Closing Bell - Closing Bell Overtime: Nvidia hits all-time high 10/2/26

Episode Date: October 2, 2026

Charles Schwab’s Kevin Gordon breaks down whether the recent yield roadblock for stocks is beginning to clear. Former Council of Economic Advisers Chair Jason Furman weighs the latest jobs data and what it means for the Fed’s next move. Needham’s Laura Martin questions whether enthusiasm around Meta’s Muse has gone too far as investors weigh its early traction against rising capex, competition and monetization risk. CrowdStrike President Michael Sentonas looks back at a wild week for AI and what the latest developments mean for cybersecurity. Meantime, our Annika Kim Constantino examines the next big obesity drug opportunity beyond GLP-1s and the new hormone attracting attention across biotech. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The bell's bringing an end to the trading day at the NYSC, the Manufacturers Alliance, ringing the click bell, and closing out the week here at the NASAC QT imaging. Welcome to closing bell over time. We're live from studio be at the NASAC market site. I'm Melissa Lee, along with Mike Santoli. The NASAC can be an intraday all-time high today after the softer jobs report. The Dow up at 250 points. S&P 500 up 3 quarters of a percent.

The NASAC up more than 1%. InVIDIA, a leader of the tech rally, it also hit an all-time high during the day. its first record since mid-May, the market cap of the company solidly above $5.5 trillion now. To get to $6 trillion would take about another 7% rise from here. Those gains in tech coming, as many now seem to think a Fed hike in October is less likely. But despite that, bond yields moving up off their morning lows. Today's gains boosting the NASAC composite and the NASAC 100 to about half percent gains on the week. The Dow, S&P, and the Russell with small losses.

Let's get the first take on the close right here. I thought the most interesting thing was that rebound in rates. I mean, it did not last, and we've seen that time and time again. We saw that with PCE. The impact on rates is fleeting. Yes. And then the rates resume the trend. It is very interesting. Now, yesterday there was enough of a downside reversal in yields that people thought that could have been a consequential peak. And maybe we are testing that proposition a little bit. We did not get to yesterday's high. But it's a pretty fragile setup in terms of, I think, positioning in yields. And people were not willing to extrapolate, you know, a moderate shortfall in jobs to say that we now therefore know that the Fed is either on hold

or, in fact, that just the other factors driving bond yields to where they are have abated suddenly. So stock market, you know, obviously we got a lift. Beret was better than it has been, but not overwhelmingly to the upside. So I feel like the stock market was unable to fully gain pace after you saw the bond market, you know, show a little bit of trepidation. But obviously the standout, the NASAC, NASAC, 100 new intraday all-time highs. Philadelphia, a semiconductor index really leaving the charge there, and you can make the

argument, well, this is the most cyclical part of the stock market. Yeah. And here we are, with nice, strong performances, this doesn't feel very defensive in light of 5.3% on the 10-year-year-year-year-year-year-e. Within tech, it's not defensive. I think you could say, look, since the bull market started almost four years ago, semis have been the strongest part of the strongest part of the stock market. So, Invidia, kind of catching a little bit of a bit, does make sense.

I think it's the better way for tech to dominate, which is it goes up more than other things, but the rest of the market at least participates a little bit. That's kind of what we had today. Even though banks did nothing special, actually down significantly on the week, by the way, Micron today gave up most of yesterday's gain. It's down a couple percent on the week after having a stellar quarter. So I still think there's enough for people to kind of squint at and wonder what's going on there. All right. Let's get more on the bond market reaction to this morning's job number.

Rick Santelli in Chicago with more on a pretty interesting. day, Rick. Yeah, pretty interesting indeed. And, you know, I'm not sure that I think the report is all that week. Let's look at a couple of issues real quickly. If you look at a year-over-year, average hourly earnings, you know, we started out the year at 3.7% for both January and February. My point here is that the weakness there isn't unique to this month or last month.

We've been losing wage prosperity both on the month-over-month and year-over-year all year. This is definitely a trend to keep in mind. And as many have pointed out, when the Fed is fighting inflation, dropping wage growth, and raising interest rates, really it seems a little bit counterintuitive because there really doesn't seem that pressure building that they can address by higher rates. The other issue is quite simple, that the unemployment rate went up to 4.2. However, 4.1 is the low of the year. The high of the year is 4.4. And most of the increase there was because of the good news of labor force participation going up.

