NVDA 全部观点

I detailed my full thesis in a video last week where I discussed why I think Nvidia will be the first $10 trillion company. for context, it's currently at a $5 trillion market cap.

The growth of Nvidia's actual business has been historic over the past few years. They have beaten earnings every single quarter going back to early 2023. That's 14 straight EPS, and they have beat 19 of their last 20 revenue estimates. In the most recent quarter, they beat their earnings per share number by 6.3%. That quarter came in at $187, up 130% year-over-year. And in the past 5 years, their quarterly earnings per share is up 1,458%. That's an 84% compound annual growth rate.

However, just because Nvidia keeps growing and beating their earnings, that does not mean that the stock always pops after reporting earnings. In fact, Nvidia has had four straight earnings beat and guidance raise quarters, and the stock price fell each time the following day. The most recent one, the stock fell 0.8%, the one prior 3.25. 2 5.5 1.8 despite beating earnings per share expectations among analysts raising guidance and seeing historic earnings per share growth. In fact, Nvidia stock hasn't had a post earnings pop in the stock price since May 2024.

When Nvidia stock falls, the whole market, especially the NASDAQ 100, falls with it. And that's because Nvidia is the top AI semiconductor name and it can move that entire trade, the entire supply chain.

So, is Nvidia going to go up or down after earnings? I think the answer to that question has to do more with market sentiment than the actual earnings results because it seems fairly obvious they're going to beat and raise guidance again, especially since we've already gotten the earnings reports from all the big tech companies putting out the large capex spend in the data center buildout.

But as Warren Buffett says, in the short run, the market is a voting machine. And the market has been voting in the past four quarters for Nvidia that they don't believe the growth will keep going.

Nvidia's stock price is only up 15% over the past year, while its trailing 12 months earnings per share is up 64% in that time. They earned $48 billion of free cash flow in the latest quarter.

The other part of that Warren Buffett quote is that in the long run, the market is a weighing machine. Over the long term, the market will weigh the intrinsic value of the company. So, while Nvidia's stock price hasn't done much over the past year, it certainly is becoming a heavier and heavier business. The second most free cash flow generated in the entire stock market over the trailing 12 months, $119 billion. and the company is still growing quarter after quarter after quarter. Nvidia will soon be generating the most cash of any company ever, but that doesn't mean it's going to pop on this earnings report.

In today's video, I'm going to give you my thoughts on this upcoming earnings week, the stocks I'll be watching and that I'll likely cover in detailed follow-up videos that includes an extended look at Nvidia and some beaten down software stocks that could be value plays companies like Inuit and Salesforce.

Some dividend growth stocks that have been slowing down and a semiconductor company that Nvidia CEO said a few months ago could be the next $1 trillion semiconductor company.

But then the big story of this earnings week is Wednesday after market closed. That's Nvidia, ticker symbol NVDA. Now, I dove into this a ton at the beginning of the video, but that's because it's the big story and it's also very interesting. I also happen to own a stake. Full transparency. Do your own research. Don't just blindly follow me. And it's one of the most impressive stories in the history of capitalism, the growth of the data center buildout over the past 3 or so years. You can see the expectation is $92.2 billion in this quarter for revenue and $29 of earnings per share. I would expect them to beat in both of those. And it's the growth of Nvidia, the actual growth of the fundamentals, which is the reason why the stock price has been up so much over the past 5 years. 838% price return. That's a 56% compound annual growth rate. But as I mentioned earlier, the stock price is only up 15% over the past year despite the actual cash flow the company generating being up 65% year-over-year in that time. Earnings per share up 83% year-over-year. And this is all still ramping up. They just had $ 48.5 billion of free cash flow in the latest quarter. As I went over in my full Nvidia video, Jensen just a few months ago, I forget what month it was, like May or something, they had a GTC conference and he was up there giving a speech and he said they had line of sight to $1 trillion of revenue over the next two years. And if you've been following basically all the guidance they give, they end up beating. So, they're kind of conservative with the guidance they give out. Now, that wasn't specific financial quarter guidance, but he did say it on stage in front of everyone. Over the trailing 12 months, the company's generated $253 billion of revenue, $81 billion in the latest quarter. That's up 85% year-over-year. And that was the first quarter in the time of where he said they would do $1 trillion of revenue. So what does that mean? It means they're going to be getting to an average of $125 billion a quarter over those eight quarters. Not going to get there this quarter, not going to get there next quarter. So that means they're going to be well above that on the back end. Right now, analysts are expecting $394 billion in fiscal year 2027 and $570 billion in fiscal year 2028. Add those up, you're a little below a trillion. So, right now, I would say analyst estimates for revenue are a little low compared to what's likely going to happen. All of the demand signals are still super strong in the

That's Nvidia, ticker symbol NVDA. Now, I dove into this a ton at the beginning of the video, but that's because it's the big story and it's also very interesting. I also happen to own a stake. Full transparency. Do your own research. Don't just blindly follow me. And it's one of the most impressive stories in the history of capitalism, the growth of the data center buildout over the past 3 or so years. You can see the expectation is $92.2 billion in this quarter for revenue and $29 of earnings per share. I would expect them to beat in both of those. And it's the growth of Nvidia, the actual growth of the fundamentals, which is the reason why the stock price has been up so much over the past 5 years. 838% price return. That's a 56% compound annual growth rate. But as I mentioned earlier, the stock price is only up 15% over the past year despite the actual cash flow the company generating being up 65% year-over-year in that time. Earnings per share up 83% year-over-year. And this is all still ramping up. They just had $ 48.5 billion of free cash flow in the latest quarter. As I went over in my full Nvidia video, Jensen just a few months ago, I forget what month it was, like May or something, they had a GTC conference and he was up there giving a speech and he said they had line of sight to $1 trillion of revenue over the next two years. And if you've been following basically all the guidance they give, they end up beating. So, they're kind of conservative with the guidance they give out. Now, that wasn't specific financial quarter guidance, but he did say it on stage in front of everyone. Over the trailing 12 months, the company's generated $253 billion of revenue, $81 billion in the latest quarter. That's up 85% year-over-year. And that was the first quarter in the time of where he said they would do $1 trillion of revenue. So what does that mean? It means they're going to be getting to an average of $125 billion a quarter over those eight quarters. Not going to get there this quarter, not going to get there next quarter. So that means they're going to be well above that on the back end. Right now, analysts are expecting $394 billion in fiscal year 2027 and $570 billion in fiscal year 2028. Add those up, you're a little below a trillion. So, right now, I would say analyst estimates for revenue are a little low compared to what's likely going to happen. All of the demand signals are still super strong in the data center buildout. You had basically all the hyperscalers raising capex. They're starting to see operating cash flow growth. Their stock prices finally went up. they were being punished for the capex increases. Now Amazon and Microsoft, they finally got a little bit of a pop from it. The cloud growth is really good. And on the actual demand side of AI usage and AI inference and token generation, it's all just exploding. I went over this a lot in my Nvidia analysis, but I'm just doing some back of the napkin math here for you guys. So latest quarter, they did $81.6 billion. That was $48.5 billion of free cash flow. So, if they're going to do a trillion dollars of revenue over the next two years, that means they're going to have roughly $500 billion of free cash flow generated in that time. That will be the most amount of free cash flow ever generated over a 2-year span. That will give Nvidia one of the best balance sheets in the history of corporations, funded by amazing cash flow. And Nvidia is not a capexheavy company themselves. They get the chips manufactured by TSMC. Super optimized for free cash flow generation. The fab semiconductor model. It's a beautiful free cash flow business model. They have started building some data centers themselves though to train open models which I think is a good thing because it just helps spur demand in general. But if we look at Nvidia over the trailing 12 months, they've spent $47 billion in share buybacks. I would expect their return to shareholders to go up massively as they continue to generate more and more cash. Again, my quote at the beginning, short-term, the market's a voting machine. In the long run, it's a weighing machine. Nvidia is becoming a heavy business and their capacity to reward shareholders and their intrinsic value of the company is growing. And that's why they just raised their dividend payment, which they rarely do. They raised it by 2,400%. And now it actually has a reasonable yield of 0.48%. All the signs for Nvidia are that growth is set to continue at least for the next year or two. People will have tons of debate of what year three, four, five, and even 10 years from now is going to look like on the demand side for Nvidia. We don't have enough time to go into all that in this video. Right now, I would say the consensus in the market with how things are being valued is that Nvidia, we're close to the peak and there's not going to be a lot of growth from here. That's why the stock price hasn't gone up much in the past year. It's also why it's trading at a very reasonable valuation. The PE over the trailing 12 months is 31.9. The P forward-looking is 23.12. So if we look at the forward-looking P ratio over the past five years, this is actually historically low for Nvidia. The median is 38.27. So we're in the 15th percentile right now. And let's just say we base it on the trailing 12 months P ratio. The current is 35.7 for adjusted earnings per share. The median over the past 5 years is 53.7. So if we assume that as being the fair value for Nvidia stock, the median multiple, Nvidia is currently trading 34% below that median. If we do it based on forward-looking earnings per share, so next year's earnings, the median has been 38.27 over the past 5 years for Nvidia. This implies an even larger discount for Nvidia stock. And if the company got rerated back up to that 38 multiple, that'd be 65% upside from here. But the thing you have to remember is that Nvidia is growing very fast still. It's just unclear how much longer it can continue growing. And frankly, I was wrong about this initially. I didn't think Nvidia could even get to where it is here. I was not aware of how big the market could be. And I think that's still might be the fact that people are being early in calling AI bubble. We still have many, many years of growth ahead there. I'm not saying there can't be an overbuild in like 2030 or at some point in the future, but it's not the current conditions on the ground. And I think people struggle to think in exponentials. We're talking about very high growth rates right here for Nvidia. The current analyst expectation is 92% growth year-over-year in fiscal 2027. And then they're expecting that to go down to 42 and then 20. But really, the 3, four, and 5 year completely changes all the time. People have no idea what to think in those later years. And given the multiple Nvidia is trading at, which is cheaper than like Coca-Cola and Proctor and Gamble on a forward-looking basis, when its actual fundamentals are growing 80% and the stock price is only up 15% over the past year, I don't think the market is pricing in the scenario where Nvidia grows 80% year-over-year again and then it does that the next year and potentially does that the next year. If that happens or even if it goes 80%, 60%, 50%. That's a ton of growth and Nvidia would look very cheap today based on that. And then the question is, is the company done growing after that? I'll let you decide that one. So, this is definitely the most interesting earnings report to cover during the week. It will definitely impact a lot of the broader market in the coming months because it has to do with the entire tech and AI trade, which is a large part of the overall market. So, I look forward to reading through the earnings report and listening to the earnings call and I'll be back with a video to analyze that. So, if you want my thoughts post earnings, subscribe to the channel and I'll be giving you an update on that.

