Speaker 1Today we have some real spy novel stuff, as the CEO of NVIDIA admits to having an affair with a woman the U.S. government believes is a Chinese spy. And while this is not the normal coverage for this show, NVIDIA is not a normal company, and this is not a normal circumstance. We are talking about the biggest company in the world, the anchor of the AI build-out, whose leader, who is implicated in all of this, has been perhaps the tech industry's strongest advocate for more normalized trade relations with China when it comes to AI. The implications, in other words, are big. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, Robots and Pencils, CDATA, and Blitzy. To get an ad-free version of the show, go to patreon.com slash ai daily brief, or you can subscribe on Apple Podcasts. And to learn more about sponsoring the show, send us a note at sponsors, at ai daily brief dot ai. Now, friends, I am excited to be back in the Hudson Valley recording as normal. Thank you to all of you who put up with the additional nature sounds from last week's episodes. And the one more thing I wanted to mention before we get into today's episode is that the third edition of the new podcast, How We Got Here, has just released. We are just coming off an election cycle where more than $3.2 billion was traded on the prediction markets, Calci and Polymarket. Calci saw nearly double the amount of betting on these midterms than it did on the AI daily brief. So, I'm excited to be back in the Hudson Valley recording as normal. Thank you to all of you who put up with the entire 2024 election cycle, including the presidential election. Our third episode of How We Got Here, then, is about how gambling on everything from our phones became normal. It's a look at America's long and complicated relationship with gambling and betting and all their various incarnations, leading up to the prediction market moment that we're in now. You can find it all, including the companion essay at howwegothere.com, or look up How We Got Here with NLW on any podcast app, or just follow the link in the show notes. Thanks to everyone who's checked it out so far. And now, let's get back to the show. Welcome back to the AI Daily Brief headlines edition, all the daily AI news you need in around five minutes. And man, between this one and the main episode today, we are in a bit of a chaotic moment in the AI story. One of the big questions recently was with all of the negative narratives swirling around the markets, Anthropic would proceed with their IPO. The past few weeks have featured collapsing valuation estimates for what was expected to be the largest IPO in US history. Two trillion dollars seemed right off the table as investors told the press they expect 1.6 or 1.5. The number went even lower after last week's sell-off, with Bloomberg reporting on Friday that some institutional buyers were looking for 1.1 trillion. Below the headline, the article said that some buyers were discussing 1.2 or 1.3, but we now have a new low appearing in the financial press. Now, Anthropic raised at $965 billion back in May, making 1.1 perilously close to even money for the most recent batch of investors. And that's, of course, assuming the stock holds value through the first few months until lockups expire. Michael Field, the chief equity strategist at Morningstar, thought Anthropic would likely take what they can get, commenting, the cost both physically and reputationally of delaying the IPO might hurt the valuation more in the longer term, whereas if markets hold up at the level that they are, they'll just get it done at a slightly lower valuation, suck it up, and move on. And that certainly seems to be the path Anthropic is taking. Later on Friday, the Information reported that Anthropic will begin their roadshow this week. Sources said that plans are still in place to IPO ahead of the Thanksgiving holiday and as soon as next week. The Financial Times reports that Anthropic is still aiming to raise $80 billion at a valuation of $1.5 trillion, despite the negative headlines. Referencing the news that is the subject of our main episode, open model researcher Nathan Lambert wrote, if tech stocks continue sharply downward on the Jensen news, looking back, this could be what we think of as the first major correction in this boom. Debt went too far, the election, Jensen all in short succession, and Anthropic gonna IPO at the bottom? Andrew Curran, though, pointed out the good news. The AI news aggregator wrote, Anthropic's valuation has now been cut roughly in half from the lofty $2 trillion heights that were being spoken of in financial circles only a month ago. On the plus side, if you are a believer, you can now get in for about half price. Now, I'm not sure it actually lands all the way down at $1.1 trillion, but I do think it is absolutely the case that the lower the valuation is, the better it is for retail investors who are in any circumstance providing an awful lot of exit liquidity from the early investors who had access to the Anthropic deals. OpenAI has also closed their latest funding round, also below initial expectations. When the round was first reported