Real quickly, if you see the intradate charts, 12-hour charts, when was the low made it 5.15 in a 10 year? You guessed it, like one minute after the number came out. And then all the other variables that have pushed us higher, took control of the market. And right now, as it stands 520, what, 8 in a 10 year, that means we're down 8 basis points on the week. And considering that last week we closed it, you know, a yield of 516, excuse me, 524, that really is. not a big drop considering everything. And finally, when you do look at the one-week chart, today that 515 low, that was a new low of the week before it reversed. And I do draw your attention to how much the Fed not real guidance has come out. We have seen Mr. Williams, Mr. Jefferson

from the Fed, push those percentages from what, over 70 percent early in the week to right around that 20 percent, which goes to show that Mr. Worse does have a vehicle to get the word out when the markets in the Fed aren't on the same page. Back to you. For sure, Rick, yeah. I mean, I think that's a fair assessment that maybe the market was getting ahead of itself and building its own narrative here. And to your point about, you know, maybe it wasn't so weak, the unrounded unemployment rate rose to 4.175 from 4.141. So it really was just a rounding effect that even gave you that cosmetic increase. You know, as you finish things up here, Rick, it reminds me that you go back to when you started, 10-year yields are in the same zone here. We keep talking about

how this has been a normalization back to a long-term, kind of pretty common level. I wonder about your thoughts about what's driven us to this point, how it relates to back then when we had, at least on paper, we had a surplus in the federal government, and where it leaves us as you walk away. You know, I've always been disturbed that debt and deficits haven't mattered enough. And I am sure that that isn't the only reason rates are up here. But when you look at it from a global plate, yes, all the deficits of all the big economies around the globe are experiencing the same type of increase in rates.

So that's a part of it. What's going on in Mideast with energy, that's a part of it. What's going on between the refineries and the notion that we don't have enough finished products? That's adding into it. And I do believe when the war ends, you're going to see a downtraft in yields. But I unfortunately think that downjointed yields.

is a down draft you're supposed to sell looking for yields to move higher. I continue to say it's also a huge part of normalization. Managed interest rates globally is something you can't get away with forever. I've said it a million times no matter if you're talking about entitlements or interest rates, there's no free launches. And when you look at the 46-year average of a 10-year no yield, it's about 5.6%. So normalization will take us higher, and I think that's going to be a driving force for the rest of the year and into next year. Rick, I can't believe that this is it in case viewers

missed it. Rick Santelli is retiring and Rick will miss you dearly. Really, we will. We'll miss you guys and I'll miss the markets as well. Oh my God, I'm going to miss it all. You're welcome. Markets will still be here, Rick. And I always say as I've got it. I know, but it won't be the same. I know. It won't be. I won't have any word of vent. I won't have any word of vent. Take up the phone. Yeah. Yeah, that's true. That's true. True. We'll be here to the comment section online. I mean, look, as long as I've done this, the more I do it, the more I choose to listen to the markets as opposed to suggest what they ought to be doing, and you've been doing that for a very long time. So a great lead and all the best as you head out. Thank you. It's been a pleasure to work with both of you. Thank you.

Thank you, Rick. Rick Santelli. Oil prices pulling back today. WTI just above 90 bucks a barrel, as G7 nations agree to release diesel and cruise. stock pit. As Stevens joins us here with more on this. So this is 100 million barrels and it's going to begin immediately span for four months and they did say it'd be front-loaded and release substantial portions of diesel. But the announcement was light on some key details, including the breakdown of that 100 million. How much is oil and how much is diesel? And of course for diesel, all that matters is how much is diesel, since oil is not what's leading to the surge in product prices. So France previously had in a had reportedly proposed 50 million barrels of diesel and

50 million barrels of oil. So let's just say that that is what it is, and that it's over four months for argument's sake. It's spread evenly over that time frame. That's about 400,000 barrels per day of diesel. That is only half of what we're currently missing from Russia. So it's not all that meaningful in the grand scheme of things. Now, this would probably pressure U.S. prices, U.S. futures, given that that does reduce the cargoes that will be going abroad, and so that could keep more supplies at home. Probably won't see that reflected at the pump immediately, more so in a span of a few weeks. But again, this is more of, you know, a Band-Aid measure. It is definitely meaningful here, but it's not a long-term solution to all

the oil and diesel specifically that we're still missing. If you divide those barrels over the course of four months, because we don't know, there's a lot we don't know yet in terms of how the IEA will actually, you know, supervise a release. But what is the impact in terms of how much Europe is in deficit on a consumption? Because they consume much more than they make. So how much does that cover? Yeah, so they are incredibly import-dependent. So the idea is that this is not going to cover that because they typically import about 1.5 million barrels per day. So it's going to dent that. I mean, the U.S. is now exporting on average 1.5 million barrels per day, and that's how much they're importing. And, of course, all of these flows go abroad. But it's not

going to make up for all of the barrels that are missing because it's 800,000 barrels per day from Russian diesel and then also about 1.5 from the Middle East. And again, when we say that flows have recovered through the Middle East, you know, J.P. Morgan says only 58% of product flows have recovered. So that is where the tightness remains. And so this is, something they can do and clearly a response to ongoing pressure from the U.S. There's a lot of people now saying that the diesel export ban was floated as, you know, a way to get Europe to tap into their reserves. Well, it does all explain the interday action in the diesel prices, right?