And it's the growth of Nvidia, the actual growth of the fundamentals, which is the reason why the stock price has been up so much over the past 5 years. 838% price return. That's a 56% compound annual growth rate.

But as I mentioned earlier, the stock price is only up 15% over the past year despite the actual cash flow the company generating being up 65% year-over-year in that time. Earnings per share up 83% year-over-year.

And this is all still ramping up. They just had $ 48.5 billion of free cash flow in the latest quarter.

As I went over in my full Nvidia video, Jensen just a few months ago, I forget what month it was, like May or something, they had a GTC conference and he was up there giving a speech and he said they had line of sight to $1 trillion of revenue over the next two years.

And if you've been following basically all the guidance they give, they end up beating. So, they're kind of conservative with the guidance they give out. Over the trailing 12 months, the company's generated $253 billion of revenue, $81 billion in the latest quarter. That's up 85% year-over-year.

And that was the first quarter in the time of where he said they would do $1 trillion of revenue. So what does that mean? It means they're going to be getting to an average of $125 billion a quarter over those eight quarters. Not going to get there this quarter, not going to get there next quarter. So that means they're going to be well above that on the back end.

Right now, analysts are expecting $394 billion in fiscal year 2027 and $570 billion in fiscal year 2028. Add those up, you're a little below a trillion. So, right now, I would say analyst estimates for revenue are a little low compared to what's likely going to happen.

All of the demand signals are still super strong in the data center buildout. You had basically all the hyperscalers raising capex. They're starting to see operating cash flow growth. Their stock prices finally went up. they were being punished for the capex increases. Now Amazon and Microsoft, they finally got a little bit of a pop from it. The cloud growth is really good. And on the actual demand side of AI usage and AI inference and token generation, it's all just exploding.

I went over this a lot in my Nvidia analysis, but I'm just doing some back of the napkin math here for you guys. So latest quarter, they did $81.6 billion. That was $48.5 billion of free cash flow. So, if they're going to do a trillion dollars of revenue over the next two years, that means they're going to have roughly $500 billion of free cash flow generated in that time. That will be the most amount of free cash flow ever generated over a 2-year span. That will give Nvidia one of the best balance sheets in the history of corporations, funded by amazing cash flow.

And Nvidia is not a capexheavy company themselves. They get the chips manufactured by TSMC. Super optimized for free cash flow generation. The fab semiconductor model. It's a beautiful free cash flow business model.

They have started building some data centers themselves though to train open models which I think is a good thing because it just helps spur demand in general. But if we look at Nvidia over the trailing 12 months, they've spent $47 billion in share buybacks. I would expect their return to shareholders to go up massively as they continue to generate more and more cash.

Again, my quote at the beginning, short-term, the market's a voting machine. In the long run, it's a weighing machine. Nvidia is becoming a heavy business and their capacity to reward shareholders and their intrinsic value of the company is growing.

And that's why they just raised their dividend payment, which they rarely do. They raised it by 2,400%. And now it actually has a reasonable yield of 0.48%.

All the signs for Nvidia are that growth is set to continue at least for the next year or two. People will have tons of debate of what year three, four, five, and even 10 years from now is going to look like on the demand side for Nvidia.

Right now, I would say the consensus in the market with how things are being valued is that Nvidia, we're close to the peak and there's not going to be a lot of growth from here. That's why the stock price hasn't gone up much in the past year. It's also why it's trading at a very reasonable valuation.

The PE over the trailing 12 months is 31.9. The P forward-looking is 23.12. So if we look at the forward-looking P ratio over the past five years, this is actually historically low for Nvidia. The median is 38.27. So we're in the 15th percentile right now. And let's just say we base it on the trailing 12 months P ratio. The current is 35.7 for adjusted earnings per share. The median over the past 5 years is 53.7. So if we assume that as being the fair value for Nvidia stock, the median multiple, Nvidia is currently trading 34% below that median. If we do it based on forward-looking earnings per share, so next year's earnings, the median has been 38.27 over the past 5 years for Nvidia. This implies an even larger discount for Nvidia stock. And if the company got rerated back up to that 38 multiple, that'd be 65% upside from here.

But the thing you have to remember is that Nvidia is growing very fast still. It's just unclear how much longer it can continue growing. And frankly, I was wrong about this initially. I didn't think Nvidia could even get to where it is here. I was not aware of how big the market could be. And I think that's still might be the fact that people are being early in calling AI bubble. We still have many, many years of growth ahead there. I'm not saying there can't be an overbuild in like 2030 or at some point in the future, but it's not the current conditions on the ground.

And I think people struggle to think in exponentials. We're talking about very high growth rates right here for Nvidia. The current analyst expectation is 92% growth year-over-year in fiscal 2027. And then they're expecting that to go down to 42 and then 20. But really, the 3, four, and 5 year completely changes all the time. People have no idea what to think in those later years.

And given the multiple Nvidia is trading at, which is cheaper than like Coca-Cola and Proctor and Gamble on a forward-looking basis, when its actual fundamentals are growing 80% and the stock price is only up 15% over the past year, I don't think the market is pricing in the scenario where Nvidia grows 80% year-over-year again and then it does that the next year and potentially does that the next year. If that happens or even if it goes 80%, 60%, 50%. That's a ton of growth and Nvidia would look very cheap today based on that. And then the question is, is the company done growing after that? I'll let you decide that one.