in late September, OpenAI was said to be targeting a $30 billion raise at a valuation of $1.4 trillion. By late last month, both numbers were coming down quickly, with only $11 billion in funding commitments. Sources have said that OpenAI could extend the offer to family offices and high net worth individuals to make up the difference, but Bloomberg now reports that the round has closed, with OpenAI securing $11 billion in new funding at $1.25 trillion. Sources said they will continue fundraising in the near future, but they are not sure if they will continue to do so. At this valuation, looking to get to $20 billion in total, they are reportedly working with investment banks to raise from high net worth individuals and retail investors. Still, it looks like they will be well short of their initial $30 billion target. The institutional portion of the round included Kotu, T. Rowe Price, SoftBank, Sequoia, Founders Fund, A16Z, and a syndicate of backers from the United Arab Emirates. One interesting sub-story was existing investors like Kotu and SoftBank putting in relatively modest amounts in what seemed to be a defensive move. SoftBank invested $60 billion over the past two rounds. The New York Times reports they were only in for hundreds of millions. SoftBank's stock price was down 20% over the past week, contributing to a 60% drop since their June high. In addition, reports circulated on Friday that a mystery seller was dumping half a billion dollars in data center debt. Initially believed to be linked to SoftBank, the seller was later revealed to be Dutch bank ING. Only $25 million of the half a billion was tied to the New Mexico Stargate facility, jointly banked by SoftBank, Oracle, and OpenAI. Still, Elon Musk threw some more confusion into the mix, posting on Friday that Now, with sentiment beginning to sour on the Frontier Labs, OpenAI CFO Sarah Fryer was deployed to CNBC to shore up the narrative. Touching on the fundraising, Fryer commented, A good CFO always raises money when she can. Fryer said that OpenAI's business grew 70% over Q3 with enterprise revenue doubling. She commented, When I think about the midterms, I need to keep thinking about how do we keep investing to meet that customer demand. If the midterm means more focus on data centers and communities, we're going to keep doing the work to get into those communities, and to build trust with them. Still, Fryer argued that AI growth isn't in question, stating, The demand for AI is real. I don't think anyone's arguing with that. Or, if they are, I think it's a losing argument. Another big blustery headline comes from Microsoft ending a compute rental arrangement with Oracle. Bloomberg reports that Microsoft declined to renew the contract, which has been in place since 2023. This deal was one of the first signs that GPU supply was getting tight as ChatGPT adoption skyrocketed. Sources actually said Microsoft had ended the deal because the aging, chips involved with it no longer suited their needs. Reporting suggests that the deal might have been focused on A100s, which are at this point three generations old. And what's important about that is that even if it's not nearly as dramatic as the media headlines might make it seem, this would be the first time we've seen any real evidence of older chips hitting the end of their depreciation cycle, given that that has been one of the key bear cases from people like Michael Burry over the past year that in and of itself could have some interesting implications among investors. Certainly, Microsoft has been the most conservative of all the hyperscalers during the AI buildout. And this move is not in isolation. Last month, they pulled out of their agreement with Nscale to build a data center in the UK, reportedly over power delays. And more recently, sources told Bloomberg that Microsoft is rethinking all of their neocloud agreements after facing numerous delays and rising costs. The TLDR here is that there isn't any clear evidence that Microsoft's decision making has anything to do with AI demand. Instead, it seems like they're signaling that data centers with power connections and the latest GPUs are vastly more valuable than those that are still under construction or even worse, just promised for the future. Now, one other similar story showing how the hyperscalers could stand to benefit from neocloud wobbles. Google has taken over a troubled data center site from Crusoe. The unfinished Wyoming project struggled to secure tenants after running into energy and construction delays as well as escalating costs. It was initially announced last year with plans to scale to 10 gigawatts of power. Google, who had been a prospective tenant, has now bought out the project and plans to invest 17 billion into the development. Sources told Bloomberg that Google had been pressured in Crusoe to leave the project since the summer over cost and timetable concerns. Crusoe is still working with Google on a separate buildout in Amarillo, Texas, and is also the developer behind OpenAI's data center campus in Abilene, Texas. Lastly today, if you want to get a sense of where AI leaders' heads are at, they're not just thinking