Well up off the lows of the morning. Thank you, Pippa. Today's records for the NASDAQ coming despite yields continuing to rise. So are yields losing their grip on stocks? And could that clear the way for more new highs? Kevin Gordon, head of macro research and strategy for the Schwab Center for Financial Research, is here. Kevin, good to see you. That's nice to see you. We're all becoming experts on the correlations between stocks and bonds

and what seems to be driving both. I guess there's no ironclad perpetual rule on this, but where does it stand right now as we did see this, you know, a little bit of a head fake action in treasuries? Yeah, I think, you know, some of it, at least from the market, the equity market's reaction to even yields moving up today, or maybe lack throughout, because now we have a positively correlated move for both today. Part of it, I think, is just getting out of this macro exhaustion phase, you know, over the past couple of months where bond market volatility was consistent with levels that you had seen only in these sort of crisis periods, even actually worse than liberation days. You kind of have to go back

to the pandemic. You go back to periods, you know, during the financial crisis to see those kinds of moves in bond market volatility. So I think the fact that that has started to recede a bit. Plus, to me, I think that more of the move at the long end this week was actually driven more by some of the firmer spending data that we got and how that was confirmed, not just on the business side, but I think particularly on the consumer side, that personal consumption expenditure number for August was pretty firm, especially in inflation-adjusted terms. So to me, I think that's actually more consistent with the bond market, you know, selling off for the so-called right reasons of growth looking relatively firm. So you think we're in a normalization phase? Are we in this phase where

we're finding that footing at a higher level that we will remain at? Yeah, I mean, I think just looking at Admonal GDP alone. I mean, you've got it running closer now to 7% year over year. Whether we stay at that level in perpetuity, I think a lot of that will still hinge on how much of a pinch the consumer feels as energy prices stay as high as they are. But when you think about that and structurally higher deficit spending and also relatively sticky inflation, all together in equation, I think still gets you to a bond market that is closer to pricing that in sort of as a new normal or maybe back to an old normal a little bit. Keep trying to sort out which is the direction of surprise from here in the sense that we really did get in this mode of thinking, you know, economic overheat was a little more of a risk. You know, we start third quarter, Atlanta Fed, GDP now, 5.7 percent. Of course, that's coming down toward the low threes probably. And, you know, consumers have been able to maintain this spending without necessarily, you know, having it all come from wage growth. Yeah. And I think what's fascinating, I mean, if you just kind of break down labor into very simplistic kind of nominal terms. I keep it relatively simple in thinking if people have jobs, they'll spend money. Sure. I mean, they'll figure out how to move that around if they need to because of a spike in gasoline prices.

But basically with us living in this sort of fully employed labor market, it reinforces that strength month to month. I don't think the average consumer is inflation adjusting their paycheck every single time and, you know, thinking about how they have to adjust spending that way. Not that inflation doesn't matter, but that labor income growth and the fact that you have that income proxy still, you know, moving up as it is. I think speaks to some of the details actually in the jobs report today that pointed to a little bit of this stealth tightening. You know, the number of job losers continuing to roll over. It's at its

lowest since the summer of 2024. The fact, you know, Rick was talking about the increase in the unemployment rate being driven by labor force participation. All of that is still indicative of a labor market that continues to recover. It's not perfect. It's not even, but it's still recovering, nonetheless. In terms of the toll, though, of higher energy prices on the consumer and on businesses, for that matter, the longer this goes on, isn't that toll greater? I mean, it's not just a period of a couple months. I mean, once we get into six, seven, eight, you know, and so on, it's more detrimental. I think, yeah, and I think the key now, especially when we're going to start getting the

September inflation data rolling in, I mean, that was the month when you saw, you know, a 13% increase in diesel prices. So the fact that you're now going to probably get a, you know, a little bit of a stronger flow through from and pass through from companies in passing on those higher costs, I do think that has to be a factor moving into the. end of the year. But I also think that combined with the wealth effect, you've got still that nominal job environment looking pretty good from an income standpoint. I don't think that you really derail the growth train until you start to chip away at the labor market.