So, this is definitely the most interesting earnings report to cover during the week. It will definitely impact a lot of the broader market in the coming months because it has to do with the entire tech and AI trade, which is a large part of the overall market.

Because Wednesday night isn't just Nvidia earnings, it's an earnings report on the entire AI trade. And for me, I do own Nvidia shares. It is a top 10 holding inside of my portfolio.

Nvidia is currently worth roughly $5 trillion. Again, it is the largest company in the S&P 500. And over the past 12 months, shares of Nvidia are up about 20%.

One company's earnings report can now materially affect the major indices simply because of how enormous its weighting is.

However, bringing it back to Nvidia, shares of the AI chip giant have held up rather well during that same time span, with shares of Nvidia up 5%. But shares have been pulling back into earnings, down 8%, which isn't necessarily a bad thing. It's lowering expectations is how I see it.

And that's why Nvidia earnings matter. If Nvidia tells us demand remains extremely strong, capacity remains constrained, customers continue ordering aggressively, and visibility extends deep into 2027, that provides another piece of evidence that this AI investment cycle remains intact and still in the early to mid innings.

And I think that's the story that CEO Jensen Huang is going to tell when the company reports on Wednesday evening. However, if Nvidia begins talking about customers delaying orders or weakening demand, excess capacity, or slowing deployments, well, then we have a very different conversation.

Well, Wall Street, as you can see on your screen here, currently expects approximately revenue of $92 billion. That would be roughly 97% growth year-over-year. Adjusted EPS is expected to come in around $2.09. That's roughly 99% growth year-over-year. And that is center revenues alone is expected to come in at 85.3 billion, representing growth of approximately 100%.

So, revenues, earnings per share, data center revenue, all pretty much doubling year-over-year. Stop and think about that. We're talking about a $5 trillion company potentially growing quarterly revenue and all of that by 100%. That is incredible.

But those numbers aren't the real bar. Here's where Nvidia earnings get interesting. Wall Street already expects Nvidia to beat expectations. That's no secret. Nvidia has beaten revenue estimates in 19 of the past 20 quarters and EPS expectations they've beat in 18 of the past 20 quarters.

So simply reporting 92.1 billion in revenue and $2.09 in EPS isn't necessarily going to make investors happy. In fact, they'll probably be pretty upset. The market wants beat, raise, and strong guidance and confidence that demand remains strong.

That's the unofficial Nvidia earnings standard because if the company comes in and reports $2.10, sure it beats EPS by 1 cent and then revenues of 92.3 billion with no change to forward guidance, expect the stock and the entire sector in fact to fall and potentially fall pretty hard.

That would be a huge jolt that could spark a market-wide sell-off. The first is going to be data center growth. This is the heartbeat of the AI thesis. If data centers continue to dramatically outperform, that's bullish not only for Nvidia, but the entire AI infrastructure ecosystem.

The second item is going to be gross margins. This one isn't getting enough attention in my opinion. Nvidia previously guided towards approximately 75% non-GAAP gross margins for the quarter. That's an incredible margin number. But investors need to watch whether increasing system complexities and component costs, including memory, begin putting pressure on profitability. Analysts have recently flagged modest gross margin pressures as something to watch. If revenues beat, but margins disappoint, the market could focus on margins.

The third item is going to be Q3 guidance and this might be the most important of the entire report. Wall Street currently expects roughly 104 billion for the next quarter, which would be in Q3. If Jensen guides meaningfully above that, that's potentially very bullish. If guidance comes in at 104 or below, you could see the stock sell off even if the Q2 numbers were fantastic because stocks don't trade on what happened yesterday. They trade on what investors think is going to happen tomorrow.

So, no doubt guidance is key, and a strong guidance number is what can ultimately jolt this stock price higher, and not just this stock price, an entire AI sector.

The fourth item that we're paying close attention to is going to be Vera Rubin. This is probably the biggest product discussion I'm watching. Nvidia is transitioning towards its next-gen Vera Rubin platform, and investors want confirmation. Is the rollout on schedule and is customer demand remaining strong? Nvidia recently announced a massive partnership with SK Group involving Vera Rubin infrastructure and next-gen HBM memory with its first planned AI factory coming online in 2027. Rubin matters because Nvidia has to continue giving customers a reason to upgrade. Blackwell, that was incredible, but Nvidia can't stop there. The road map is the moat.

So, the fifth item that I'm paying close attention to is going to be something that could be very new to the company, and that would be China. Nvidia's ability to sell advanced AI products into China has been one of the biggest uncertainties surrounding the company, and the current guidance that we have seen thus far has implied nothing coming from China. The company gets zero in data center revenue from China. Imagine if those floodgates open to a degree, meaning older model chips, and recent reports suggest Chinese customers have begun receiving approvals for H200 purchases, which could potentially reopen an important revenue opportunity for the company. I expect analysts to really push Jensen during the Q&A session. Well, will we get it from them? Well, I don't know. We haven't in the past, but if we do get something, that's another angle that could jolt the stock and the entire sector even higher.

So, now let's take a look at part three, which is going to center around valuation, because this is where the Nvidia conversation really becomes interesting. We're talking about a company doubling their revenue, doubling in earnings, strong margins over 70% huge amounts of free cash flow. Look at this chart here. The start of the pandemic, this company was generating 4.6 billion in free cash flow. That's a lot of cash. In the past 12 months alone, they've generated nearly $120 billion, and that number is only going higher when they report. Insane numbers. Yet, when you look at valuation, it's just absurd. Nvidia trades at a next 12 months earnings multiple of 20 times. And if you go out 18 months from now, we're talking about a 16x multiple. That is far below a market multiple. It's lower than many companies providing single-digit earnings growth, which is wild. Now, compare that against Nvidia's growth, we're talking about a company that's expected to double their quarterly revenue, double their earnings, with earnings still expected to grow 50% all of next year, even. If looking at those figures and we just take the conservative approach and just assume 50% growth, we are still talking about a company with a PEG ratio of less than 0.5. I love stocks with PEGs below one, but we're talking about below 0.5.

So, with that, why is Nvidia so difficult to value for investors? Well, for me, I think the stock is incredibly cheap. It's not that difficult. And investors have a hard time though wrapping their head around the fact that the company continues to put up these growth figures. Growth for this year and the year prior, we're going up against insane comps, and they blew away those expectations. But, if earnings growth does in fact slow, and slow dramatically, well, that's when the multiple can expand very quickly. And, I guess that is the risk that some investors are looking at.

Now, we're going to move on and let you know the reason I chose today's title. Nvidia isn't operating in isolation. It's essentially sitting in the middle of the entire AI supply chain. Taiwan Semi, well, they manufacture advanced chips for the likes of Nvidia, might I add. And then Nvidia and AMD, well, they provide compute.

Nvidia sits right in the middle of all of that. So, when Jensen talks, Wall Street listens for clues about all of it. Because if, and that's a big if, for some reason the demand is faltering, all of these companies are likely coming down. The entire demand story starts with Nvidia.

If Nvidia says AI accelerator demand remains enormous, hyperscalers continue increasing their spending, well, customers can't get enough compute, that's evidence that AMD is competing inside an enormous and expanding market.

So, strong Nvidia demand can actually reinforce my AMD thesis. The danger would be if Jensen starts talking about weakening accelerator demand because then investors aren't debating market share anymore. They're debating whether the entire market is slowing.

Nvidia needs someone to manufacture these advanced chips. Taiwan Semi is one of Nvidia's most critical manufacturing partners. If Nvidia raises expectations for Blackwell and Rubin, that potentially means more advanced wafers, more advanced packaging, more capacity requirements, more demand for Taiwan Semi.

So, strong Nvidia guidance is another data point supporting Taiwan Semi's leading-edge manufacturing outlook.

