, there is a reason why the best product teams rely on Statsig, whether they are iterating on their core product features or shipping AI-powered experiences at scale. Yep. In the crazy speed of today's AI world, shipping fast is just table stakes now. It's basically trivial to build and deploy your app constantly. The real advantage is how quickly you learn what changes actually created value for customers. And how fast you can use that signal to guide what you shipped next. This is where Statsig comes in, it brings experimentation, feature flags, and product analytics into one unified system so teams can ship safely, test rigorously, and directly link what they changed to how users actually behaved. So if you want to make learning your competitive advantage, whether you're building new AI experiences or just evolving your existing core product, go to Statsig.com/acquired to get started. Oh, listeners, that was a total blast. David, crazy to do a live show like that with no guest for two and a half hours on stage just you and I. Yes. And a professionally operated boom arm camera. Yes. If you haven't watched the video version of this, just go check it out on YouTube or Spotify or anywhere just to see what that looked like, it was a very fun spectacle to get to do that. Our huge thank you to the Salona Foundation for hosting us at Breakpoint this year. It's a really great event and fun to be in Lisbon. When you finish this episode, come talk with us acquired.fm/slack13000 other smart, thoughtful kind people. If you want some of that sweet acquired merch, everyone is talking about go to acquired.fm/store. I know in the next few weeks there's going to be a couple of new designs dropping inspired by catchphrases from episodes where I applied my graphic design skills for better or for worse. Is the perfect time to sign up as a customer for Brex to get one of those? Yes, that's right. If you don't want to pay for your t-shirt, brex.com/acquired, much cheaper way, you also get to be a Brex customer. So, wins all around. Win, win. If you want to listen to the LP show, we have had some awesome, awesome episodes recently. We just interviewed Jay Hogue, which is a super rare interview to get. Jay is the founder of the $21 billion firm TCV, formerly technology crossover ventures, about their story and his personal philosophies. TCV was a major investor on much of the journey of companies you know, like Zillow, Spotify, and Netflix, which we spent a lot of time talking with Jay about. You can search acquired LP show for free, publicly in the podcast player of your choice to catch that. Fat listeners, we'll see you next time.