11AM Hour: Senator Richard Blumenthal on AI, Early Facebook Investor Roger McNamee & Clarity Act Fails to Advance 9/16/26

Good Wednesday morning and welcome to Squawk on the street. I'm Sarah Eisen with Karl Kintinia, live from post-Night of the New York Stock Exchange. Today, the AI Safety Debate, Connecticut Senator Richard Gleumenthal warning that quote, "we're dealing with real peril, not science fiction, and the senator will join us to discuss his call on Congress to deal with AI legislation before taking any breaks." Plus, early Facebook and Google investor Roger McNameez with us, while he argues the AI game is about to end and thinks this recent fear factor is an attempt to scare the government into saving them. And Bitcoin trading near four week lows here after a major crypto regulation bill suffered a defeat in the Senate. We'll discuss the path ahead for the crypto sector. "Maintime pretty tight range as we await the Fed decision in a couple of hours. Dow is only down about 30 points, SAP hanging onto 76/14. See, the tenure has backed off of five by a few points, even with a very strong retail sales number, whether that was on headline, exotdos or exotdos of gas. Anytime the growing debate around AI safety, Mark Zuckerberg, the latest to chime in, siding with Nvidia's Jensen Wong, our Kate Rooney has some details. What a week we're having, Kate." "It really is, Carl, so we did hear from Metis CEO pushing back on this growing crowd of AI leaders now arguing for an industry coordinated slowdown amid some of the safety concerns that we've been talking about. Mark Zuckerberg chiming in, saying competition and the threat of liability already gives AI companies enough incentive to build the tech safely and that each lab has an individual responsibility to slow itself down, if necessary." Zuckerberg did point to meta-recently delaying its news AI agent for months he said over security concerns. Other labs should note have done the same, notably Anthropic, with Fable and its model. The comments, though, very much put Zuckerberg on the other side of a few of his industry peers here, Anthropic CEO Daria Amadeh did kick all of this off over the weekend. He called on the entire industry to quote, "Pace the frontier," and argued that some of the rapidly improving AI out there could create catastrophic risk without more time for testing. Open AI, CEO Sam Altman, quickly agreed over the weekend saying, "Open AI is now discussing its own need to pace the development internally and wants to work with others." Elon Musk on board for some of the industry coordination, he backed Amadeh's idea here, floated the idea, though, of companies sort of grading each other's homework instead of grading their own and there is now a queer counter here. With Zuckerberg and Nvidia CEO Jensen Wong, he was on mad money last night with Jim. "The fact that we need new laws, new antitrust laws, or new regulations so that these companies could do their fundamental engineering and do what properly before they release products, that is just completely unnecessary. And I think that that's what the president is saying, that we don't need new laws. We have plenty of laws, we have plenty of regulations." And guys, the biggest open question now is about global coordination. We've talked a lot about what the CEOs think, but it's very unclear if China would agree with any sort of slowdown a lot of CEOs suggesting that's not the case and obviously President Trump weighing in on that as well. "Kate, people say, well, we don't trust Exxon to regulate their own carbon emissions. Meta had to be taken to court to do some content moderation. When Jensen Wong says there are plenty of laws, what is he talking about?" "The most cynical take I've heard on that carl is the tobacco industry, suggesting that they self-regulate." I mean, some have talked about Newell and Musk brought this up, the movie industry, which has done its own form of self-regulation. Others have talked about the FDA, which of course is federal, FAA, as part of this. But the problem is that there's not consensus on what each company is agreeing to. So when Zuckerberg says, "Each company is responsible," that also implies that each company is going to agree and act on the behalf of all of us and really not take too much risk in sort of giving up any sort of profit here. I mean, the underlying factor here as well, as most of these AI companies, the big ones, at least anthropic and open AI, are also trying to grow revenue at a time when they're about to go public. So that's making a lot of people nervous. I will also say, you have these big tech companies, but there are smaller players who also don't want the big giants in Silicon Valley making the rules. We talked to co here, CEO yesterday, Aiden Gomez, who's based in Toronto and said, "Wait a minute, he called the tech companies here a cartel." And he said that he's very much with some of the smaller players saying, "We agree on safety, but we don't want the biggest players being the ones that come up with the rules here, but unclear if there will be rules as we talk about all of this." Yeah, that's the whole regulatory capture argument that has sparked some criticism around all this, as well. Kate, thank you. Kate Rooney, much more on AI