These companies are more nuanced because Broadcom and Marvell benefit from AI spending, but it also represents one of Nvidia's biggest long-term threats. Why? Custom silicon. Google, Meta, OpenAI, other hyperscalers, these companies increasingly want AI accelerators designed specifically for their own workloads. Broadcom and Marvell help enable those chips. So, if Nvidia says overall AI infrastructure spending continues exploding, that's great for both Broadcom and Marvell. But if Jensen starts discussing competitive pressures from custom accelerators, that's another important positive signal for both of these companies.

This is why I own exposure to both sides of the equation. I own Nvidia, AMD, and then Broadcom and Marvell, but position sizing is different for all of them. Nvidia benefits from general-purpose accelerated computing. Broadcom and Marvell benefits as hyperscalers increasingly design specialized silicon. I don't need to perfectly predict which architecture wins.

If Nvidia continues talking about larger AI factories, rack-scale systems, and enormous clusters, that's bullish for the networking demand. So, Nvidia's commentary could effectively set the tone for both reports.

Nvidia's recent expanded relationship with SK Hynix specifically includes long-term collaboration around next-gen HBM. If Rubin demand remains exceptionally strong, HBM demand should

That includes the likes of AMD, Taiwan Semi, Broadcom, Marvell, Credo, Micron, and several others. And before we jump any further, do me a huge favor, smash that like button down below, and let me know in the comment section, do you already own Nvidia? Are you a shareholder or not?

And with that being said, let's jump right into it with part number one with where things stand currently. Before looking at earnings expectations, we need some context. Think about how absurd that number is.

So, what was once the leaders is now the laggards. I don't think the market is suddenly believing that AI is going away. I think investors are asking a much more important question. How much of this AI boom is already priced in to many of these stock prices? Because there's a huge difference between saying AI is going to change the world and saying AI stocks are attractive investments at today's prices. Those are not the same statement.

Now let's move to the $500 billion question. We've now reached the point where hyperscalers are spending extraordinary amounts of money building AI infrastructure. Whether you're talking about Microsoft, Amazon, Google, Meta, Oracle, AI labs, sovereign governments, everyone wants compute. But Wall Street is increasingly asking, "Well, where is the return in all of this spending?"

Is it possible? Sure, but that's not the angle I'm taking. Now let's move to the next part of this video and look at what Wall Street expects. What are the expectations heading into the numbers when Nvidia reports at the closing bell on Wednesday?

And for those of you that have taken my Investing Accelerator program, which is my 12-week self-paced program, you can find a link to that down in the description below, but those that have taken it have heard me say a number of times, when it comes to stocks reporting, it's less about what you've done for me, meaning the most recent quarterly report, and more about what you're going to do for me, meaning the forward guidance.

Think about our AI investment stack from last weekend's Investors Weekly Playbook video that I put out on Sunday. And if you missed that, go check it out. I'll link it down in the description below, and you'll see it up here up top. But, here's a look at the AI investor stack I introduced in that very video.

Broadcom, Marvell, and Credo, they provide networking and custom silicon. If you want to throw in a company like Micron, they offer memory, and that supply chain is critical to all of this. Cloud companies deploy infrastructure, data centers house it, power companies support it.

So, when Jensen talks, Wall Street listens for clues about all of it. But, what are some of the other companies to have on your watch list that will be impacted by these earnings on Wednesday night?

That's much worse for AMD. And we already heard from them. And they reported strong earnings and strong guidance, setting up the stage for strong demand.

That's why I'll be listening closely to Jensen's discussion around networking.

So, with that, there's really two scenarios that could take place after the report on Wednesday night. Scenario one, Nvidia crushes it. Revenue beats, EPS beats, margins hold, Q3 guidance comes in well above 104 billion. Ruben remains on schedule, and Jensen says demand continues exceeding supply. If we get that combination, I think you could see a significant relief rally across AI, Nvidia, AMD, Taiwan Semi, Broadcom, Marvell, Credo, Micron, potentially even AI infrastructure and data center names because the market gets confirmation the AI spending cycle isn't over. And to be clear, this is the scenario I'm not only hoping for, but also it's my expectation.

But then there's scenario two where Nvidia disappoints. This is where things could get real ugly. And this would be a surprise because it rarely ever happens. And again, Nvidia doesn't necessarily have to report bad numbers for stocks to sell off. Imagine revenues beating, EPS beating, but Q3 guidance disappoints or gross margins compress, Rubin experiences delays, or Jensen starts talking about customers becoming a little more cautious. Suddenly, the conversation changes completely. And given how much money has flowed into AI infrastructure, investors could start taking some risk off across the entire sector. That's why Wednesday could create such a major jolt.

Options markets, well, they're currently pricing approximately a 6% move in Nvidia in either direction around the report on a company worth $5 trillion dollars. 6% represents roughly a $300 billion market value one way or another from one earnings report.

So, there's one final thing I want investors to understand. Even if Nvidia reports phenomenal numbers, the stock could still fall. In fact, Nvidia that share their shares have declined the day following their report over the course of the past four earnings report. Why? Expectations. Everyone knows Nvidia's going to report enormous growth. Everyone knows AI spending has remained strong. Everyone expects another good quarter. The question isn't will Nvidia beat? The question is, can Nvidia beat expectations by enough to exceed the expectations already embedded into the stock and raise the bar in terms of forward guidance? That's a much higher bar. And that's why I wouldn't try to gamble on the immediate post earnings reaction. I'm much more interested in what the report tells me about the next 3 to 5 years for the stock.

So, in conclusion, Nvidia reports Wednesday, and this is probably the most important earnings report remaining in the season. I'm watching 92 plus billion in revenue, $2.09 EPS, hoping for higher, 75% gross margins, and Q3 guidance versus roughly $104 billion. What's the take on rumor demand? Is there any sales heading into China? And most importantly, Jensen Huang's commentary around infrastructure spending. Because if Nvidia tells us demand remains extraordinary, that doesn't just strengthen my Nvidia thesis, it strengthens my conviction across much of the AI investment stack. AMD, Taiwan Semi, Broadcom, Marvell, Credo, Micron, and potentially many more. But if we start seeing cracks, I'm not going to ignore them simply because I'm bullish on AI. That's the entire purpose of following earnings. Our thesis should follow the fundamentals, not the other way around. And Wednesday night, we're going to get one of the most important updates of those fundamentals we receive all year.

Now, I want to hear from you. What happens after Nvidia reports? Do you believe it's going to break out to new all-time highs? Or does the AI trade experience another major sell-off? Let me know down in the comments below. And more importantly, if Nvidia does sell off after earnings, are you buying the dip?

So, if Nvidia says overall AI infrastructure spending continues exploding, that's great for both Broadcom and Marvell. But if Jensen starts discussing competitive pressures from custom accelerators, that's another important positive signal for both of these companies.

Scenario one, Nvidia crushes it. Revenue beats, EPS beats, margins hold, Q3 guidance comes in well above 104 billion. Ruben remains on schedule, and Jensen says demand continues exceeding supply.

We did talk about Nvidia. Nvidia is trading up just a little bit in the pre-market here from yesterday's big decline. There was some news that I caught on Nvidia that I thought was interesting. There's a firm in India that's buying 9,000 uh Vera Rubin chips. So that's the newest chip that Nvidia is going to be debuting I believe later this year. And again there's already a 9,000 chip order and those chips are expensive. So there is demand out there. The question is when we look at what Nvidia says, how much are margins when they report and obviously earnings will be important but the guidance the guidance is going to be the driving factor. What investors want to see is no slowdown at all and they want to see margins still at their or near their peak. If you see anything about a slowdown or margins contracting, it's going to be a huge negative for Nvidia, which by the way is trading at a reasonable PE at this point. I think it's only at like uh uh 15 PE, forward PE or so. So, it's very very reasonable, but again, it's the law of large numbers with Nvidia as well. You know, they can add five billion in revenue and it's a drop in the bucket at this point.

If we look at the charts and you bring up even Nvidia, Nvidia has been a very poor performer and arguably it did make a high back here, but since that high in May, it has really been struggling substantially.

There is a trend line I'm watching very closely which goes back to the April 2025 lows through the pivot low here in March of uh March 30th, 2026 to this pivot low in July 29th. And if we come down into this, will this trend line hold or will Nvidia break to the downside?