I mean, index investors have never benefited more from the fact that the stock market is not the real economy, right? I mean, just in terms of 40% is 10 stocks. The AI theme is swallowing the entire index. I think the past couple of months proved that. I mean, the average maximum drawdown in the S&P over the past two months has been 14. percent, markets barely off of its high. So you're getting that diversification offset maybe in a narrower sense because of tech, but from an index and a cap-weighted index standpoint, you're benefiting from that. Kevin, great to see you. Good to see you guys. Thank you. Kevin Gordon of Schwab. And take a live look at Chicago. Youngstown State University's student investment fund ringing the

closing bell there in Chicago at 4 o'clock, 415. There you go, on the dot, ending the regular trading day for options. Well, the September jobs report. was a surprising one after it came in much softer than expected. Non-farm payrolls came in a 29,000 versus the 84,000 that was expected. August payroll numbers were revised lower. July saw its gain of 21,000 revised to a drop of 10,000 payrolls. But there may have been another number in the report that makes an even bigger argument for the Fed not to move in September. Joining us now, Jason Furman. He is a former chair of the Council of Economic Advisors, currently a professor at Harvard Kennedy School. Jason, great to see you again.

Great to see you. You're going to point to wage growth. So tell me how this is the number. Yeah, you just had it up by the screen. Over the last year, wages have grown 3%. Basically, every month, the 12-month growth have been lower than it was in the past. In a 3% wage economy, that's the type of economy where you have maybe 2% inflation, possibly even lower than 2%. And so for me, this is just more confirmation that all the inflation we're seeing.

is I did not use the word five years ago. I never said transitory. This time it really is. It's tariffs. It's Iran. It's not underlying overheating in the labor market. We just don't have that sort of inflation anymore. But annual wage growth has not kept up with inflation for the past six months, Jason. So if you're an average person, you're thinking, I may be getting paid a little bit more. My wages may be holding steady at least, but inflation continues to march higher. Oh, look, if you're a worker in this country, it's been a bad year for you. You know, moreover, you're not focused on core inflation like economists are. You actually probably buy gasoline, certainly eat food, and all of that is more expensive than it used to be.

You are probably behind where you used to be. But here's the good news. If it doesn't look like inflation's being driven up, we don't need to compound that with another rate hike in October. Maybe we'll need one later in the year. I don't know. They're going to have to look at the data, but there's almost nothing I could imagine in the data over the month in October that would change my mind that we don't need a rate hike now. We can wait and see because it looks like inflation's headed back down. Yeah, that message certainly has gotten into the market. I think it's only like a 22 percent market implied odds of a hike in October. We still think maybe one comes in December, but it is open-ended based on the data. I guess the question, Jason, is nobody's really. really suggesting who thinks that the Fed should raise rates that we need to curtail consumer demand, that wage growth is getting out of hand. It's sort of this squishier idea of, well,

we have massive demand for capital, this sort of Cappex boom, is raising prices of some real-world things, and maybe even tightening up certain labor pockets of the labor market. And the Fed just got to get in the general zone of what the new neutral rate is, if in fact we're going to kind of keep up with what this boom might mean? Look, I mean, we have a boom, and that boom is raising long rates. The Fed doesn't really need to do anything in order to make that happen. It's an equilibrium in the market. So for the Fed, for me, the biggest question is not as the labor market overheating. The answer there is clearly no. It's whether their credibility requires this. I think it is possible it might. Their credibility probably did need the hike that they did last month.

But inflation expectations are in pretty good shape right now. Wage growth itself is actually a measure of expectations. It's how the price and wage-setting process is working in the economy. So they bought themselves enough credibility and enough room that they don't need a credibility hike anytime soon. They don't need a labor market hike anytime soon. And the AI hike is happening already. It's just happening in the long rates where it should be. When you say the Fed can afford to wait, Jason, what kind of time frame are you thinking?

how many more data points do you need to see or do you think the Fed needs to see in order to say, you know what, maybe a pause, a longer pause, is a right way to go? I like a little bit less forward guidance from the Fed. I don't totally mind that the chair is not giving us his dot. I do think we need more reaction function and to understand that. And he did a better job with that in Jackson Hole and I think at the last presser. So for me, though, they shouldn't be deciding what they're doing in December, what they're doing next year. The data should help them decide that.

For October, we have the main data point I care about. We'll have the CPI. Maybe that could change things. But I think we mostly know what we need to for October. But by December, we'll have a lot more data. And they should have an open mind. And they should have an open mind at every meeting next year. But for me, it would really be wage growth, increasing or inflation broadening in a way that looked like the,

transitory factors that we know about aren't really the issue. How focused are you, Jason, on the longer-term, slower-moving AI impact, whether it is on productivity or on demand for labor? Look, if you look at the data this week, GDP growth was revised up, the inertial parts of GDP were even stronger, combine that with the downgrades to job growth, and they cover slightly different periods. I'm not saying this all lines up. What you get is a story where growth is being handed off. There'd been a lot of labor growth during the Biden administration, during the Trump administration. It's been a lot of productivity growth and much less labor growth.