about markets and data centers. They're also discussing, at least in vague terms, a huge speedup in AI-driven scientific discovery as new models come online internally. On Saturday, DeepMind VP Pushmeet Kohli posted, "Across our projects at Google, DeepMind, we are witnessing a genuine phase transition in the pace of scientific discovery. You will see a glimpse of this in our releases in the next few weeks. As this AI-powered acceleration unfolds, our focus will be on empowering the broader scientific community to harness these tools responsibly." Others quickly chimed in. Meta's Tom Mann commented, "Wait till you see what we've been up to at Meta." And OpenAI's Greg Brockman posting, "Frontier models with lots of test-time compute are becoming the engine of scientific discovery." The sense is that we are in for big things. News aggregator Andrew Curran wrote, "I think all three big labs have achieved autonomous research, and their internal models can now choose their own research directions. We will begin to see the fruits of this work this month." AI educator and content creator Riley Brown echoed that sentiment, writing, "It feels like nothing has happened in AI since Opus 5-5, but that is about to change. The next two weeks will be very, very fun." For now, however, that is going to do it for the headlines. Next up, the main episode. A new study from KPMG in the United University of Texas at Austin found that when people work with AI, Similar skills don't guarantee similar outcomes. Researchers studied more than 500 early career professionals and found that the best performers consistently amplified the value of AI by guiding, evaluating, and refining its outputs. These top performers, called AI amplifiers, weren't defined by what they knew alone, but by how they worked with AI. Learn more about what separates AI amplifiers from everyone else at kpmg.com slash us slash AI amplifiers. The best teams don't have a single star carrying everyone else. They know their own strengths and each other's weaknesses and play to both. That's the team Robots and Pencils has built on purpose. Nobody there is grinding through busy work to pad a headcount number. People come for the hard problems and they stay because everyone around them is leveling up at the same time. In a market full of companies that are just trying to hire fast, that's worth a look. Check out robotsandpencils.com slash careers. Token spend is climbing, security is holding agents back, and no one knows how AI is truly being used. The root cause? AI is cut off from the systems and context your business runs on. That's why CDATA built Connect AI Gateway. One control point between your AI and your systems. Built on the connectivity layer powering Palantir, Google, Salesforce, and hundreds more, each agent gets exactly the data and tools it needs. It governs down to the record and lets you finally see your org's complete AI usage in one place. Accurate answers, a fraction of the token cost, and a gateway that learns your business with every interaction. See how at cdata.ai. Every AI coding tool on the market does the same thing first. It starts writing code. Blitzy does the opposite. Before writing a single line, Blitzy spends days reverse engineering your entire code base. Thousands of agents ingest millions of lines, mapping every dependency, every undocumented constraint, every architectural decision made over the last decade. The result is a dynamic knowledge graph that understands your software the way a principal engineer would after 30 years in the building. Other tools guess at context with grep searches and markdown files. Blitzy never guesses. It builds true understanding first, then delivers over 80% of entire software epics autonomously. Validated, end-to-end tested, production-grade pull requests. That's why Fortune 500 engineering teams trust Blitzy with the code bases that matter most. See for yourself at Blitzy.com. That's B-L-I-T-Z-Y dot com. Welcome back to the AI Network. The AI Daily Brief. And good lord, this episode. In a plot twist ripped straight from the pages of a Tom Clancy novel, NVIDIA CEO Jensen Huang has admitted to having an affair with a suspected Chinese spy. Now, you might be thinking to yourself, this is a bit more salacious than what we normally cover here at the AI Daily Brief. And indeed, if this was just a question of Jensen's moral failings, and whatever the implications were for his personal life, we would steer well clear. The fact, unfortunately, that it has such geopolitical dimension, and the fact that it has major market implications, unfortunately means it's something that we are going to have to cover. Now, the New York Times first reported on the affair at the end of October, in a highly detailed and carefully constructed article. At the time, they didn't have the sourcing to fully describe the nature of Jensen's relationship with Alice Huang, who, of course, despite sharing the same last name, has no relation to Jensen. Alice is an executive at Megaspeed, a Singaporean cloud provider that is a large customer of NVIDIA. Megaspeed was known to provide Chinese AI labs with remote access to cutting-edge