Closing Bell - Closing Bell Overtime: Stocks Bounce Back But Still End Week Lower as Markets Weigh Rising Risks 8/21/26

Episode Date: August 21, 2026

Roosevelt Bowman, Senior Investment Strategist at Bernstein Private Wealth Management, assesses the broader market backdrop and where investors should focus heading into the final weeks of summer. Former Federal Reserve Vice Chair Alan Blinder breaks down the outlook for rates and the economy ahead of Jackson Hole. Crypto markets remain volatile despite a recent rally: Andrew Moss of Jefferies explains why he believes it is still too early to declare the correction over or expect a sustained run to new highs. Plus, the increasingly complex financing behind the AI boom and how off-balance-sheet investments are helping fund the infrastructure buildout. Warren Pies of 3Fourteen Research warns that macro risks are rising and outlines what could challenge the market’s momentum. Finally, Harsh Kumar of BMO on his launch coverage of semiconductors and quantum computing and explains where he sees opportunities across both technologies. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The bell is bringing an end to the trading day. At the NYSC, Manchester United, marking the start of the Premier League season and at the NASDAQ, stable coin X ringing the bell to close out the week. Welcome to closing bell overtime live from Studio B at the NASDAQ market site. I'm Mike Santoli. Melissa Lee is off today. Stocks higher across the board to finish out the week, rebounding from yesterday's losses. The Dow up about 500 points, the S&P 500 and NASDAQ both up around a half a percent,

nearly a 1% gain for the Russell 2000. That was not quite enough to wipe out this week's losses, though. The Dow down nearly 1% on the week. Bigger loss for the S&P 500 and the NASDAQ, which fell about 2%. Rising bond yields, putting pressure on stocks, the 10-year up about 10 basis points this week, despite the announced Treasury intervention. Uncertainty around treasuries and the dollar also contributing to a more than 20% jump in Bitcoin. That was its best week in more than two years.

The reaction to the move in bonds echoed in the massive. metals as well. Christina Parts-Nevelas joining me now with more on that and some of the other market moves. Yes, we had a broad-based rally, but materials really led as copper, gold, and silver all climbed, and the miners just amplified that move today. And we saw that with Freeport Mac Moran, and one of the markets' clearest liquid proxies for copper closed more than what, 7% higher today on reports of tighter supply, Newmont Steel Dynamics also climbed higher. For the week, though, if we look at all the sectors, it's health care that topped powered by, of course, Moderna and Merk with HCA Healthcare also helping lift the group specifically today.

Energy finished near the flat line, but higher on the week, the oil and gas, ETF, XOP, notched its 10 straight positive session. It's best run since April 2020, and just Shaiba, an 11-day streak that we had all the way back in 2010. Crypto, you mentioned climbed, which looks like risk appetite, yet some of the high-flying momentum names or even the M-TUM-E-U-M-Eat. really didn't catch a bit. Intel Micro, and we often talk about GE-Vernova, all roughly lagged today with the, you can see, Bitcoin only green on the screen there. And it's Friday, so maybe you're enjoying a cold one. I like to say brusky, but Boston beer shareholders may not be toasting necessarily. The stock fell after the company said at CFO is stepping down, naming

an interim wallet searches for a permanent successor down two and a half percent. Summer may be winding down. You know, there's a theme here because it's Friday. But next week is a big one. The Fed Chair speaks to Jackson Hole. What he says could make or break the mood. The Fed's preferred inflation report out on Wednesday. And then, of course, NVIDIA earnings could reignite the AI trade, especially after shares fell this week, just a little bit less than 1%. For sure. You mentioned Bitcoin. Sometimes it is a risk appetite. Today and yesterday, it acted more like a commodity. It acted more like the inverse of a falling dollar, obviously gold. So really a reflection on what's happening with the longer date. It would seem more of a macro. Absolutely.

And then, you know, you say the AI trade, it did have another little stumble. I think these semis were down like 5.5% on the week. And it felt like, I'm not sure if that was reacting to specific new developments or just kind of this hangover and we had a big rally off the lows. And also just the concern about all of the data center buildouts. If there's going to be a slowdown, you know, the financing that's been coming through with the NVIDIA, Broadcom announcing their deal. I know we'll be talking about it later on in the show. But I think with NVIDIA specifically next week on Wednesday, gross margins will actually play a larger role. And I say that because so far in this earning season with a lot of these chip names,

AMD, Cisco, we'll throw it in the chip bucket, but hardware. A lot of the, even though sales climbed, their guide go up a little bit more. It's the gross margins that just failed to impress, and that's where you saw the stocks drop. The other thing, too, specifically with NVIDIA is China sales. It wasn't part of their guidance. Maybe this time around, it will be because they were able to ship certain hopper chips to China, reportedly just over the last two weeks or so. So it could be a positive.