coming up this hour when we speak with Senator Richard Blumenthal and early Google and Facebook investor Roger McNamee. And then tomorrow, we've got Palantir CEO Alex Carp here on the show to talk about all of this. That's at 11 a.m. Eastern. The other big story today is the Fed decision. And our next guest warns that Fed could be facing a slippery slope if it does decide to hike today. Joining us now is Rosenberg Research Founder and President David Rosenberg. So you're in the camp that says they shouldn't do it. I don't think that they should do it, but I think that they will do it, and I don't think there will be any dissents. The key is going to be really, you know, although Kevin Morse doesn't like to give guidance, he does give body language and tone. However, I think what's going to be more important really is what the dog plot show. If they're going to ratify the aggressive tightening that's priced in right now when it next year. But no, I've never really a fan of a central bank tightening into a supply shock because this isn't about an overheated economy. The man led inflation. You know, they talk about how inflation's been above target for five years, but I don't know what makes sense about, you know, tightening policy based on what's happened in the past five years. That's in the rear view mirror. So I guess they could argue that beyond just that inflation is above target that, you know, there's evidence it's seeping through the AI build out, for instance, pretty inflationary. And there's not much out there that shows that, you know, their policy is restrictive, I think outside of the housing market. I mean, if you look at, you know, some of the consumer data, the employment data, I think that they could feel confident that there's more risk on the inflation side than on the job side. Well, if you actually did the sector size, which I did, which you add up on the interest sensitive sectors, the sickly sensitive sectors of the economy, which is cyclical consumer services, durable goods spending, housing, non-residential construction, year-to-year, it's now running flat. So I would say that, you know, there's a lot of evidence that the economy beneath the veneer is responding to policy that actually is on the restrictive side. I would never make the claim that policy is not restrictive. I would say that I don't see anything out there attacking the Fed's credibility. The dollar is not cramping out. Market-based inflation expectations are really within the range they've been in the past five years. And we're going to be ending, you know, Kevin Worsh is talking about inflation data, but he told, that he already told us he doesn't even trust, which is why he's got this, these task forces, and one of them is on inflation to come up with a better barometer. So even when he talks about that the share of the PC deflator doing this against that, it's ironic that that's based on the same indicator that he doesn't even have faith in. So I think outside of oil, you mentioned, you know, the AI impact on memory chips, so on and so forth. But you know, I did a bottom-up count of the last CPI number that came out last week out of all the sub-components because Worsh likes to talk about breadth. So this is the August CPI number. Did you know that 45% of the components were flatter down and historically in any given month, that number is 40%. So there's a lot of things out there, Sarah, that are actually just inflating or deflating. That don't come out imagined. We've been presenting this argument. Yeah, I mean, we just got NAHB today, where David, I think what, 38% of builders cut price up from 35, average cut, 6% for six straight months. Yes, and 100%. Well, look at, look at median new home prices, negative 0.8% year of a year. You know, look, the, okay, so I guess Sarah believes in the 162,000 on from payroll number for August that we all know we're going to get revised. If you look at the 12-month trend in household or payrolls, it's roughly flat. I guess that's the no higher, no fire economy. Your less stability means no growth in employment. But, but are we getting inflation out of the labor market? I mean, the year for your trend and average early earnings is barely more than 3% instead of 5 year alone. So everybody's got inflation on the brain. This is if you have an industry that professionals out there and the people that you interview that have inflation on the brain because of oil, you know, but there's no inflation out of housing no inflation on the labor market. No, is there monetary inflation with 5% plus M2 growth with money velocity having rolled over since March? answer is no none of this I mean it's real oil is real but it's not it's just it doesn't have any sustainability to it I do think that the fed raises rates today and I think there's gonna be bar there's gonna be bars for more several months or now so I was just gonna say so no matter what you think if they go today what what ultimately happens how damaging do you think it'll be and how will that show up it'll show up in the sense that the it's on about the action today that's baked in the cake and today's retail sales number is gonna make anybody on the FOMC the voting class comfortable to raise rates I don't think they'll be any dissents I'm