NVIDIA GPUs that were banned by export controls. While that remote access is not technically illegal, China AI hawks have been complaining about it as an obvious loophole for over a year. U.S. officials suspect Megaspeed has also smuggled chips into China, but that allegation hasn't been proved. The biggest bombshell from that initial reporting was that Alice Huang was suspected of being a Chinese intelligence agent, and that President Trump had taken no action after being briefed. As for Alice's relationship, the Times merely referred to it as a close relationship and presented Jensen as a target of foreign espionage. Even at the time before we had all the details, concerns were being raised that Jensen might have shared confidential information, or even state secrets with Alice, gained through his close relationship with the White House. NVIDIA rejected the speculation at the time, claiming that mere customers do not get access to sensitive information. During an interview with Bloomberg last week, Jensen said his interactions with Megaspeed were, quote, completely normal and not unusual. He added, we've studied all their deployments of our GPUs and counted every one, and I think the U.S. government also has as well. A Chinese foreign ministry spokesperson said last week that the rumors were, quote, unsubstantiated and purely a smear against China. And yet, in their bombshell follow-up on Friday, the New York Times had all the sourcing they needed. They concretely reported that a memo circulated the White House last spring, which identified Alice as a spy for China's Ministry of State Security. The existence of a widely circulated memo and the assessment from U.S. intelligence agencies had not been previously reported. From there, things escalated quickly. The Financial Times reported that the U.S. intelligence community believed that Jensen's relationship with Alice was romantic in nature. The article said that national security officials had privately expressed frustration with Jensen and noted that he was almost left off the president's trip to Beijing in May. You might recall that the president intervened personally and picked Jensen up in Alaska before continuing the trip. On Sunday night, Jensen confessed. In a company-wide email, he wrote, I was unfaithful to my wife. My actions hurt her and our children. There is no excuse for what I did. I am deeply sorry. The rest of the email was a lengthy apology to his family, his employees, his board, and business partners. He did not address the geopolitical angle of the scandal. In a separate communication, the NVIDIA board announced an independent investigation assisted by outside counsel. They presented Jensen as a target of a foreign intelligence effort and wrote, quote, The board said that Jensen is committed to cooperating with the investigation. But at this stage, the board also expressed confidence in Jensen and the rest of the leadership team. As of Monday afternoon, China time, officials are still denying involvement. The Chinese Foreign Affairs Ministry spokesperson said, We have already responded to this matter. The allegations regarding so-called Chinese spies are pure groundless speculation and slander. In terms of reactions, many in tech were quick to come to Jensen's defense. Sequoia partner Sean McGuire said, One, Jensen is one of the greatest entrepreneurs of this generation. Two, he did something incredibly dumb. Three, it was bad for America. Four, but my guess is there was no bad intention. Five, two mistakes don't make a right, and I believe Jensen is irreplaceable for America's AI leadership. Investor Bill Ackman wrote, On alert, Jensen Huang is human, and he is an incredible human. Even incredible humans make mistakes. He takes full responsibility as he should. We should all support him and his family. Onward. Folks disagreed with that assessment. New York Magazine writer Alex Brunzini-Venda wrote, Tech guys will log on here and declare, unprompted, that cheating on your wife doesn't count as bad intention. Matthew Iglesias wrote, I don't think anyone gets seduced by a foreign spy with the intention of harming national security. That's just not really what the issue is. And this idea that the intention wasn't the point was a sentiment shared by many. AI commentator Samir on X wrote, The fact that he was successfully seduced is one thing, but the more problematic thing is that he let his affair guide his personal life. and professional actions in ways that have endangered American interests. It's so incredibly reckless because it will now require us to question every decision Jensen made with regard to China and foreign policy with China. It's not even clear if Jensen knew he was being played in making these statements now, and I'm sure he doesn't know either. If removing him nukes our AI policy, we have a key man risk issue that's an us problem. Think tanker Ryan Fedesiek made a slightly different but related point. The issue at hand, he wrote, is not whether sensitive information about NVIDIA or its products might have been compromised by Chinese intelligence services. Though that is obviously a concern. The issue at hand is whether