So still some suspense over what we're going to hear exactly from NVIDIA. All right. We'll be getting into that for sure. Well, the Treasury Department, thank you, Christina. Treasury Department's attempt to lower yields not quite achieving its goal this week. Yields back to where they were before the intervention was announced. And actually, they're higher than they were a week ago. At what point could rates start to become a threat to the equity rally?

Joining us now is Bernstein Private Wealth Management, Senior Investment Strategist, Roosevelt Bowman. Good to see you. Welcome. Thank you for having you, Mike. Let's just start right there. I mean, you know, we have yields higher, maybe not absolute levels that are threatening in particular. But I guess the why kind of matters if it's just kind of this voracious demand for new debt. And I guess at some point what it might mean for somebody with an equity portfolio.

For sure. And I think the why kind of comes down to two factors. Number one, you think about the inflationary pressures that are coming from the ongoing conflict in the Middle East and higher prices of oil. The other would be kind of what's the credibility in terms of both monetary and fiscal policy. I think part of the reason you saw that retracement in yields is that many investors kind of view that buyback announcement as having a beach ball in the pool, putting your hand on it. As soon as you move your hand, it goes right back up. And I think that's really the challenge for the fiscal policy right now is that often when you're doing those sorts of measures, you're trying to buy time

and you're looking for other macro or monetary forces that were more powerful and durable to take over. Unfortunately, that buyback announcement didn't buy a whole lot of time. And now, you know, the Treasury is kind of in a tough position. And certainly part of the story, everyone keeps pointing to this is just the private sector raising so much money for the AI buildout in larger amounts and maybe for longer duration than the market is immediately equipped to digest. How does that feed into the AI theme as a driver of equity returns, of earnings? And we've had this fabulous run of earnings that's supported the market. Now it's a matter of sustainability, I guess.

Right. I think that's where we've been speaking to clients about it being a multifaceted AI trade, rather than this monolith. For sure, you can focus on those hypers and kind of what's the return on investment of these big data center buildouts, but we've been speaking to clients about diversifying into other parts

of the AI trade, if you will. So we think about healthcare innovation and some of those companies that are using AI to forecast whether someone might have a heart attack or discover new antibodies that can be the foundation for medicine that has fewer side effects and greater efficacy.

So that's, I think, one example of kind of moving away from just the producers and AI infrastructure in terms of the models, and more moving towards the effective end users of those models that can help improve the profitability in different sectors, whether it be health care or maybe even manufacturing as well. And certainly, I mean, healthcare has been, you know, kind of a big outperform in the last little while. Part of that is some biotech trial results and things like that. I guess more broadly, are you bracing for potentially a little bit of a risk-off moment in the markets, or do you feel as if we sort of have enough cushion behind us fundamentally?

Well, I think over the very short term, it's a pretty light week data-wise. And so I think all eyes will be on Treasury still and certainly some attention towards Jackson Hole. When you look at the title of the symposium, it really speaks to much more kind of wonky times. Sure. Some of the research papers I used to write at the New York Fed and not a signal towards monetary policy. Fed Chair Warsh has talked about moving away from forward guidance anyway. So I do think at least over the next couple of trading sessions could, be pretty calm. But as you move into the end of the year, that's where we're likely to see maybe

this crossover point where inflation, headline inflation, is actually outpacing wage growth. And that would be the first time in four years that we've observed that. That would lead to slower consumption in the quarter's head. So, you know, for certainly we've been speaking to clients about and positioning them a bit more defensively leaning into some of those consumer staples and health care that do well in a slowing economy. And when it comes to bonds themselves, You know, keep pointing out, the market-based inflation expectations that are built into the treasury curve are not particularly onerous. So you're almost, at least on paper, being compensated for buying fixed income around these levels. Does that make sense?