And by the, I'm not saying this, two presidents caused either of those. But immigration shift means, yeah, growth isn't coming so much from workers. Break-even job growth is quite low. But GDP growth has held up and it's held up because we are seeing the beginnings of the productivity benefits of AI. Jason, great to see you. Thanks. Great to see you. Jason Furman. C-Gate and Western Digital have soared on demand for storage, but those stocks sinking today on a potential new threat to their pricing power,

plus Tesla rising today after reporting delivery numbers, but is that really the reason for the gain? We'll discuss those stories next on closing bell overtime. shares of Tesla rallying today reported deliveries of 486,000 in the third quarter, slight decline from last year, but better than the street was expecting. We should point out, though, that whether the... The stock is going up or down. It often trades in things in line with other things, obviously, not just deliveries here.

There's a big focus on Robotaxy Rollout, on robotics, et cetera, broader themes of technology. And there's also another current in today's market in terms of what could be driving a Tesla kind of stock. I think there was a little bit of, you know, kind of noticing that we got some help on jobs and were past the Fed and there was a risk appetite. I mean, SpaceX was up 7%. That's obviously in the muscle. I obviously think it was an upside surprise on deliveries. And the notion that in a quarter of high gas prices, maybe on the margin, you do have some more demand, you know,

in the absence of the credits that consumers enjoy for buying a Tesla, that maybe it can bolster things around this level. Yeah, I mean, you see stories and numbers indicating that people are looking to sedans, which had not been the choice for a very long time. So really, gas prices are factoring into their... We're all going to be close to the road again. And by the way, you're... Europe was a strong spot for Tesla sales this time.

So that's where they're feeling at the most. Well, two AI build-out high flyers getting hit hard today. Storage company, Seagate and Western Digital, falling in a report that Japanese rival Tshiba, plans to double its capacity to produce hard disk drives that are used in data centers. That increase in production could ease the supply crunch and ensuing pricing power, which have led both Seagate and Western Digital up more than 200% over the last year. you know, the way you know that it's a commodity business, these two stocks are like the same chart. You know, they just really do move in lockstep. It's really an aggregate macro story.

So I guess naturally you have this potential new capacity coming on. It's not good. And the want of investors to jump to the worst case scenario on this is really telling in terms of the trade because Tashiba is a very small maker of hard-disc drives in this point. And if you take a look at how much they plan on increasing spending in order to expand this production, It's like $380 million for fiscal 2027. And for Seagate and Western Digital, they're spending $1.7 billion in the same year. So it's really, it's a tiny amount of production potentially that could come along.

I guess it tells you how sensitive the investment case is to how long this exact level of shortage is going to last. It's fascinating. Well, Invidia hitting an all-time high today, putting even more distance between itself and Apple in the market cap race. Up next, what this divergence can tell. us about investor thinking on the buildup. The NASAC 100 hitting an all-time high today, topping 31,000 for the first time ever before pulling back, closing at 30,800. The index has closed higher in 11 of the last 13 sessions. It's best stretched since April when it had a 13-session winning streak.

We've had a little health today from the biggest stock in the index. That would be Nvidia, which also hit a new high today, and that's where we start off at the dashboard. Yeah, Nvidia been playing some catch-up on Apple. We know Apple serves as a kind of defensive play within MnVDVDIA. mega-cap tech, and it's been pretty solid near its ties here. And you see it a little bit of curl higher in NVIDIA and slight pullback in Apple as kind of risk appetites and attention toward the AI trade start to revive. Take a look at the valuation that's really dramatic here because the earnings power of NVIDIA has done nothing but gone up. And therefore, the valuation has compressed. The P.E.

is down to 17 for Invidia on a forward basis, still up in the mid-30s for Apple. So you see they We're mostly in sync for a while. This, to me, just represents concerns that Nvidia is kind of over-earning in the near term. Arguably, Apple is maybe trading it too high a premium given its growth rates, but also you don't have the CAPEX risk in there of saying if this boom starts to fizzle that Nvidia and the rest of the food chain's in trouble. Yeah. I mean, you sort of see, we were talking about Micron earlier, but also Western DigitalSgate,

also sort of the same sort of phenomenon where you have this amazing, I mean, for Micron's numbers, they were amazing. in terms of growth, projected growth, but that's not doing it at this point. The market's really reluctant to just extrapolate this much more than, you know, through next year, let's say. Yeah, real skepticism. Time now for our CNBC News Update with Pippa Stevens. Hey, Pippa. Hey, Melissa. New court documents say Tennessee death row inmate, Krista Pike, was on a ventilator and unconscious