Mr. Huang's confidential personal relationship with his customer might have weighed materially on any decision to export NVIDIA's products to Megaspeed or to defend the company's broader client relationship with Megaspeed, even after it was revealed the Commerce Department was investigating that company for surreptitiously diverting NVIDIA's products for resale to China in contravention of US law. Any independent investigation should get to the bottom of that very obvious apparent conflict of interest and spillover risk to US national security. Politics commentator Dmitry Alperovich put it more bluntly. Not only did Jensen Huang have an affair with an alleged Chinese MSS agent, putting himself in a classic position of being set up for all types of blackmail that could have been hugely detrimental to both NVIDIA and more importantly, the United States, but he had an affair with the managing director of what is an effectively and very clearly a Chinese shell company, Megaspeed, a company that had been illicitly funneling billions of dollars of compute to Chinese AI labs that should have gone to American companies with what is almost certainly his full awareness and support. Unless he is a complete idiot, which he is plainly not. This can be called many things, but one thing it is not is not doing anything to disadvantage America. Indeed, the point that some are making is that hold aside all the trade secret stuff and the accusations that she was working as a spy. Even without that, there's still the export control dimension. A Rao on X writes, I suspect she wasn't a spy after state secrets. She merely wanted GPUs that are under trade restrictions. Now, where I think a lot of folks will ultimately come down on this was the position summed up by Signal who wrote, no one in their right mind should condone fidelity. That said, if an independent investigation finds that Jensen's actions posed no national security risk and compromised no company interests, then I think the matter should remain between him and his family. At that point, it's simply a private moral failure. Still for many, the idea that there are no implications and that this is just a personal matter seems highly unrealistic. And some in fact are arguing that the sourcing of the key details is itself not just good reporting, but its own strategic maneuver in the battle over what the US's relationship with China on AI and chips should be. So, Taiwanese semiconductor analyst Andrew Liu Liu, as reposted by Citrini's Jukin, wrote, quote, The purpose of this scandal is by no means to expose an affair. It is to force through three critical objectives that had made no headway in the tech war between the U.S. and China because Jensen Huang opposed them. Now, the three objectives that they articulate were one, to completely destroy the legitimacy of Huang's lobbying against sanctions in Washington, making the point that Jensen has been the most successful business leader in opposing a total ban on chip exports to China. The second objective, they argue, is to reverse the H-200 export approvals for China and sever the link between Chinese shell companies and Southeast Asian smuggling networks. And the third objective, they argue, is to, quote, put NVIDIA on a national security leash through congressional oversight and an independent investigation. They wrote, As the world's largest company by market cap and the dominant force in AI, NVIDIA has until now enjoyed broad autonomy in how it runs its business. But in the wake of this incident, the three congressional committees responsible for China, foreign affairs, and intelligence have begun examining NVIDIA's internal security measures directly. They have even pressured NVIDIA's board into forming an independent investigation committee to look at the company's own founder and CEO. In effect, Washington's national security establishment has planted a watchdog right inside NVIDIA's board. From now on, the goal is to make sure that everything, down to where each and every one of NVIDIA's high-performance chips ends up, conforms strictly to U.S. national security strategy. Now, all of that is based on the idea that some amount of this story was leaked to the press with some level of intention from the intelligence community. And seemingly compounding that point, NVIDIA's national security establishment has planted a watchdog right inside NVIDIA's internal security system. The New York Times dropped another bombshell on Sunday night, reporting that due diligence from NVIDIA and the Commerce Department has been extremely lacking. Writes the Times, The Commerce Department has approved dozens of Chinese companies to handle NVIDIA's H200 chips, some of which have ties to firms under sanctions over national security. Several of these firms are reportedly just empty office buildings. The Times focused on one company called New Matrix, whose listed headquarters in northwest Beijing has no light fixtures, let alone desks and chairs. The Commerce Department approved an export license for New Matrix, but the Times writes there is little evidence the firm actually exists, raising questions about who in China is behind the