You know, I think over the longer term it does, so you kind of think two to four quarters ahead. And mainly for two reasons. Number one, when you look at where yields have risen to, it's actually kind of over the past six weeks, it's in the face of weaker economic data. We've seen data kind of come in below expectations. Often that kind of divergence resolves in itself in lower yields, right, where the economic data really drives. I think the other part of it, too, is, again, when you think about the wage picture, for sure, the current labor market picture looks fine and solid, but those kind of forward-looking

indicators, on-my-job postings, the quits rate, you know, separations, that sort of leverage between the employer and the employee beginning of the year. It was kind of back in the employee's corner, and now it's shifting more to the employer. And if you have those lower wages, or at least wage growth, it's not keeping up with inflation, that's a pretty good setup for fixed income over the next two to four quarters. Yeah, all right. We'll see how that all develops. Great to see you.

Thanks for coming in Roosevelt. Thank you so much. Appreciate it. All right. This week's bond action by the Treasury comes as Fed Chairman Kevin Worse gets set to speak at his first Jackson Hall meeting on Friday, as we were just mentioning. What can investors expect, especially after the July FOMC meeting, which includes included what Warsh called a good family fight.

And did Besson's bond action complicate things for the Fed? With us now is former Federal Reserve Vice Chairman Alan Blinder, currently an economics professor at Princeton University. Alan, great to see you. Thanks for coming on. Good to be here. Thank you. I mean, I guess we always say it's a tricky moment when the Fed is speaking or has a decision to make.

But it seems maybe more than most this time. I think the market's pricing in close to a 40 percent, chance September is a hike. We have this move by Treasury to try and suppress longer-term yields. The reaction by the market was dollar down, gold up, expressing some kind of policy flux there. So what do you think the message that Warsh maybe ought to be delivering on Friday? Well, what he ought to be delivering or what he will deliver, I think, are very different things. First, let me point out what you said right there. 40% chance of going up. That's very close to 50. that's very close to a coin flip.

I don't think Kevin Warsh is eager to push that coin one way or the other verbally. Eventually he's going to have to decide. We've become accustomed in recent years with Powell and before that, Yellen and before that Bernanke, to getting some hints often at the Jackson Hole conference. Everything that Kevin Warsh has said about, Fed talk since he was confirmed even before suggests he's not going to do that. And by the way, that's easier to do when you're not sure yourself. And he may not be sure himself at this stage.

For sure. I guess just in terms of the market setup, I made it sound like, oh, there was some kind of a mini panic here with what was happening with the dollar and gold. But if we look at just the interest rate setup, you know, for. 4.7 and change on 10-year yields, 4.2-ish on twos. Obviously, Fed funds rate 3.5 to 375 is the range. That's not an unusual yield curve shape. In terms of absolute levels, maybe it's not something that's particularly out of whack with the growth rates of the economy or where inflation has been. So is there anything there that you feel that the Fed ought to be responding to? I don't think so. As you say, these are not very very.

This is not a very unusual yield curve. The two-year that you just cited is, what, 60 basis points above the Fed funds? That's not unusual at all. Doesn't look very high. Remember, the inflation rate that the Fed worries about, which is the PCE inflation rate, is over three and a half percent right now. So, I mean, where would you expect the two-year rate to be, unless you think the inflation is going to come down dramatically in that two-year period, which is, I think, not the expectation that most people have. I make one more side remark. I don't think it's such a smart.

Your previous speaker was making the point that Secretary Besson's attempted to an impulation of the market didn't bring very much. I think that's right. I also think it's not such a good idea to be monkeying around with the treasury market. You know, eventually you're going to start undermining confidence in U.S. contribute to the undermining of confidence in U.S. Treasuries if you keep fiddling around as opposed to letting the market function as it normally functions. But to you, I don't think it will affect the Fed's decision at all. Right. And, you know, obviously we have, you know, four plus weeks, I think, till the next Fed meeting. We'll have to see how the numbers come out. But, you know, interestingly, since the July meeting, the inflation numbers at least have been a bit more benign or friendly and maybe contributing to the case that some have made that this kind of inflation is one that you should probably wait out as opposed to move quickly to tighten in the face up.