not saying that I agree with it they have their reasons but if they ratify what's priced into the futures curve that's a bit of a problem it means that this is not like one and done or two and through they're they're gonna be tightening it in the next year and the problem I have with the economy is this because there's lags is that you know 90% of this run up in treasury else has not even been inflation expectations it's been the real rate for a variety of reasons it's one thing to have the real rate go up as much as it has into an environment of accelerating real growth but that's not happening at the margin real growth is actually going down it's not contracting but we've had a real shock here layered on top of the layered on top of the oil shock is a real interest rate shock now so I don't know is the Fed gonna ratify what's priced in to me that's gonna be what's important because the markets have gone I mean a lot of the reason why treasuries are gone up there's been this massive reset of the Fed and the Fed's let it happen so really quick the question is what is what is what does he do today mr. no guidance the sort of say to the market you know you may have over left what our intentions are gonna be down the road yeah really quick before we let you go David I got to get you on Canada two headlines in the times one is world economy becoming wary of the US and the other is about bonder land offering this associate membership to the EU you think that flies where you are I think there's really broad support for what Mark currently is doing across the board and his popularity is very high but you know let's face facts what Canada's share of trade with Europe is like 5% with Asia it's like 10% so we're talking about you know a full diversification away from the US it's it's an extra impossible because that's where all the supply chains are connected but you know at the margin at the margin I think he's doing the the right thing to protect the economy here but you know Canada's trade times with the United States are always gonna be an extremely linked we'll be able to walk away from that but at the margin I think there's things he could do to question the blow and one of them is to diversify our trading relationships but is it gonna spend the dial in a major way no okay David Rosenberg thank you for joining us with the with the opinion the critique on this Fed day we appreciate it from Rosenberg research still ahead this morning Senator Richard Blumenthal's with us will talk about how he's right looking at regulating AI right now plus early Google and Facebook investor Roger McNamee joins us why he says AI companies are trying to scare the government into saving them we're back in a moment welcome back to squawk on the street calls for federal AI regulation intensifying following multiple warnings from insiders last hour we spoke with the defense department's chief technology officer who says that AI companies need to learn to regulate themselves listen let's hope they they do the right thing and become a normal company with normal normal companies with normal products with normal safety standards that they test and spend money on just like every good company does when they mature these companies are three four years old they need a little more time to season in my view to understand how business really works our next guest says however that one half forward could be to establish an agency like the FDA but for AI companies and joining us now first on CNBC from Capitol Hill is Connecticut Senator Richard Blumenthal what why would why would Congress or the government know more about how to control the safety issues here than the than the companies themselves senator it isn't a question of who knows more it's a question of who has an interest in protecting all of us ordinary everyday people in the world from the potential dangers of AI we are becoming more and more aware of what those dangers are but in fact we started having hearing three years ago Senator Holly Republican of Missouri and myself with Sam Altman partly a household word at the time and all of the other AI leaders who said then we need some oversight standards government scrutiny and analogously to the FDA which regulates pharmaceutical drugs and medical devices those are well established mature companies but we've decided that there ought to be reviewed for whether or not these products are safe and effective not a slow down not a pause not excessive government intrusion but simply expertise and experience through a group of scientists that decide whether or not these kinds of drugs or medical devices ought to be put on the market same idea here before these products new technology are deployed there ought to be some objective review and we should learn by the way from our experience with social media where the horse is out of the barn algorithms are driving toxic content of kids and we're still trying to catch up do you not worry though about potential harm to innovation and global competitiveness which are the primary warnings against over regulating this that's a great question and of course I do worry about it we have to remain competitive China is understanding that it needs to regulate AI because it could be a threat to their regime in fact even