following an execution attempt earlier this week. Pike's lawyers say they believe the execution drugs flowed into her body instead of her bloodstream because of issues with the IV lines. Governor Bill Lee has ordered a review of the incident. The Omani co-pilot accused of stabbing his captain on a fly Dubai plane headed to Tel Aviv earlier this week was reportedly banned from flying by Oman over concerns. He adopted radical ideological views. According to the Wall Street Journal, it is not clear if Oman shared those concerns prior to his hiring by Fly Dubai.

Israeli officials say he intended to crash the plane. And the FDA announced a 10-point plan today to prevent future cyclospora outbreaks, The U.S. recorded nearly 20,000 lab-confirmed illnesses during an outbreak linked to the parasite this year, many of which were connected to lettuce in Mexico sold by Taylor Farms. The changes include boosted surveillance of high-risk foods and a bigger lab presence in Mexico to collect, analyze, and process samples for contamination. Mike? Pippa, thank you. Ometa, just posting its best month in four years, thanks to optimism over its AI agent Muse,

but our next guest isn't buying the hype. Laura Martin explains why she's. She's retaining her hold rating on the stock when closing bill overtime returns. Meta shares soared more than 20% in September posting its best month since 2022. The move was driven by strong investor enthusiasm over its new consumer AI agent, Mews. But one firm on Wall Street isn't buying the hype because the monetization isn't quite there. Joining us now is Needham Senior Internet and Media Analyst, Laura Martin. Okay, Laura, I mean, look, number one in the app store out of the gate.

I feel like I'm the only person who hasn't downloaded Mews. why is that not enough for you to get excited that now meta has a good AI strategy here? Right. So first of all, he hasn't said he's going to charge anything for it. And we know he's giving dedicated resources to this for computing. So that means he's using tokens that he has available. He said he had excess capacity. He was talking about leasing that excess capacity to open AI or anthropic. He's going to use it for Muse instead. So his costs are going up in the face of $140 billion, of CAPEX this year, and he doesn't have a monetization model that he's articulated yet. So this is a man who typically his track record is to grow usage for two or three years before

he starts monetizing an asset. So we just think it's too early. So you think that he's not going to come out with some sort of strategy to broadcast the street for a couple of years? I think he's going to garner new users. So they have five million downloads so far of Muse, according to Censor Tower. But there's no monetization model. So all he's doing is adding cost to his cost structure to his token costs. Because remember, he doesn't have a cloud. He's renting all of that space from Amazon or Google or Anthropics. He's renting it from somebody else.

So these are costs on his income statement, but there's no revenue associated with Mews or with the Connectors platform. Not sure we have to say it. He here is Mark Zuckerberg, who's making all the calls here on meta. You know, they have said, I guess, that they expect down the road. they'll just take a piece of all e-commerce transactions, perhaps, of muse users as these agents go out there and find a better deal for the end customer. I'm not sure if that's going to be enough. Obviously, not enough to quantify it.

But your idea on cost, I have been interested in since the launch, you know, if people are using this for, you know, relatively prosaic purposes or whether it's just to monitor their emails or whatever it might be, what is the cost of delivery in terms of the compute that they have to burn? Well, it sort of depends on your use case because the whole benefit of agents is it works in the background for you. So, for example, like right now if you do a Google search, cheapest flights to Tahiti, that's a Google search right now it has compute. But really what an agent does, what a personal agent does, which is Mark Zuckerberg's vision, is you put it on in the background and every five minutes it searches for cheaper flights to Tahiti. So you would actually give your agent the job when you find flights to Tahiti under $300, notify me, which means that agent is working in the background using tokens indefinitely until there's a flight under $300 to Tahiti at which point it notifies you. So tokens are meant to be used in the background like a personal assistant,

which means they use compute over and over and over like always on. So it's sort of compute intensive. Do you have an underlying skepticism of all agents and sort of the agent offering, brings to come, Laura? I mean, do you just sort of question the whole idea that this will be profitable, that this will be monetizable for these huge spenders of money? So, good question. So the answer is no. I am not skeptical. It is my view that consumer behavior is going to change radically over a decade, because that's how long it takes Americans to change behaviors. And they will all be using agents. And the goal of these agents is to disintermediate and take