Now, looking at the options market, the options market have an implied move on earnings of about 6% one way or the other. So, that gives us a little general idea of what the option market expects in terms of a move off earnings from Nvidia for Nvidia.

There was some news that I caught on Nvidia that I thought was interesting. There's a firm in India that's buying 9,000 uh Vera Rubin chips. So that's the newest chip that Nvidia is going to be debuting I believe later this year. And again there's already a 9,000 chip order and those chips are expensive. So there is demand out there.

The question is when we look at what Nvidia says, how much are margins when they report and obviously earnings will be important but the guidance the guidance is going to be the driving factor. What investors want to see is no slowdown at all and they want to see margins still at their or near their peak. If you see anything about a slowdown or margins contracting, it's going to be a huge negative for Nvidia, which by the way is trading at a reasonable PE at this point. I think it's only at like uh uh 15 PE, forward PE or so. So, it's very very reasonable, but again, it's the law of large numbers with Nvidia as well. You know, they can add five billion in revenue and it's a drop in the bucket at this point.

And it'll be the day before Nvidia, so we'll check in on them as well. Remember, it's not that big of a market mover. It never has not been recently.

So, we're going to get a lot of content and a lot of things coming out from Nvidia. with Nvidia earnings coming up, I'm actually pretty happy that the stock is not as excited, right? We were over $220 about a week ago. Now, we're at 208.

I'm a long-term investor of Nvidia, right? This is the company that has completely changed my life. Now, I don't know when Nvidia's going to have that quarter that's going to completely flip the switch and make it exciting again for the AI space.

And the reason I'm saying that is because year-to-date, the stock is pretty much flat, right? Up 10%. That's nothing, especially if you are the AI leader.

I don't know if it's this quarter or next quarter or the quarter after that, but one of these quarters, the market is going to wake up and say, "I was wrong with Nvidia and this, even though it is the biggest company at a $5 trillion, it can actually be bigger."

Now, in the intro of this episode, I mentioned that there's a market Nvidia is pretty much saying competition is dead and competition is so dead that that you need to come back to Nvidia. And that competition is the XPU market.

But Nvidia says that chip design is hard. And if you're building an XPU, you are wasting your time and this is a market that's dead unless unless you enter the Nvidia ecosystem.

Now, the reason that they mentioned this is because back then, right back then in the chip design process or in the kind of just the compute market, all you really had to do was sell a chip. And if you sold a chip, you got revenue, you got collected and that chip worked well.

Now, in this AI market that's completely changing because you're building factories, right? You're building full factories. And when you're building full factories, you need to be able to design you're no longer designing a chip, you're designing a system.

And most companies are still just designing a chip and that's not even half the battle, right? Probably not even a third of the battle because if you're building we're building these gigawatt data centers now, right?

And historically, a gigawatt data center, we can see it's collecting roughly $50 of revenue per year. Give or take plus uh plus or minus $10 billion. Regardless, these massive data centers are crucial and the number of revenues that they're collecting is extremely high.

So, if you are building an AI chip, you need to make sure that you're able to build an AI data center. And that's no longer just a chip itself.

You need to include the scale up, the scale out, all the networking types of solutions. You need to be able to understand how you can racks at your rack scale architecture production factory of software. You need to have kind of this crazy supply ecosystem that allows you to not just say, "Hey, I'm just building a chip, but I'm building this whole AI data center."

Because money is time or time is money, right? NVIDIA you can call them up and say, "Listen, I need a 1 kW data center." NVIDIA says, "No problem. I know how to handle the cooling. This is the blueprint that I have to do all this. You'll be able to set up your data center in no time. You know these chips are great, and you're going to be collecting revenue in no time."

If you're doing an ASIC with somebody internally, had they thought about all that already, right? And if they haven't, what is going to be the timeline of finally getting all that ready?

And then what if it doesn't work? You missed out on tens of billions of dollars on revenue on a 1 kW data center because you were trying to save a few margins on an ASIC solution.

Uh so, NVIDIA is saying, "If you're in this ASIC space, it's hard work. But, if you really want to be in this ASIC space, come with me, right? I'll help you out. I have this thing called NVLink Fusion, which allows you to bring your ASIC into the NVIDIA ecosystem."

So, while maybe you're no longer a 100% of my customer, you're still 60, 70% of my customer, and we both win because you feel like, 'Hey, you don't aren't not buying all my chips.' And I win because I'm still growing and growing with you as well.

Now, NVLink Fusion allows, like I mentioned, the connection of these XPUs to NVIDIA's infrastructure. And NVIDIA actually gives a lot of reasons why this is pretty important.

And like you kind of connect with an NVIDIA's ecosystem, you're connecting with their scale up, their scale out, their CPUs, and the list goes on. But, but the main reason NVIDIA does this is because or they're offering this and they're bringing it up is they're saying that this is a great way to manage risk, right?

A data center locked to one chip can become a schedule risk. And like I was talking about, imagine you're building a data center that only is meant for one type of ASIC chip that you are designing.

And what happens if there's a hiccup? That data center is not making any revenue. But, what if you are part of the N D I N D link fusion program? A customer can deploy direct level solution with N D I GPUs to start off with, right?

We're going to build we're building this data center my AI chip is not ready yet. Let me build with N D I's GPU and the whole ecosystem.

And then, they can decouple the development of their XPUs and put it at a different pace. So, maybe month one, month two, month three, as you're building this data center and you're putting things on, you're going to start with N D link and N D I's main solution.

But, now five months have happened and you've actually feel more comfortable with your AI chip and you are able to put it into this N D I ecosystem where now these two servers can live together.

So, it's the best of both worlds because you're able to start collecting revenue as soon as possible because N D I is the go at designing chips. But, at the same time, you're also going to be able to make your shareholders happy.

You may be able to increase your margins because as times progress, you're going to be able to start adding your own ASICs into these solutions as well, but still be part of the whole ecosystem.

And I think that is extremely extremely extremely bullish. Now, there are a few partners that I think can benefit from this. N D I has actually given us the N D I link fusion list of the stocks of the companies that are working.

In the CPU side, you have ARM, you have Intel, you have Fujitsu, you have Sci-Fi.

Now, one thing that I found pretty interesting was Qualcomm used to be a part of this list. And I don't know where they went. I don't know if that's bullish or bearish, but they used to be part of this list.

Now, in the custom silicon, you have Astera Labs, you have Alchip, you have Samsung, you have MediaTek, and you have Marvell. In the technology partners, you have Cadence and Synopsys. These are the ones that help you design these.

In the optical partners, you have Aear Labs, Light Matter, and Marvell. So, you can kind of already guess what are the markets that I believe can benefit from these, and it's going to be the parts within these ecosystems.

And the one one of the players I believe is very high likely of winning is going to be Marvell. It's not a guarantee, but very high likely. The reason I bring that is because in March 31st of 2026, Nvidia brought Marvell they they joined forces in the in the kind of expanded partnership.

And one of the partnerships was the optic side. The other things was the NVLink fusion. Nvidia ended up investing $2 billion in Marvell. Around the same time frame, we did hear that AWS, who is building their own chips to train them for is actually going to be using Nvidia's NVLink fusion.

And the one that designs and helps AWS with the training for is, you guessed it, Marvell.

Right, Marvell is the one winning here. And and now Nvidia I I believe this is really good for AWS understood. AWS understood that if we're going to have to be if we're going to be able to be the a good performer in our AI chip, good total cost of ownership, we have to join with Nvidia.

So, it's interesting. But then you have already announcements like this where Nvidia is actually still keeping that customer that designing their own AI chip, and they're going to be able to license a lot of products.

The other one that I thought was pretty interesting is Marvell and Google announced a partnership. Now, Marvell Marvell is not going to help Google design their own TPUs, but they're going to help specific AI accelerators.

Is this kind of the entry-level point for Marvell to somehow sleep slip the NVLink into Google? Because at the moment, there is a company that's very popular in the ASIC space that does not is not part of the NVLink fusion space, and that's Broadcom.