more so to their kind of autocratic rule than it would be to the United States and it has to be international in scope it has to avoid stifling innovation and invention or entrenching the kind of monopolistic control that we see in social media and now more and more in AI and let's face it AI is pervasive in our lives already we need to make sure that the benefits are accomplished without the dangers and that's the reason why the artificial intelligence risk evaluation act a bipartisan measure I think has a lot of promise I think a lot of people look to Europe and they say look what regulation does regulation stifles innovation there's a reason they don't have any big AI players and they're nowhere in this race and that's because of the regulatory state nobody wants a regulatory state what we want is in effect some scrutiny to prevent the autonomous road box out of the sandbox infiltrating sites without the control of the developers who are responsible for them and I'm just done with trust me that's what the industry says it is and set it for years and we've seen the results in social media it's the reason why the kids online safety act has attracted such strong bipartisan support Senator Blackburn and I proposed it a number of years ago it passed the Senate by a 91 to 3 vote it was stymied in the house because of big tech's opposition its armies of lawyers and lobbyists we need to make sure that democracy works that we get things done and part of it is to make sure we protect people against the dangers that are out there so if we're to go down this road on this FDA sort of like metaphor I mean certainly we we regulate drug discovery and no one would argue we're not number one in drug discovery but would this look like a new model in a phase three trial when it gets reviewed how do you see that pipeline working well part of what Congress needs to do is avoid heavy-handed or intrusive intervention and we need to give the agency whether it's the department of energy or some new agency the flexibility and freedom to make sure that the process targets potential evils and done it does it in the most expeditious way so the answer to your question is I think we need to delegate some of this responsibility to the experts we are hardly experts in Congress on AI and I think we need to understand our own limitations in this area where a highly complex and challenging technology is involved where we want to be at the tip of the spear we want to be at the forefront America should be just as we are in development of new medicines and medical devices and I think there is a balance here to be struck but the point is that right now Now, we're on the verge of losing control completely. >> Yeah, I mean, there's another risk with that, that, you know, critics warn, which is, you know, as you say, it's not Congress that are the experts, that there's this regulatory capture, that basically the agency gets captured by the industry and the dominant players when it comes to a fast-moving industry like AI, because they understand what's going on, and that ultimately will stifle competition and upstarts and just give the power to the hands of the few. >> And we ought to learn from the FDA that regulatory capture is indeed a danger, and that there are failings that we've recognized in the FDA process, sometimes styming the flow of new products to be available to people who want to use them. And the FDA has revised its process, learning from experience. So the FDA is far from perfect, but it does provide a model for looking at new products, whether they are safe and effective, safe and effective. >> Do you have an idea of what capabilities basically would become dangerous enough that the model shouldn't be able to be released? >> We've seen some of the dangers in so far as the bots or agents can be out of control, the idea of RIS, the self-improvement, recursive self-improvement, the misaligned agents, you know, all of these terms apply to, in effect, development that no longer is within the oversight of control of human beings. That's one area where I think people are very concerned. But you know, in a broader sense, we've been working on this issue for a number of years, three years. Senator Hawley and I came up with a framework for making sure there are export controls, assuring that there is safety, other kinds of review, and we've been proposing legislation that requires, for example, the Guard Act, that the bots disclose that they are not human when they're talking to people, that there be age verification, they're a set of protective measures that I think our ideas whose time has come. I'm glad to see the rising awareness here after we have been doing it for three years, and I'm hoping that we can get some measure across the finish line. But I think very importantly, there's a bipartisan awareness and support. Well, there certainly is awareness now, and we thank you for coming on to talk about some of the ideas here. I appreciate it. Keep us posted. Senator Richard Blumenthal. Still, come this morning, early Google and Facebook investor Roger McEnmail joined us, why he says what's going right now, and on right now, in AI, is ridiculous. Plus, the Clarity Act failing to advance in the Senate yesterday, we'll discuss what that means