revenue away from the internet winners, which are Google and Amazon and meta. So you see, those three are really working and not have themselves disintermediated, but it's a whole new, it's a whole new disruptive technology. And that's the only way you could justify the 800 billion that will go into CAPEX this year and the trillion one next year is if it disrupts and collects money from everything that came before. So I do believe in agents, and I do believe they will be adopted, but right now there's no revenue use case for them. What about the argument, Laura, that even if Mews itself is not going to be some kind of short-term profit bonanza, that meta stock just got a little bit too cheap, even based on the core

business, it was like 16 times forward. You had a little bit of a repricing higher, maybe just because you now have that user base engaged again on this level, and it's just back to its kind of normal valuation. Look, I think a fair point for META is when's the last time it had a new product, it didn't copy from somebody else or buy. Like WhatsApp it bought. You know, Instagram was copying Snap. Like most of its products have been sort of copying somebody else's product. This is a whole new vision for them.

So I think if it works and user adoption implies that it works as a new product point of view, you know, I think that's a positive. It means you have a new product like engine that is an asset here. But, you know, I'm Wall Street. I want to see money. Like, don't tell me about usage adoption. tell me about the money. And to your point, he's saying someday we'll take a piece of e-commerce, well, you saw Amazon just shut them out. They said, no agents allowed on Amazon. We want to keep

the relationship with the customer. So if the agent can't get into Amazon, it's going to be, that's the biggest, you know, rev share you could get if you were meta. So you just got blocked out of one of the biggest rev share platforms. And just quickly, Laura, I mean, to Mike's point about valuation, you're saying also that that investors should really take into consideration the off-balance sheet obligations that META has because that really should change the way you take a look at the stock compared to others that don't have as much of an off-balance-sheet obligation. Right. So I think one of the really key points about META is a lot of the way they're funding these enormous infrastructure builds is they're doing it through off-balance-sheet

financings that are buried in footnote 9. So we published today that they're on balance sheet debt and leases are round numbers 100 million and they're off balance sheet. Liabilities that are senior equity are 600 billion now. And all of those are going to show up in the income statement over the next decade. So your P.E., your earnings is going to have headwinds to it. And we think when you look at an EV like an enterprise value to revenue or an enterprise value to EBITDA multiple, you're 35% too low because you're not counting $600 billion of off balance sheet, really debt, because that's how the accountants let them do it. They don't make them put it on the balance sheet. So we call shenanigans. And so we would actually recommend for all the hyperscalers that are spending money on infrastructure

builds that you add back off balance sheet liabilities to the debt number to actually calculate and adjusted enterprise value before you do multiples. Laura, great to speak with you. Thank you. Laura Martin. Up next, the president of CrowdStrike on how the cybersecurity industry is tackling rising fears of rogue AI agents. And speaking of Crowdstrike, it's one. one of several S&B 500 stocks hitting new highs today, as well as Palo Alto Networks, HPE, NetApp, and EverPure. Closing bell over time, we'd be right back. Welcome back even by the usual high standard.

This was a big week for the AI industry. Monday, OpenAI said it would hold back its newest model after internal researchers raised a number of security concerns that didn't stop OpenAI's dev day the very next day, revealing more than 20 new products and features as it tries to push beyond its chatbot and into software that can work independently on users' behalf. That same day, the White House convened a Breton Woods kind of meeting of AI leaders with Sam Altman, Jeff Bezos, Satya Nadella, Alex Karp, and Sundar Pichai among those in attendance. While most seem to back the morally binding self-policing deal they signed on to, Anthropic CEO Dario Amade sang a slightly different tune.

The mechanism, how we address those risks, is still under discussion. But we all need to work. We all need to work together to make sure that we can win and we can win safely. If we do this right, we work with the president and everyone here, we can win safe. Let's talk more about the security challenges of the AI arms race. Joining us now in an exclusive interview as Crowdstrike President. Michael Santonis, Michael, great to have you with us. Thank you, having you.

We've heard so many different things in terms of there are no real, you know, huge dangers that we can't contain to. It's going to be the end of mankind. from your perch, what is the one risk that you are most worried about? I think the biggest thing right now is every week we're hearing these models are doing things that they did without us realizing. We hear stories of organizations where an agent has carried out a task in the news this week where Open AI are holding back a model. Their internal testing showed higher levels of deception. and AI failing to accurately disclose the action that it's taking.

So we're learning every week about what these models can do, evasion oversight, deceptive behavior, scope violation. And I think we're just starting to understand the breadth of these issues. So I think these issues are going to get worse. We're going to see big incidents. And we're going to keep learning on how we have to deal with these moving forward. Mike, from your perspective and from the perspective of your clients, would it be helpful if there was some concerted effort to slow things down,

or is the urgency created by the fact that we have this unchecked development of new models all the time helping you? I don't think you can take the technology back. I mean, you've got open weight models, models that are being built around the world, that don't have guardrails, that have nearly the same capability as the frontier models. They're in the hands of adversaries. They're in the hands of the people that carry out attacks. So unfortunately, at this point in time, you're not going to be able to slow down or take things back. We need to make sure that the same technologies that's in the hands of the attacker are in the hands of the defenders.