Look, that's my general view about supply-side-induced inflations, but I have to say, and as you probably know, I'm known as an inflation dove, not an inflation hawk. This inflation rate seems very, very persistent. So we like to look through these supply shocks because they're, to coin a phrase, transitory. This one is not looking so transitory, and it may not be over. You know, who knows what's going to happen to the price of oil over the next, I was going to say, year. How about month? You know, we have Donald Trump in the White House, and Lord knows what's going to happen. Yeah, it's been all over the place, of course. And then also, I suppose the other possible driver of just all of this, you know, private capital spending that's, you know, increasing demand for debt, increasing demand for, you know, real assets even, and land might be a contributor to some inflation. and maybe that's also something

that's not particularly rate-sensitive right now. I think the everything, as far as I can tell, now I'm not in the market in the middle of it, as far as I can tell, everything that's going on having to do with AI, et cetera, which involves acquiring land, acquiring water rights, acquiring electricity, among other things,

has nothing to do with interest rates. Yeah. And it would take an extraordinary move in interest rates to deter some of them. Yeah, it does seem that that's the case. That's got that momentum. Alan Blinder really appreciate you being with us. Thank you very much.

You're very welcome. Watching there, the CBO close out in Chicago, that brings the end of options trading for the day and the week. And by the way, don't miss CNBC's full coverage from the Jackson Hall Conference, including key newsmaking interviews. That starts on Wednesday of next week. This week's bond market drama sending some investors looking for alternatives, at least some of the money, piling into Bitcoin, which rallied more than 20 percent, Still about 50 grand from last year's highs.

So is this the beginning of the climb to new records? You're watching Closier Bill Overtime, live from the NASDAQ market sector. Shares of raw store is up 4% today following results last night, beating on revenue and same store sales and raising guidance as well. The analyst at William Blair saying the company's strategy of not raising prices is helping it take market share from rival. Shares up 4.4%. Target also had a good week following its results,

putting those two stocks pretty much level, terms of market cap right now, but it was a rough week for Walmart, losing 10% of its value. Its worst week in four years, seems as if nearly every firm on Wall Street cut its price target on the stock today did manage to slightly finish above the flatline on the day. Bitcoin up 23% this week with the rally starting on Wednesday after the Treasury Department's intervention in the bond market. Other cryptocurrencies and crypto-linked stocks like Coinbase Circle and MicroStrategy, also moving higher on the back of this rally.

But Bitcoin is still down, double digits this year. So is the crypto winter really over? And could we see a climb to new highs? Joining us now is Andrew Moss. He is the head of digital assets research at Jeffries. Andrew, it's good to have you with us. It clearly was kind of coiled up, Bitcoin was, and other cryptocurrencies for quite a long time, kind of going sideways at these relative low levels.

We released higher. What's next? Mike, so there are really three primary catalysts powering the current rally across Bitcoin, tokens more broadly, than blockchain native equities. The first is what you just mentioned. Secretary Besson announced increase in Treasury buybacks. That triggered $3 billion in short liquidations that drove upside volatility. But it's really important to remember that Treasury buybacks and forced liquidations are not sustainable long-term drivers of industry growth. So we would be cautious at these levels. Trading volumes across Spot and perpetual futures are at three-year lows,

spot token ETFs have seen sizable net outflows. And if you look at on-chain data, small and medium-sized Bitcoin holders are actually selling into strength, which partially explained by the rally is fading. It's interesting. So you take those signals of relatively low volumes and smaller traders selling into this rally as being net negative on a forward-looking basis as opposed to, you know, a contrarian view of saying, hey, nobody likes it. Nobody's sponsoring this move? Well, the on-chain data shows what it shows. And until you start to see trading volumes and per volumes increase, and you start to see sustained ETF inflows, it's hard to make the case that

Bitcoin will begin to, you know, climb the next leg of growth back above 120,000. Right. You need it to show itself first, I suppose. So what is now your preferred way of trying to play some themes in digital assets then? If it's not sort of purely long, the dominant coins? Sure. So we actually are long-term bullish on Bitcoin. We don't do price objectives. But we would also look at many of the defy applications and the tokens that power them.

They've seen significant adoption and growth that's driving upside to the token prices there. And we'd also look at blockchain-native equities, Coinbase, bullish, securitize, and Robinner. it. How dependent is this whole space on the passage of the Clarity Act? I mean, obviously, the prices move in response to the perceived prospects of it passing, even though I don't think the betting markets have very high odds of it actually getting through. But, you know, to what degree is that sort of make or break for a lot of these assets? Sure. So if you look at polymarket odds, they're at about 30 percent probability of passage by year ends. Now, we view a comprehensive U.S. regulatory framework as critical for mainstream digital asset adoption.