for the future of the crypto industry, back in a moment. Welcome back, Fed Day. S&P slightly higher this morning as we await the decision this afternoon. We do have some names in tech like coherent Dell, once again, Marvel, all-top gainers. Maybe Hans the biggest lagger today, stock is tumbling after the trucking company did warn of an earnings drop between 5% and 10%, kind of notable since they're not classic guidance givers to begin with, but Diesel certainly hurting transports today. Hearing from some airlines, too, also at Morgan State. That's a story, right? High diesel, high fuel prices, that's really going to hurt truckers, and also Norfolk's Southern talking about gaining market share from the truckers because of it. Now for a news update, Julia Borsten has it first this morning, Julia. Sarah the House today is expected to vote on a new sanctions bill aimed at increasing economic pressure on Russia. The bill named after the late Senator Lindsey Graham passed the Senate last month. The legislation would target Russian defense and energy industries in the shadow fleet of tankers debating current sanctions. President Trump is expected to sign the bill into law if it passes. Amazon announced today it's raising the minimum hourly pay for eligible operations workers by $1 to $20 an hour, including the value of a new benefits package. The company says total compensation for those full-time employees will now exceed $32 per hour. And travelers are apparently happier with their experience at the airport, even as record number of people are flying. JD Powers annual survey finds overall satisfaction with airports increased significantly from last year, in part because of newer gates and terminals. The top airports on the list, Minneapolis, St. Paul, Detroit, Phoenix, New York, LaGuardia, and Harry Reid International in Las Vegas. Back over to you. LaGuardia. What about? It's been such a big upgrade lately. They rebuilt it. And the food choices are still good. I think that's very important. Thank you, Julia. Julia Borsten. Still ahead. Early Google and Facebook investor Roger McNamee joins us. Why he says AI companies are trying to scare the government into saving them. SK Heinext reportedly intakes with Intel to manufacture memory chips in the U.S. for the first time. Gets into tails on that story when we come right back. Longer back, a squawk in the street. Our next guest out with a bold call saying that the same AI leaders who are now calling for a slowdown ultimately picked quote a dead end technology that cannot do what they wanted to do. Let's bring in elevation partners co-founder Roger McNamee, of course, an early investor in Facebook and Google, as well as a former advisor to Mark Zuckerberg, Roger. It's great to have you. It does sound like you think these guys are up against the wall on open source competition and they are pulling any lever they can pull. Yeah. So, the issue Carl, I think is this, you know, anthropic wants to go public next month. And if I were an investor today, I would be really thoughtful about the risk factors that would exist here. And the first of those is the one you just mentioned. Which is that LLMs appear to be a dead end technology that for all the intensive purposes we've put, you know, trillion dollars plus into this whole thing to produce a technology that has error rates 10 to 25 percent. Which means you can't really apply to high value stuff. The second problem they have is they went with a general purpose product architecture. And you guys remember the last thing that worked with general purpose architecture was SAP with its ERP product in 1993. And I'm not sure anybody ever got a positive ROI from that. Because it's too expensive for corporations to implement general purpose architectures. They like you to solve a problem with a quick payback. These things do not do that. The third problem is they spent almost a trillion and a half dollars on cap equipment without commoditizing the hardware first. I mean, that is absolutely nuts. And you know, I think that to the internet, by the time Google came along, six years into the internet, they had taken three orders of magnitude of cost out of the infrastructure. And we still have rising unit costs here. And then the fourth thing is they're 10 guys competing for what? Two slots. So a lot of guys are going to get hurt no matter what. The final point here is that there's not one customer out there that's paying more for their AI than it cost the AI vendor to deliver it. And that's not a sustainable model. So what should the response be to their complaints? Well, I think the core issue here is investors. We should be sitting there and going, "You must be kidding. Cannot take a company public when, you know, they're sitting there saying we have 80 percent growth margins as long as you exclude all of our operating costs." And I'm sitting there going, "I'm sorry, that's not how markets work. That's not how accounting works." And, you know, there is potential to all of this stuff. But the problem that they face is that they made a bet. The bet hasn't worked. The open source guys are eating their lunch. They've just had a disaster's PR cycle with data centers and a disaster's PR cycle with hugging phase. And now they're sitting there