That's the only way we're going to be able to keep up. And look, there's economic realities here. We want to make sure that we are keeping up with other countries that are building these models, sometimes building it for not the greatest of reasons. And I think that's something that the reality is. we're going to have to keep pushing forward at this point. How much is AI helping you battle rogue AI, Mike? And how does that change sort of the dynamic in terms of, you know,

those AI agents that can go rogue will eventually try and learn how the AI agents that are trying to go after them are going after them? Well, Crash Strike was started with AI in mind. I mean, when Crash Rock was started, the traditional security vendors weren't keeping up. They weren't keeping their customers safe and secure. the idea was to use this technology to be able to find attacks without traditional signatures or legacy techniques to put a simpler term on it. So it's something that we've always used. And here we are today. We're using these models. We're using the frontier AI models. We're

partnering with the companies. We're partnering with Nvidia. We're using open weight models. We want to make sure that we are using the most up-to-date technologies together with our platform so we can keep our customers safe and secure. The amount of attacks that we see every week where adversaries are you using AI to accelerate their attacks to carry out steps that normally would take hours to days to weeks and they're doing it in seconds to minutes, this is the only way forward and this is what we've always done at Crowdstroke. It's essentially the same in terms of the principles of defense. I mean, is it, you know, detection and then the credentials making sure that you actually have your data protected in the same way or is it an entirely different game? It's a great question because a lot of

the traditional techniques hold true in today's world. Remember, when an agent runs, it runs with the permissions of the person that has asked it to do something. A lot of the times, those permissions are they're too much. They have too level, too much levels of authorization. The agents have that authorization and they can keep it for a long period of time. And that's where things go wrong. The big focus that we have is to make sure that we remove all of that permission. We only provide the permission for the task that is needed. And when that task is carried out and completed, you remove that permission. So you remove all standing privilege.

That solves a huge amount of the problem. I can give you a lot of examples where some of the traditional approaches of blocking and tackling hold true in today's world. You just need to make sure that technology can scale into these modern environments. with agents. And it's really, it's about agent observability and it's agent explainability, I should say, because you not only have to watch what they're doing, but you have to understand what they're doing so that if they are being deceptive, you can detect it, you can stop it, and you can make sure that you don't have a problem.

Good perspective. I really appreciate the time today, Mike, Mike Sintanis. Up next, we'll discuss encouraging new data on the next big bet on obesity drugs and how much is at stake for the pharmaceutical industry. Closing bill overtime, live from the NASDAQ market side will be right back. Drug makers are making a big bet that the next generation of weight lost drugs will not replace the current blockbusters, but rather compliment them. Annika Kim Constantino has the story. Anika. Hey Melissa, one of the next big obesity bets beyond GOP-1s involves using a new hormone called amylin. Lily, Novo, and other companies are developing shots, pills and combination regiments that target amylin. And that's a hormone released along

alongside insulin when you eat that helps regulate hunger and fullness. The idea here is that amylin could boost weight loss on top of existing treatments or provide another option for people who don't get enough benefits from GLP1s alone. New data from Lilly this week supports that. The company's experimental amylin drug delivered greater weight loss when combined with terseptide, the active ingredient in Zepound and Minjaro, in a phase two trial of patients with obesity and diabetes. Wall Street sees a big opportunity here. Learing Partners estimates Lilly's broader Amalyn franchise

could generate more than $23 billion in annual sales by 2035 and points to a potential market of more than 10 million people who don't respond adequately to GLP1s. Lily's drugs won't reach patients for years, but Novo's own combination treatment, Kagrisema, will likely be the first major tests of the Amalin class when it enters the market next year, Melissa. How should we think about this? Are they going to be combined when they're sold with the GLP1? And what does that do to patent protection? That's a great question, Melissa.

And so with NOVA, what they're doing is having a sort of dual product where there's two different sort of injections that you'll be able to inject at the very same time. The drugs aren't going to be put together. I do think the patents will still be separate in that case. Same thing with Lilly, they're going to do a sort of co-formulation product that's going to have the two sort of injections available. But as of now in clinical testing, Lily is testing them as separate injections that patients do separately right now. Attica, thanks. Attica, Kim Constantino. Another weekend for the bomb market to think about what it needs to do. And we were talking about sovereign risk and all the

rest could get interesting. That's going to do it from overtime.