Now, at the same time, you've seen SEC rulemaking proposals and guidances, and those definitely move the needle, but they're not a replacement for legislation, given that a future less supportive administration could roll them back. So, I mean, clarity is critical. Right. Okay. So we do have to, I suppose, kind of monitor every step of that process. and assume the market's going to move pretty dramatically in response to how that goes. Andrew, I appreciate you checking in today. Andrew Moss from Jeffers. Thank you. Thank you.

Coming up, Broadcom is making another big move to finance its spending amid this AI boom. Those details are ahead. Plus, the bullish case for Nvidia ahead of its earnings results on Wednesday as it closes lower for a sixth straight session. Closable overtime is back right after the time. Welcome back. Samsung getting a nice pop in overseas trading, the company announcing a blockbuster shareholder return package, saying it expects to return between $65 billion and nearly $80 billion. This follows South Korean rival S.K. Heinex announcing its own nearly $30 billion buyback earlier this week.

HSBC named Heinex a preferred pick this week, writing this is the beginning of a full-fledged shareholder return plan. Let's stay with tech and AI. Broadcom is making its biggest move yet to finance the AI boom, but this time it is not. not coming directly from its balance. She. Christina Parchinev is back here with more on this one. Hey, Christina. Yeah, they want to sell essentially billions in AI chips. The problem is its customers just simply cannot purchase it.

And the issue with that is that it's just so expensive. So it's helping finance these purchases and keeping the debt off of its own books. CNBC has learned Broadcom is in talks to raise more than 60 billion. David Faber even saying 70 billion for a new AI chip deal, a package that could hit $100 billion, first reported by Bloomberg. It would benefit Anthropic and others and could resemble XPV, which is this financing platform Broadcom set up in June with Apollo on Blackstone. Also why the stock may not be reacting as much because they kind of alluded to this just early June. So what it is, it's a separate entity, SPV, that raises the money, leases the chips to the customers.

So Broadcom's role, it co-signs more so this senior level of tranche of debt. If a customer can't pay, Broadcom would cover the gap. CEO, Hock Tan, has rejected the word backstop on earnings calls in the past, even just this past June. But Bank of America estimates that exposure could reach roughly $370 billion by 2029. Broadcom isn't alone, of course. InVDHA has a similar platform targeting $500 billion, but only backstops up to 25% of each deal. Med has done a version of this, too, raising roughly $27 billion for a data center. It now leases back.

It all works, while AI demand pretty much. much outrun supply. The risk is the day that it doesn't. And so that is the jittery comments that we were talking about just within 20 minutes ago and why it's affecting the AI trip. Yeah. I mean, it's almost everywhere you look kind of the AI build out is sort of outgrowing whatever infrastructure we had to deal with it before, right? And yet it hasn't even hit the ground with boots and like no, exactly right. I mean, I guess on some of you could say, you know, equipment leasing facilities exist in the world. You know, you know, you're not. You know, have captive finance arms of vehicle makers.

You know, in other words, this isn't completely novel, but it is fascinating that it creates this chain of incentives, obviously, to have, you know, a lot more money flowing through to these customers to buy the chips. Everybody's hoping the value of those chips holds, right? And then, as you say, demand for renting these, you know, this capacity remains. It's not novel, to your point, within, you know, construction, et cetera, but it is relatively new to the tech world. And I think for companies that have been asset life for so long, now linking themselves to data center buildouts, that has created some concern to your

depreciation comment. You constantly hear from core of CEO, all of the Neocloud CEO, and of course, Jensen Wong himself, the CEO of Nvidia, just commenting on how they're still selling out of chips that are at least six years old in the market. And then you look at spot prices that continue to trend, you know, upwards for the last little while on a lot of these GPUs. They're arguing that everybody's going to need these older models. But then that makes me think, what does it mean for the cadence of Nvidia chips? He's trying to get these new chips out every single year. If all of these old ones are still just as good, will that change the cadence, especially when you have so much debt linked to these older GPUs? Like, do you need to come out

with the, you know, new ones every single year? Right, exactly. So what the kind of obsolescent cycle is, or maybe the way they would put it is, you know, if the new chips are that much better for certain extreme workloads, and we have to push the models in a certain direction, but it's good enough for everybody else. I don't know. I guess that's the hope. It all comes down to, I guess, money right now

and then keeping, hitting that $3 billion over target. Exactly. And even though this chunk is not slated to come directly and be put on Broadcom's balance sheet, it's still money being demanded from the capital markets. And if I can make the point that we often hear, we have these headlines, these flashy big numbers, and none of it is actually concrete, right?