And that's the Broadcom company that I'm a little bit I want to say worried about, but I think they need a change certain things and we'll talk a little bit about there.

Microsoft is also working with them for the Maya 300, right? So, Marvell does have a lot of I wonder if we're going to get an announcement where one of the upcoming Mayas from Microsoft's A6 might end up being part of the NVLink Fusion world.

So, few stocks that I believe can benefit. First, Marvell, right? I think NVLink Fusion is going to be big, especially as we're getting into these sizes and this would be a good kind of revenue driver for Marvell.

Not only because it's going to be a partner for the NVLink, but most likely if you're trying to work on certain networking solutions or certain optic solution, Marvell is going to be that partner because they're working with like, "Hey, you're already doing the NVLink Fusion with us. Let us also work with you on some of the networking and optic solution as well."

So, Marvell is the first one. The next one is Astera Labs. This is another partner from their kind of list of of custom silicon. You have Astera Labs here and I do believe they can benefit dramatically.

Again, with all these types of approaches and Astera Labs right now down roughly 42 from its 52-week high or a year-to-date high, Astera Labs is another one that I believe can benefit in the long term of things.

Next the two are the kind of hyperscalers that are already working with Marvell. Amazon is the first and Microsoft is the second. Now, I'm a little hesitant on buying these because we did buy them a lot cheaper when we were kind of in the 220s below, we bought Amazon in the sub 400s, we were buying Microsoft.

So, bullish long-term, maybe not right now for individual stocks unless I'm dollar cost averaging.

Marvell and Astera Labs, I don't mind being a little bit more aggressive on these high beta names because they have pulled down a nice amount from their 52-week high and I am very bullish.

Again, this doesn't mean they're going to be winners. The market can always shift, but based on the data points that I have right now, I would be excited on these two stocks at the moment.

Now, like I mentioned, there is one stock that I I think needs to change how they're working with and Broadcom does not is not part of the NVLink Fusion.

Right, they're not part of the NVLink Fusion.

But, they have a huge customer base, right? They have Anthropic, they have OpenAI, they have Meta, they have Google and they have other customers.

What if those other customers start to leave Broadcom or start to work with other players because they're so hard-headed and don't want to be part of the NVLink Fusion world.

I For example, if I was OpenAI, I was if I was Nvidia, I'll go to OpenAI, I'll be like, "Are you sure jalapeno's going to do good? Are you sure you guys have the supply chain unlocked? Are you sure about this, that, and this, and that? Wouldn't it make sense for you to go talk to Broadcom to make sure you are still part of the ecosystem because you're going to make sure that everything's right, right? We have all this worked out."

So, if I'm a company like OpenAI and Anthropic, right? These are two players that unfortunately don't have the cash flow like the hyperscalers.

So, if a chip fails and it chip fails really bad, especially if they're building something at massive scale, that is going to be a huge hit for them.

For Google, if one of their AI chips fails, nothing bad happens, right? I mean, I think they they'll take a hit and maybe their stock might be punished, but it's not end game over.

If Open AI and their optics spent 500 MW on a chip that didn't scale up properly, that is bad news. And I'm pretty sure Nvidia is letting it be known to those players.

And I wonder if Broadcom is going to eventually be a part of the NVLink Fusion domain or partnership. If actually it does, I would actually be more bullish on Broadcom than I am right now.

Now, I'm not bearish Broadcom, and I don't think it's game over completely for them, but it is something that investors should keep a closer eye on. And I would rather right now focus on companies that are part of the NVLink Fusion partnership list.

So, I hope you guys enjoyed today's episode. Take care, have a good day, and see you all next time.

That's why SpaceX is then able to pay 30 billion to Nvidia for chips for the buildout. That's why Nvidia is able to But the reality is Elon's probably spending, you know, 50 for the entire data center, right? We forgot that there are other components to the data center. It's not just the Nvidia GPUs, it's the CPUs, it's the memory, it's uh the flash storage, the hard drives, uh it's the server racks, it's the building or whatever else. You know, 30 billion might very well just be the GPUs, but realistically, that's probably going up to closer to $40 billion on the latest generations of GPUs.

Well, then you got all the other crap that goes into the data center as well. And this hundred billion dollar figure is not validated. So, let's be clear, like he sounds really good, and I I respect Zach for what he's done and what he does. All this sounds really good, but see how quickly we gloss over these things. I mean, he says three words. I got to freaking pause it and explain.

Token generation, when you look at the Google IO event, has already turned negative on the second derivative. Now, part of that is law of large numbers. But if you're already going negative on the second derivative already, and we're supposedly early, then you're no longer exponential. You can't have a negative second derivative and call it exponential. What is a negative second derivative? It's a really positive second derivative is the rate of change is getting faster. when the second derivative, so the rate of the growth, how much the growth rate is changing, turns negative. It means you're still growing. So the first derivative is positive, but you're growing slower. So it's still going up, it's just going up in a more flat way, [laughter] right? Uh so you can't you can't just say, "Oh, exponentially."

Hey, you're protected against that. But Jake, could I could I shift gears for a second to tell you what's so brilliant about what Jensen did here? There's there's one part. There's one other part in return for that.

They get a revenue share above that floor. Who who gets the revenue share above the floor? Yeah. So they I think the way I think the deals will work is say we're willing to guarantee offtake at you know nowhere near $30 or $50 a watt. We're willing to guarantee

that the TAM is getting constrained by the ability to finance this build out. And what he's doing is alleviating that finance constraint so that he can grow uh as big as the the TAM actually is, right? It's removing that constraint. So for just to take one example, Elon wants to add somewhere around 6 to 8 gawatt next year. We know that that would cost3 to400 billion dollar of capex.

Remember what I wrote on my SpaceX documents? On the SpaceX documents when the earnings came out, I said on the filing, they're going to have to come up with about $414 billion of extra debt. And their preference is debt over stock issuance, which is good and bad. Uh there you go. It's the same number. It's no secret. We're all seeing the same stuff. It's just how we essentially you make draw conclusions from that information. Like that's a lot of debt that still needs to be sucked out of somewhere. And we saw last time we had a suckening. It was June. It led to the Leo fold in July. All right. The company just raised a hundred billion in its equity and debt offering. So obviously they would have to go out and finance that somehow. And as we talked about on our previous episode, the simplest way to finance it would be to get seller financing from Nvidia, especially given that the payback period could be as quick as one year. Could be. Could be.

See a lot of the assumptions that have been layered in here. Okay. Politics will stop overbuilding. Token growth is exponential and the payback period could be as low as one year. That those are huge assumptions. and you know the payback period is related to that $30 to50 billion of of revenue that Elon wants to pick up from compute maybe it'll happen but look at the assumptions that are built in and when you see the assumptions then you can kind of make the determination like okay we getting closer to uh you know the top doodla they actually well sachs comes up with a very similar thing that I've been harping on for a while in terms of the potential top uh and how to measure it and it comes after 3 minutes of yapping from Gavin and the others. And it is right at 109. I'll play from about here. Yes.

No, they're not. Okay. Again, this is where David Saxs reads the crap on X and I don't think he's putting his critical thinking hat on. I think he's scrolling on X too much. They're not making hundred billion of ARR right now. Annualizing out the last day of July, they're at 65 supposedly. Okay, we don't know if that's just because of the co-work explosion. We don't know uh if that's going to last, if that's going to get whittleled away by openweight models. We don't know. But David Saxs here just literally told you they're making a hundred billion. That's enough to pay Elon's 30 to 50. But they're not. They're making on one or two gigawatts of compute, they're making about 6465. Now, what's the problem with that? Well, the problem with that is they have massive compute coming online. What I also wrote over here that I didn't talk about yet is they have 5 gawatt coming from Amazon and another 5 gawatt coming from Google and Broadcom. Now, not all of those 10 gawatt will get built. But even if half of this gets built, that's another 5 gawatt, which is 2 and a half to five times as much power as they have now. So, is their revenue going to be able to scale to keep making what they're making now this many more times over? Well, according to David Sachs, demand is exponential. So, in that case, sure, why not? But that's not what we're seeing. We're not seeing exponential demand. We're seeing declining second derivatives on token usage. There's still really big numbers. [laughter] But that's a concern.