going, "We are screwed. The markets don't want to give us more money. We need the government to bail us out." We're still getting news though of pre-IPO funding rounds, at least on the open AI side. That's just today, right? No, no. Hang on. This is the great thing about markets. You know, there are lots of people who disagree with me. I'm just saying before you put your money to work, you should do the analysis. Look at the five points I just gave you there and ask yourself, objectively, in any other context, what would you pay for companies with those facing those issues? And, you know, again, there are lots of people who paid big money for we work right up until it blew up. There are people who paid real money for N-Rod before it blew up. And you know, FTX, same thing. You know, I just, markets have manias. This is a mania. And, you know, the key question is not, will this be a big deal someday? The question is, can you make money from the valuation you're paying today? And my argument is that's going to take a miracle. And I think what these guys are signaling is they want the government to force a slowdown and they want to force it a way that blocks open source AIs from coming to market and where they get bailed out by the government for the delay. To be clear, you mean your comparisons to N-Rod and FTX? I mean, those are illogality, as you were not talking about that stuff. Well, actually, Carl, I would tell you, if anthropics way of describing its revenues is the way they're going to bring this thing public, you're going to have to explain to me how this is different than N-Rod. Why? Because you think it's-- Well, they're excluding that cost-- >> Yeah, they want to exclude the cost of training, they want to exclude all these really basic costs. They're trying to treat stuff that they received as an investment from Microsoft as though it's revenue. I mean, that's not how revenue works. Again, we haven't seen the hard data, right? But we've heard what they said about the hard data. >> Here's the thing. I don't think that most people would agree with you, with your whole premise that LLMs are a dead end. I mean, they have shown measurable, scalable progress. They're being enabled into all sorts of new functionalities and capabilities, and they remain central to AI research and how we're interacting with it. >> I'm not suggesting there isn't interesting stuff going on. I'm really saying they've spent a trillion and a half dollars. They have three trillion in commitments on their valid sheets right now. We're talking about the big tech guys. And so far, the industry's aggregate revenues are under $200 billion cumulatively. I mean, the math just doesn't work. >> So you're suggesting they're just overvalued? >> Grossly overvalued, but the point is, we're talking about markets. We're talking about an anthropic IPO. So the valuation is actually a really important thing to talk about here. >> Roger, does this mean that you don't believe them when they say that there's 10% risk of extinction or even whatever incremental risk to humanity there is? >> No, I think that's all a game that they play. These guys come from the whole world of effective altruism which believes in kind of the replacement of humans by machines so they have some philosophical things that go to that extinction risk argument. But no, I think it's total nonsense. You look at this, I mean, the products that went into hugging the face were programmed to hack into other companies. That's actually a crime, right? That's a felony. They were programmed to do that. They were put into a so-called sandbox that was not architected properly, but all the controls were stripped off by OpenAI. And I'm sitting there and saying, that is not an indication of extinction risk. That's an indication of incompetent software development and really bad security practices. And I look at this, I go, this, that stuff didn't, it didn't go rogue, it didn't think, it didn't conspire. It was programmed that it did exactly what it was programmed to do. It's just it was programmed by people who either failed to understand how to manage a sandbox or they did something that was blatantly illegal. >> So do you think there's vulnerability for the vendors who are expecting counting on revenue from these guys? >> Of course, don't you? I mean, do you expect, I mean, I just don't expect $10 trillion, which is I think the most recent forecast we've seen out of NVIDIA and Tropic about kind of the long-term investment dollars. Right, those kind of numbers require this thing to make the world a better place. That's not what we're seeing. And, you know, what progress you're seeing here is against benchmarks that aren't real world. They're things created to evaluate a product that you're trying to sell. And so my only point here is that no matter how confident you are in the underlying technology for the long term, putting a trillion dollar valuation on one of these companies is not gonna lead to happy endings. >> Do you think that regulation is some sort of solve or are fixed here when it comes to some of these risks? >> Well, I think the first thing you do is you apply just basic liability laws. I think if these guys hacked, and there's some evidence that they constructively created a situation