So Broadcom is like up to talks, Target number. You know, MOUs for NVIDIA and they sign these deals. And who knows how much, a big dollar sign. Yeah. That dollar sign is going to be there, you know, a year from now. We'll see when the closing happens. Yeah, exactly.

All right, Christina, thank you. Time now for CBC News Update with Frank Collins. Hi, Frank. Hey, Frank. Hey there, Mike. The Department of Justice and TikTok, I've reached a $400 million settlement in the lawsuit against the social media platform alleging it violated children's online privacy laws.

According to an announcement shared first with Axios, The DOJ says TikTok has undergone significant changes since the Biden Justice Department filed that suit against the company in 2024. This is one of the largest recoveries ever obtained in a case under that privacy law. The Pentagon has fired the editor-in-chief of the military newspaper Stars and Stripes. Eric Slavin says he, along with a publisher and a reporter, were let go. Slavin said Friday that he was dismissed for insubordination after an interview he gave that objected to any potential censorship by the U.S. military. And Mark Stad has completed a deal to become the controlling owner and largest shareholder of the NBA's Minnesota Timberwolves and the WMBA's Minnesota links.

Stad is buying the majority of co-owner Mark Lorry's stake in the two franchises. Sources say that deal is valued at $4.5 billion. And based on recent deals, Mike, seems like he might have got a discount. Back over to you. Yeah, I mean, I guess, you know, geography and brand and history matter, but that's not a bad number. Frank, thank you. Thank you. Coming up, Warren Pyes of 314 research says market structure is becoming less supportive of this market rally. We'll look at why and what it could mean as we head into September. Welcome back to closing bell overtime live from the NASDAQ market site.

Stock's higher across the board today. The Dow up 517 points, 4 tenths of a percent for both the S&P 500 and the NASDAQ, still read for the week, a 2% loss for the NASDAQ as the rise in Treasury yields weight on stocks. A good day for Elon Musk, though, SpaceX, up 2% Tesla gaining 5%, continuing a recent comeback, but still the worst Mag 7 name of this year. It is down 19%. Check out Sandisk, notable for its lack of a big move, 20 straight sessions with a move of at least 2%. Six of those days, it moved at least 10% in either direction. But that volatility taking a break today, it was off just about one quarter of 1%. So with strong earnings growth,

ambitious AI expectations and some stronger that expected economic data lately, should investors continue to play offense or is it time to take a more cautious approach in the months ahead? Joining us now is 314 research co-founder Warren Pyes. Warren, good to see you. Nice to be here. You took a step to, I guess, a more neutral footing in terms of equity recommendation about a week or so ago. What were you seeing there in terms of, I guess, the dynamics, the interplay between rates and stocks and the AI trade? Yeah, so I think if you go back one month ago, there was a healthy reservoir of skepticism and pessimism out in the market. We saw it expressed through single stock

level volatility, even though the VIX was still low because correlations in stocks were moving against each other so much. Single stock volatility was high. It was over 50. And in our view is that if you look at earning seasons, as long as results come in strong, then usually you get a ball crush, which is bullish for the market. And we've had that happen. So earnings is obviously a very strong season. Nobody can, and no one can stop talking about it now. But that's in the rearview at this point. Earning season basically 90% over. And now we have to think about macro. And we're left with lower volatility and still low correlations. And so if there was a macro event that were to kick off, let's say the Fed or an S1 drop from one of these labs and then you have to look forward to midterm elections,

you could spike volatility and I think have a little bit of a hiccup in the equity market. So from our perspective, we've been overweight stocks really since mid-April. I think this is a window where you want to neutralize and start de-risking a little bit. And in terms of the way the bond market has been kind of trying to navigate a lot of what's going on with the Treasury and the Fed uncertainty, what is it implying to you in terms of rate hike expectations? And is it mispriced in some way? I mean, is it these level of longer-term yields worrisome? How does that fit in? Yeah, I think that's, when you're talking about macro, you can really divide into two big camps.