company. They claim that several similar shell companies have ties to firms that are blacklisted by the U.S. government or work as military contractors in China. Now, prior to the entire firestorm kicking up over the weekend, NVIDIA said they were stepping up their tracking of chips bound for Southeast Asia. That region has been a hotbed for data centers providing remote access to advanced chips. On Friday, NVIDIA announced that data centers in the region will need to enroll in their fleet intelligence system, which tracks the locations of GPUs. Along with other metrics, it could be just a coincidence, or it may have been NVIDIA taking proactive steps to get ahead of the scandal. Now, when it comes to the government, there are a few interesting dimensions of this. Secretary of State Marco Rubio is showing up as a character in the palace intrigue. The New York Times reported that Rubio oversaw the drafting of the classified memo that was circulated in spring. This was the memo that stated that Jensen's relationship with Alice was romantic and that she was believed to be a Chinese spy. The memo was reportedly circulated ahead of Trump's trip to China in May, although sources didn't say, Jensen's exclusion from the trip was related to the memo. Trump himself, meanwhile, so far has stood by Jensen, saying that he should remain the CEO of NVIDIA. He told reporters, I have not spoken to him about it. I just know he's a fine person, but you never know what is going on. Asked whether Trump would invite Jensen on his trip to Asia later this month, Trump said he wasn't sure but would happily bring him along, adding, I haven't heard he's done anything wrong. We will see if that continues to be the president's position as the story evolves. Certainly the Dems are gearing up, and this story showing up the week them an obvious target. Senator Chris Coons, who will chair the Senate Foreign Relations Committee from January, wrote on X, The more I learn about the insane relationship between the Trump administration, Chinese spies, and AI billionaires, the more I'm convinced there is nothing this administration does that isn't corrupt. Congress will investigate and get answers. And yet it's not just Democrats, as Republicans are also concerned. In fact, before the story broke, the chairs of the House Committee on Foreign Affairs, the Select Committee on the Chinese Communist Party, and the Select Committee on Intelligence summoned Jensen to a closed-door meeting later in November. The three Republican leaders wrote, Now, in terms of market reaction, as best I can tell so far, the market is showing that it is, on the one hand, genuinely concerned about this, but on the other hand, still full of investors willing to buy any NVIDIA dip. The news initially triggered a huge sell-off in the S&P 500 at 6pm on Sunday night when futures opened, erasing around $300 billion in market cap. However, within around five minutes, the entire market opened. was erased and futures were trading higher. The Kobayashi letter wrote, Now, I'm recording just a little bit after market opening, and it seems to be a similar story. NVIDIA was at its worst down roughly 8.8%, but then buyers came in and recovered roughly half the dollar loss, and even as I'm recording, you can see the candle swinging wildly. When I started this segment about five minutes ago, NVIDIA was down 5%, but then even as I was talking, a big chunk of that was clawed back, with about $40 billion in market cap bought back in the last five minutes. Perhaps more importantly, at least as of now, there aren't any indications that this is spilling over into the rest of the market, or even into tech, which is mostly, for the moment, in the green. And yet, whatever the initial market story is, the event still could cast a long shadow. Investor Jim Osman wrote, Forget the affair, that's a personal matter. The real issue is whether the man running the world's most important AI company has been compromised by foreign intelligence. Markets can price bad earnings. How do you price a CEO you can no longer fully trust? Even eternal optimist Dan Ives writes, The big question every NVIDIA and tech investor is asking this morning, will Jensen stay on as CEO of NVIDIA? Now for Dan's take, he says Jensen is not replaceable, and that is worry for the street. Others aren't so sure. And in fact, that indispensability is exactly the question being asked. With Andrew Ross Sorkin's dealbook article this morning being titled, How Indispensable is Jensen Huang? There is a lot more to this story, but for now we'll leave it there. For those who want to stay well away from any of the personal elements of this, really the key things to continue to watch will be market implications, further investigations by NVIDIA or external sources or the US government, and the ramification on Jensen and NVIDIA's ability to continue to argue for more dovish policy with regard to China and AI. I will of course keep you posted as it evolves, but for now, that's going to do it for today's AI Daily Brief. Appreciate you listening or watching as always, and until next time, peace. Transcribed by ESO, translated by —
PodcastTranscript AI