where companies like Hugging Face were hacked, that can be prosecuted, you know, hacking is a felony. There's also product liability things that can be applied here, we don't need new laws. These guys are breaking existing laws. We need to just recognize that there's nothing abnormal you know, about this technology. It's just the latest hype machine. And we can manage it in the existing framework. We just have to have a little bit of government- >> You say they're breaking existing laws. What do you mean they're breaking existing laws? >> I'm saying there's a law against, against hacking. You can go to prison for hacking another company. And they set up experiments that were designed to do precisely that. How is that not a violation of the law? >> Yeah, well, the report out of Reuters today that some of these agents were probing the site for vulnerabilities for nearly two months before being detected. That's a new wrinkle out of Reuters. Roger, next time maybe we can talk about how, how you're positioning around all of this interesting thesis though. We always love hearing from you. >> I'm sure I'll get a lot of blowbacks for it, but I will tell everybody out there. >> You know, I don't need you to agree with me. I'm just trying to help you think about this in a way that saves your net worth just in case things go wrong. >> Roger Mackamy, thanks. By the way, do not miss CNBC's second annual AI forum. In Dallas on October 1, we're gonna try to tackle some of these issues here from top business leaders, entrepreneurs. For that, you can scan the QR code or just go to CNBCevents.com/AI to learn more. >> Still ahead, the Senate failing to pass the clarity act yesterday. We're gonna break down what that means for the future of crypto when we come right back. >> At this point, I don't think we can wait on Congress and the Senate. It's in our interest at this point to go work directly with the SEC and the CFTC. And chairs, atkins and C-LIG at the SEC and CFTC have made it clear that they have, under their existing authority, all the tools they need to go publish clear rules for crypto. They've even said publicly I think their intention to go out there and start the rule-making process right away. >> That was Coinbase CEO Brian Armstrong speaking to CNBC's Garret Downs yesterday after the clarity act failed to advance. Garret joins us with more on what is next year for the industry, Garret. >> Yes, Sarah, well, you heard Brian Armstrong there and basically he's saying they're gonna be looking to the White House now to regulate crypto. You know, this long awaited bill to finally regulate crypto through legislation has now seemingly failed. So they're looking to the White House, they're looking to the SEC and the CFTC and the Trump administration, which has been an ally to crypto to regulate them now. >> So what would that look like? Can the White House actually put in place any kind of regulation or legislation on this front? >> Well, they can't put in place legislation or anything like that, but they can do rule-makings, they can do executive rule-makings, and those are less durable than legislation. So another administration could come in and undo some of those, but it could, in the meantime, allow crypto to get at least some rules of the road if the White House does come in and act. >> Why did it fail? >> You know, Sarah, they just couldn't get to a point where Democrats felt comfortable supporting this because of this ethics part. Democrats had a big problem with the money that President Donald Trump and his family have made on crypto and they wanted clear cut restrictions that the President wouldn't be able to trade in crypto wouldn't be able to continue profiting on crypto into the law of the land. Now, there was a lot of back and forth going on about this, last minute exchanging of papers, but they just couldn't get to a point where Democrats felt comfortable supporting it. You also had them lose a couple of Republicans too. >> How about the timing on all this, Garrett, coming within what, 46 days of the midterms? What does that mean? >> Yeah, so look, everyone's now sort of looking at what the crypto packs do. These groups like FairShake that spent a lot of money in the prior election in 2024 and actually had some impact in those races. So everyone's sort of looking, Armstrong said he would be surprised if these crypto packs were not looking at yesterday's vote as an indicator of where they should spend. So we're all gonna be keeping an eye on that. FairShake yesterday told me they have no news to share yet, but I'm sure they will be looking at some races where they can spend in and hopefully get a more crypto friendly Congress. >> Yeah, certainly some of the comments, at least on the wires out of Thune today suggests maybe this could be reconstructed or revisited. Really interesting, Garrett. Certainly Bitcoin has suffered a bit because of it. Appreciate that, Garrett Downs. Still ahead, SK Heinex and Intel jumping today on reports that they're in talks to maybe manufacture some memory chips in the US for the first time. We'll get you details after a short break. Got some action in the chips today. SK Heinex reportedly in talks with Intel to manufacture memory chips in the US for the first time Christina Parts