Growth and inflation and inflation manifests through rates and interest rate policy. And our view is that, look, set aside what the Fed should do. That doesn't really matter at this point. But what the Fed's going to do in September, I think it's a coin flip. And the market went from 75 percent chance, again, back in July, that we're going to see a hike in September and they've taken those odds down to 30%. Our view is that it should be a coin flip, that this is a very live meeting that the intervening data that we've received, the CPI and PPI, when we plug that into our PCE nowcast, core PCEs set to come in, like, say,

from our estimates, 0.28% month over month. Anything above 0.25%, I think you're going to see a violent repricing of these odds for a rate hike in September. And so that comes out here soon. We've got Jackson Hole next week. These are the macro things we're talking about. And contrary to popular belief, I think if the Fed starts hiking, you're going to see the 10-year yield move higher. There's this misconception out there that you're going to see yields come down if the Fed were to hike rates.

That's not what history suggests. So that's what we're watching for. Obviously, the Treasury's doing what they're doing, but hasn't been very effective so far. Yeah, I noted that work you did on exactly how the longer. term yields tend to respond to hikes, especially when you're kind of moving from cutting to hiking as we would be at this point. I want to be, I give you a chance to mention, you think it's probably a mistake if they hike, but they may do it anyway. Yeah, I mean, I look at it as we're basically restrictive right now. I'm on with the core of the committee. I think it's the Waller and Powell and

Williams group is the core of this committee. And what they decide is what's going to happen. I worry that a hike is really going to be more about political pressure. You know, they've looked through, and you see this in their speeches, they've looked through so many of these supply shocks over these years, whether it's COVID, and I think that was a misdiagnosis, but tariffs and now oil. And I think that, so if they do hike, I think it's going to be largely a bowing out to political pressure in how their credibility is seen. Ultimately, I do think that's a mistake, but, you know, they have their, they're, they're, they're, they're, they're, all humans just like us. Yeah, you can sort of pick up a bit of impatience building in the committee just with this wait and see approach with inflation remaining above target. Finally, Warren, I know your work on, you know, the demand for AI compute capacity is maybe softened up a little bit. I wonder if you could just shed light on that, what it might mean for that trade.

Yeah, so we track GPU availability. We've been doing this since 2023. And so we go out to the neoclounds and basically ask. every hour, how available are each of these GPUs? And we put that in a composite index. And so this last, just last 72 at four days or so, we've seen a real sharp uptick in availability, which leads us to read that as weakening compute demand. So we dug into that for our clients this week. At this moment, we see this as benign. I don't think it's a disruption to the overall AI story. It's really coming from older GPUs that we see. The availability there is

spiked. And I think that just inference economics that really favor the newer GPUs and the availability in, say, the Blackwell has stayed very tight. So demand very strong there. As long as that's intact, then we're going to be okay, I think, on the compute story. But it feeds back to this whole risk window that we're in. The market is sentiment is kind of expecting everything to go right for the AI trade, for the IPOs, for the AR of the labs, all these things. This is that seasonal period where any of these things go right. wrong with this kind of expectation, you're going to have an indigestion. Yeah, maybe not on the right footing for that.

Warren, really great to talk to you. Thank you very much. Have a good weekend for Pies. All right, ahead. A big week for tech and Nvidia getting set to report on Wednesday. We'll speak with an analyst who initiated the stock with a buy today ahead of those results and the stock that had both its best day ever and its worst day ever in the span of 48 hours. I'll tell you about that.

Close your bell overtime. It's back after this. Welcome back to closing bell overtime. Moderna shareholders celebrating their best week ever, closing up over 130 percent following blockbuster cancer drug trial results. But the week wasn't without some side effects. The day after its best day on record, Buterner posted its worst day ever. It was down nearly 25 percent. Buterna giving a lift to the health care sector as a whole, which closed out its best week since June 26th, up nearly four and a half percent.

The SOX ETF falling five percent this. week. It was its first down week in the last three. But Bimo is out with a bullish note initiating coverage on semiconductor names like NVIDIA, AMD, Broadcom, Micron, and others with an outperform rating. The firm naming NVIDIA as its top pick, calling it the de facto leader with full-stack hardware and software and market share dominance. This comes as Wall Street gets ready for NVIDIA's earnings results next Wednesday. Joining us now is the author of that note. Harsh Kumar from BMO. Great to see you, Harsh. I mean, you had a chance, I guess, to size up the whole group and sort of relaunch coverage here. What, in your view, is the market

missing about Nvidia after, you know, this period when the stock is not quite performed and had its valuation compressed? So, Mike, first of all, thank you for having me on your show, as always. Look, we look at AI demand. We're very bullish on AI. We look at AI demand as in a very simple manner. There's a near-term demand. There's a near-term demand. There's a new-term demand. there's a near-to-mitterm demand and the mid-to-long-term demand. The near-term demand is all being given by hypers. There is some enterprise demand in there, but hyperscalers are overshadowing all the press. In the mid-term, you have real companies such as banks, pharmaceutical manufacturing,