Okay. If compute, that's why they're able to pay SpaceX say 50 billion a gigawatt. No, they're losing money. They're not able to. They're losing money.

You know, a spot price for compute. the buildout is going to be way more than and this is the other problem. THE NUMBERS COMPRESS SO MUCH MORE THAN WHAT David Sachs is saying and and is fine. Maybe maybe the exponential ramp will come but it's not 30 billion that you're putting into the buildout. That's the problem. You know, he's telling you, oh, they're making a hundred and they're going to spend 50 to Elon and Elon's going to give 30 to Jensen and everybody's going to win.

That's why Nvidia is able to

Nvidia's gone down four out of four of their last earnings after earnings despite beating expectations. It's not Kevin Worsh. It's not Nvidia earnings.

So Nvidia just raised prices for their server chips by about 15%. And it's leading to expectations of oh my gosh this is going to lead to bigger margins at Nvidia. My expectation is a lot of those price increases are solely because of memory, which that makes people really bullish on memory, that higher prices can sustain for longer.

Now could that actually benefit Nvidia? That's another argument. I could make the argument that Nvidia could actually benefit from memory stocks coming down because they could keep uh GPU prices elevated but actually uh make more money because they're spending less on cost of goods sold for the memory supply. It's possible.

Uh yeah, I mean this is a hardware index. So I've got uh Nvidia, TSM, Broadcom, Micron.

you could end up getting Nvidia pull a massive rally here post earnings and and you know, once you get through Kevin Worsh and whatever, uh it ends up uh you know, skyrocketing again, you have your next hardware 2.0 rally. When Nvidia goes down 33%, it's going to come back.

But if it booms on, let's say, Nvidia earnings, it lowers the floor and you can kind of trail with it.

If Nvidia goes down, even if Nvidia goes down 80%. Just to be ridiculously extreme, it's not going to zero.

Obviously any company could go to zero, but you know I would say uh 10,000 3x leveraged ETFs will go to zero before a company like Nvidia goes to zero, right?

So that way Nvidia earnings come in and the entire semiconductor index rebounds because we get over some of the Kevin Worsh or Scott Besson concerns or Iran D-Day concerns and this puppy rebounds like it deserves to.

Best case scenario, you're allocated in such a way it doesn't matter if Nvidia goes down 80% because you look at it and go, "Whatever, it's on sale. I'll buy a little more. I know it's not going to zero." Nvidia is a great company.

Wait, I want to hear this. for Nvidia's chips. Full circle. Nvidia doesn't need to profit on the models themselves. It profits on everything that runs on them. The Grock chip going live today is the hardware half of that same bet. The speed layer for the open models Nvidia now wants to build itself. For all that though, look at the share price down about 2.5%. Stock just can't catch a bit. It's fallen wet for the seven straight days, its longest losing streak since September 2022. Down almost 7% over that stretch. As Missou's Jordan Klein put it in a note this morning, buyers are in just no rush ahead of Wednesday's night and Wednesday night's results. Why try to be a hero ahead of that print?

Uh, well, speaking of it, I mean, you're going to be talking a lot more about it between now and then, Christina, I have a feeling. But do we have any key expectations there in terms of of of what we'll all be looking at with Nvidia? Assuming as usual, they report a very strong number,

which is, see, you use the word assuming, right? The market already expects Nvidia to beat to provide these strong numbers. The market already expects the CEO to be extremely bullish saying that uh you know everybody's

which we spoke about just in the last hour and also why you saw uh price hikes you know reports this weekend saying they're going to increase prices for their systems anywhere between 15 and 17%. And of course the circular financing narrative that came up with that $500 billion uh initiative with those six finance years. You know I think the narrative that Jensen Wong and management will have to be on the call is just it's not about circular financing. This is us deploying our money in smart ways. We have a lot of money on the in terms of free cash flow and that is why we are very aggressive right now and acting. Keep in mind Nvidia concentrates their risk by uh accelerating their investments into these smaller companies. It's great when it goes up, but when it goes down, it's an accelerated risk. Now you have fewer orders and losses on your books.

Know what Nvidia built into the prior pricing. So, there's a lot we don't know at this point. Recall that that Jen Sinwang has said repeatedly that they

straight days, its longest losing streak since September 2022. Down almost 7% over that stretch. As Missou's Jordan Klein put it in a note this morning, buyers are in just no rush ahead of Wednesday's night and Wednesday night's results. Why try to be a hero ahead of that print?

But do we have any key expectations there in terms of of of what we'll all be looking at with Nvidia? Assuming as usual, they report a very strong number,

it shifts the risk. It shifts it off of Nvidia's balance sheet elsewhere. And that is most appropriate. I mean, not really because if Nvidia is taking stakes, they don't actually shift the risk at all. They actually amplify their risk. So the guy's wrong.

Generating roughly a billion dollars of free cash flow every two days. they have a responsibility to put the cash on their balance sheet to work. And if they're big believers in AI, like we are, and I'm sure they are, why not invest in the AI opportunities? So, sort of the classic eating your own dog food. I think the optics of circular financing, that is putting money into your customers to buy your product, it's not a great look. I I totally get that. But Nvidia has to put its cash to work. I think what would soften the blow for

You literally do not need to put your cash to work. You could literally sit on it. Damn. In buying back its own stock

and you expect them to do that. I do. I want to talk about the distinction between depreciation and the economic lifetime of a GPU because right now what I see is lots of Nvidia executives posting on social media about Vera Ruben being in full production at the same time like that KKR Black Rockck deal. They're trying to sell that this is a highly utilized cash flow generating asset for many years to come on older obsolete generations. How can both of those things be true at the same time?

I think they can be true because of the way the market evolves that the most leading edge applications, the most demanding will require the newest generation of accelerator and in order to be efficient, in order to have the best profitability, one wants to utilize the most leading edge tool because it will be more efficient. Basically, fewer tokens pulled. Uh however, we've seen that the older generations, the ampers, the hoppers are still being used. And so what what we're observing is the market's broadening in that the most leading edge applications, the most demanding applications will require the newest generation, but the older uh GPUs, the older accelerators have useful lives that could extend many years.

Simon, many of your industry colleagues, including our our team at Bloomberg Intelligence, expect to hear something about China. Does it matter to you if Nvidia can sell some kind of GPU into the Chinese market?

No, honestly, not not really because Yeah. No, it doesn't matter. That's when I went heavy on Nvidia and then I launched an ETF around pricing power chips.

But I want to look at the pricing for Google really quickly. I just want to look at Google's debt and see what that's trading for on the secondary. Okay, cuz they raised $84.75 billion in and that was the upsized round in June. That was basically right at the SpaceX suckening pulling all that uh money out of the market. That's where they raised some good money. Basically timed it perfectly at the te peak.

Gus is asking biggest question is what pops the bubble. I have my eyes on anthropic. uh they they are very interesting uh because if their profitability doesn't accelerate uh then then we have a problem on on what we're underwriting a lot of data centers for, you know, especially under the Elon hype of, oh yeah, we're going to be able to charge $30 to $50 long term for these compute prices. Really? Are you really going to be able to

Okay. So, uh, let's go. Let's take a peek here. Okay. Where is I want to see what Google's trading for now. The Q's Q's are still slipping a little bit here. We're sitting at what are we at? We've got Yeah. down a buck four or 14 right now. Hood's green. Amazon's green. Palunteer's only down 169. Now I've got Bitcoin's almost at 80K. That's awesome. Okay. So, okay. Got it. All right. Now, what about the googs? And then I want to go hit Nvidia.

Okay, so this is up 14 now. It's still a little bit under where Birkshire acquired, which I want to say was 359 with some of the discounts they got cuz they, you know, that was announced somewhere around like the 370s range.