Nevelos watching that story this morning. Hey, KP. >> Hi, Carl, all the fab at the center of this is one Intel started building in Ohio years ago. Still hasn't finished. Two options though on the table. SK Heinex could simply lease part of that particular plant or maybe the more ambitious path. They joined venture with Intel and Cloud Giants building out AI data centers that are essentially star for memory right now. And by helping fund the plant, they would lock in their own supply. SK Heinex isn't denying any of this. The company says only that it's exploring options with nothing final. They even told me this when they had their US listing here at the Nasdaq. So I actually believe all of these headlines. Politics and demand. Korea is on the hook for roughly a $350 billion investment pledge in the United States, the price of keeping its tariffs down. And Washington really keeps pushing to see more of it built on American soil. SK Heinex risks being singled out if it stays on the sidelines and doesn't actually commit to building on US soil. And of course, man, we know this car was white hot. Intel's own CEO just yesterday at a conference said that memory prices have jumped five to seven times and suppliers don't see relief until 2028. That's the pressure pushing cloud companies to back a plant of their own. For Intel, this helps maybe fix an expensive mistake. The company spent billions starting, maybe I shouldn't say a mistake, just an expensive endeavor. The company spent billions starting the Ohio plant then installed construction because they didn't have enough customers to fill it. They changed CEOs. SK Heinex could be that customer paying to use space until already plant to build. Investors definitely like the idea. That's why shares are up 5%. And there's already a relationship between the two firms Intel sold its own memory business to SK Heinex back in 2020, and now it may actually rent the floor from them. The hard part is high bandwidth memory. The high-end memory behind the AI trade. South Korea guards it as a national core technology and making it abroad means a government review. And with that fab not set to open until 2030, 2031, a signed deal could still be years from its first chip. But nonetheless, the market is reacting positively. And speaking of chips in closing bell over time, I'll have an exclusive interview with on semi-CEO after their investment day, hard-pid it, but in the world of chips, guys. - I was gonna ask you about the geopolitical issues here and what led to a potential deal. Intel now considered more of a national champion, SK Heinex for Korean. Just I don't know how it relates to some of the geopolitical and trade conversations. - Well, the thing that I mentioned in this story is that the sole economic daily times is reporting today that they're putting on hold their $350 billion investment in the United States. They were gonna announce it tomorrow on the 18th, or 17th I should say, and then announce it here in the United States. So things are already getting delayed on their end. That's why politics plays a big part in it because the US is threatening to add more tariffs on companies like SK Heinex and Samsung. And so they need to say that they're gonna be investing here on American soil in a big way, and specifically building chips. And many of these firms haven't done so to that extent. So perhaps it'll be used as leverage, but this is South Korea, not China, which we know is a whole other conversation starting on a September 10th course. - Yeah, it's been a huge source of irritation to commerce. The fact that the Koreans have been, I don't know, a little bit of rope of dope, a little bit of dragging feet on some initial promises to deal with as well. - Christina, thank you, Christina Parks and Evalists. Stock market is holding in there with the S&P 500 higher ahead of the Fed. And I think a few things to watch today. First of all, will they hike? The market, the bond market is baking it in with more than 90% odds. What do the dots look like? Do they matter because Kevin Worsh himself, the chairman hasn't participated in the dots, but how hawkish do they get if there is a hike or there are more hikes baked in for the rest of the year? And then obviously his news conference, even though he says he doesn't want to do forward guidance, the market moved a lot after his Jackson Hole speech in a more hawkish direction. - Yeah, we were joking about the word hike in the first few graphs as, who knows whether he wanted the algos to read that, but we'll know more in a couple of hours. - All opinions expressed by Squawk on the Street participants are solely their opinions and do not reflect the opinions of CNBC or its parent company or affiliates, and may have been previously disseminated by them on television, radio, internet, or another medium. You should not treat any opinion expressed on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of his opinion. Such opinions are based upon information Squawk on the Street participants consider reliable, but neither CNBC nor its affiliates and/or subsidiaries warrant its completeness or accuracy, and it should not be relied upon as such. To view the full Squawk on the Street disclaimer, please visit CNBC.com/Squawk on the Street disclaimer.