that kick in. That's about two years plus out, two to five years. And then mid-to-long-term, there's self-driving robotics, you know, sovereign, things like enterprise that continue to kick in. And so we look at very long legs. And then the other thing you have to take into account is LLMs are getting more complex. They're up about 10x on year-on-year basis. I can go back and tell you that the latest LLM is about $1.5 to $2 trillion parameters. Two years ago, Chad T3 was $175 billion.

The tokens used are going up dramatically. But in order to make the token economics work, you need to drop the cost for the tokens. and the way you do that is through upgrades. So even for these hypers, they need to constantly think about upgrades to have reasonably economic tokens, which is kind of the core of our thesis for hyperscalor upgrades. But then, of course, you have enterprise. You have these new applications.

We are not concerned about Nvidia earnings whatsoever. We think demand is exceptionally strong there. I suppose the demand side is not what a lot of folks have questioned about, at least in the near term. I assume Nvidia next week will be asked to elaborate further on a lot of sort of the financing efforts and the stakes it's taking in other parts of the ecosystem. I mean, how much is that part of your conversation with clients about the name? It has come up a little bit in the conversation with clients.

Look, the reality is some of these private companies are staying private for a very long time. In my initiation, I cited Anthropic, it's doing something like $11.5 billion a quarter, up almost 1,400 percent, something like that year on year, 400 percent sequentially. My point here is something that big that is growing that fast has required a lot of capital. Wall Street didn't have a great mechanism because a lot of the big banks didn't understand the business because the business wasn't making any money at that point. But companies like Nvidia did. And so companies like Nvidia were very happy to finance the growth because they understood these models.

Now, Nvidia has a very, you know, clear clause that their capital is not to be used for their equipment purchases. So, yeah, you know, I think finally we're getting some capital that is allocated through Apollo and some of these other large financing arms along with Nvidia to facilitate this kind of growth that we're seeing. but it's a function of companies staying private for long. It's also, I look at the capitalism moat. It stands right there with technology, Kuta, software, the ability to manufacture complex systems. And if you have the balance sheet, why not use it? Well, certainly it is an advantage.

No doubt about it. Real quick word on the analog semi-group, which you also think may be due for an upswing. Yeah, so look, economy, despite everybody's worry, is staying pretty strong. We are seeing a definite turn in industrial. We are seeing, we've already started to see that since the beginning of the year. We're seeing a turn in automotive. And of course, defense has been very strong. Those are the three biggest categories, if you will, of the analog space. And so between those three categories, analog is seeing a resurgence. We are seeing things like

lower inventories, ASP increases, expedite orders from customers. And my suspicion is that the cycle will probably last for many quarters from here. So we're bullish on analog, particularly those that have industrial exposure, and of course, very bullish on Nvidia with the cheap multiple and the AI complex overall. I appreciate you coming to sum it up for us, Hars. Thank you very much. Always your pleasure. Thank you. Hars Kumar. Up next, everything that could impact Wall Street next week, there's more than just NVIDIA, closing bell overtime live from the NASDAQ market site. We'll be right back. The options market is getting set for some key economic data, including the PCE price index,

as well as FedSpeak with some interesting action in both bonds and gold. Oliver Renick is at the CBO in Chicago with more on that, Oliver. Hey, Mike, this week's Treasury buyback has options traders looking much more on edge compared to most of the summer when it comes to bonds. We saw traders buy 10 times more puts than calls on the 10-year note futures, which are popular as a short-term hedging tool. Any further slippage in the last. long-term ETF TLT next week might exacerbate volatility based on what I see in options positioning

and put buying out-based calls 10 to 1 on high-old corporates in HYG. Now, what's bad for bonds looks like it might be good for gold right now. GLD is on a 14% run this month and 440,000 calls traded in GLD today compared to 150,000 puts. 87% of the nearly half a billion premium in GLD was tied to calls, including someone who spent $38 million buying $20,000 of the 450 strike calls in GLD expiring in mid-January. Mike, that was the fourth biggest trade in the entire market today. Yeah, obviously some big bets being laid ahead of next week.

Oliver, thanks very much. Talk to you again soon. That is going to do it for overtime.