Um, okay. Let's go take a peek at that. And then we have a couple of those journal articles to look at. So Google's price Google after their suckening that was their sort of top because they issued at the end of May. It was like May 29th. Yeah, that was their final. Yeah, they sort of topped out on the 22nd. So I wonder how much writing was on the wall for Okay, the Google preferred. Here we go. So, the Google preferred was two $9.6 billion $50 par value, 6.25% mandatory conver uh convertibles. Oh, wow. They actually have a premium conversion price. Uh so they convert at 440 bucks. 444 to 440. I see. Okay. Either way you get converted though, it just depends on what ratio you get. So it gives some downside protection basically. That's a little complicated, but that is very interesting. Uh, that's an interesting way to raise money because Google's raising at 6.25.

SpaceX coreweave we know is somewhere at uh over 10%. And that we could see with the Wall Street Journal. Coreweave 10%. The financing. Yeah, that's this Cororeweave raises debt. And this is another thing that's interesting about debt right now is Coreweave has a junk rating. They're raising at over 10%. But they don't actually disclose that they're raising at 10%.

Okay. Let me pull up coreweave and then we'll be able to see that. Okay, so coreweave corewave coreweave corewave coreweave they have earnings report. Here we go. August 11th. So this is the coreweave earnings set. And if we scroll down to I should have bookmarked it, but somewhere here we're going to see their weighted average cost. And I want to say it's 9%. H I don't know where it is. Weighted average. No. 9%. Nope. 9%. Dang it. Let me see if I can find it another way. We'll find it. But I think this will be very important for seeing just how impactful all these the these uh rates are that are coming out. So Amazon, for example, I'm just looking at some of their bonds. Amazon issued bonds for 4.25. Those are a five-year bond, 4.25%. 25%. Their 30-year bonds are at 5.8. Oracle's raising money on a three-year at 4.5 and on a 10ear at 5.7. SpaceX is raising on five years at 5.35. So SpaceX is raising for more. Okay. I want to see some more of these. I want to see like coreweave. Yeah. Okay. Here's coreweave. So coreweave. Here's a fiveyear at 9.75 and then they sold it for a discount. So that's how you got over 10. That's interesting. XAI was issuing debt at 12%. So they actually got a big discount by merging or getting acquired by SpaceX. They reduced their cost of capital substantially. It's actually kind of smart. Meta on the five years raising at 4.5, but they're also raising 40-year debt at 6.4. 4. All of those become more expensive unfortunately as rates go up. So that also hurts the uh what's it called? Uh the outlook for these hyperscalers and spending Besson's meeting or press conferences in like two and a half hours. We're still bond yields bond yields right now. Fortunately, they're still all in the red in anticipation of Bessence call. So, that'll be interesting.

Oh, let's see them talk about the foldable phone here. Uh, is foldable phones, but it's actually double that in China, right? So, the vast majority of foldable phones in use in the world are in China. So,

Nvidia just posted this. Nvidia's Grog 3 now in full production with worldclass speed for agentic AI. This is obviously direct competition to Cerebrus.

Nvidia today announced that Grog 3 LPX is the interactive AI inference accelerator. That's a fancy way to say our cerebrous competitor. It's now in full production. An extension of the Vera Rub platform. The Gro 3 LPX delivers a major boost in AI inference by enabling ultra fast token generation for highly responsive agentic systems.

Okay, let's go look at the Nvidia deals. Nvidia's Grog 3 now in full production with worldclass speed for agentic AI. This is obviously direct competition to Cerebrus.

That's a fancy way to say our cerebrous competitor. It's now in full production. The Gro 3 LPX delivers a major boost in AI inference by enabling ultra fast token generation for highly responsive agentic systems.

Now coming up, we're going to get back to Nvidia as investors prepare for the Super Bowl of earnings season. Stock down 2% down for a seventh straight day. We'll have the details.

Nvidia earnings coming up. Nvidia earnings on deck Wednesday

post Nvidia earnings especially with Nvidia earnings on deck Wednesday

Nvidia competes with their growth product which they just announced is in full production. And so, you know, then you kind of get stuck and people just use the Nvidia product and the Cerebrus product doesn't go anywhere, especially if Nvidia can scale faster.

And then we still have the Nvidia announcement to look at because they just announced that. I think it's kind of strategic that why they're doing this right away, right before uh oh, with Nebius, the first to adopt.

let's now go on to Nvidia for a moment.

In Nvidia discusses perplexity investment at $30 billion valuation. Wow. Nvidia and Salesforce in the spotlight this week. Oh, good old earnings. Good old earnings.

While Nvidia costs are going up, memory chip prices are rising, raising the prices to maintain. Right. Right. Salesforce will be the same day as uh as Nvidia. That'll be very interesting.

Wall Street is counting on Nvidia to keep the party growing. Going growing and growing. Okay, let's go get let's try.

Elon Musk says SpaceX's first AI satellites provide powered by Nvidia will initially launch in the fourth quarter of next year and hit significant scale in 2028. This is not a surprise that this is just, you know, 2 days before the Nvidia earnings, you know, it's sort of like the Gro announcement.

SpaceX in partnership with Nvidia has designed a space optimized Vera uh Revenuven uh Envy Link 72 system for launch to orbit in Q4 of next year with significant scale in 2028.

That sounds good. And then of course, uh, you've got Nvidia replying with a little clap emoji. So, it's like, hey, Elon's got to pump it up before uh before uh before earnings. I I think that the space compute thing is, you know, a little ahead of schedule and it it's being hyped a little ahead of schedule. Put it that way.

Micro processor made by Nvidia powering Moscow's AI drones. That's not good.

I think at the end of the day, as long as earnings are hanging in there, and they are, and I think this week we have Nvidia on Wednesday and then we have the Fed speaking on Friday. I think Nvidia is going to be the bigger story. if they can prove that story continues and especially what they're saying memory chip and how that's affecting their demand and their customers affordability. I think that's going to be a lot more important than what the Fed says or doesn't say because people want clarity.

Do you agree though that Nvidia is the the biggest story this week? Because you don't expect

And I'll just add real quick on on the Nvidia side of things as as Courtney and Steph just both both argued and I would agree, everyone expects them to beat raised that the the quality of the of the earnings report in that sense is not really a question. The issue was going to be around as as Steph alluded to the what we'll call the circular financing and the the funding of the AI infrastructure.

I imagine what she's going to say is the company's position is we have all this money. We generate a ton of money and we are seeding our future and the industry's future for the next however many years.

And look, the stock's been on a six-day losing streak. It's down again. And that was after a nice move higher. So, it maybe complicates the story for investors a bit going into Wednesday because investors right now are just hesitant what's going to be that next major catalyst to move the stock, which is maybe why they're not getting in.

Uh Dan mentioned it, the numbers at this point really aren't the story. Beat assumed across the board. We know that demand surging, lead times have stretched from 6 weeks a quarter ago to nearly 40 now. Wall Street actually wants is specifically in the guide, the first real read on how fast the new Reuben chips are ramping the next iteration.

Then there's also the margins. Nvidia told customers it's raising AI server prices as much as 17%. It's reported that uh driven by soaring memory costs. The question is whether that protects its mid70s gross margins. You can see on your screen which is estimated to decline in Q3 maybe even more or if memory still eats into it.

Watch too also for the Vera CPU a new multi-billion dollar line as AI agents take off. And then to your last point all of you on the panel hanging over it is the circular financing debate. Nvidia is bankrolling the same customers who buys these chips. Like it's up to $15 billion backs stop for an OpenAI data center in Ohio. That's just one example.

Speaking of our guy Jensen

drained by the ability to finance this build out and what he's doing is alleviating that finance constraint so that he can grow uh as big as the the TAM actually is right as talked about in our previous I think Wall Street banks and private equity firms providing financing based on the expected cash flows that will be delivered from these GPUs. In some ways, Nvidia is kind of becoming the central bank of AI.

Yeah, we've made that reference as well that they're that Jensen's powered, right? You have asset back financing and Did I shift gears for a second to tell you what's so brilliant about what Jensen did here? That's a Steve Jobs line.

Unlike the activity of animals which are observed, outputs of LLMs are not a reliable guide of what might be going on underneath. This is because the training of models requires feeding to them trillions of words of human produced language.

These contain countless accounts of consciousness and how humans convey feelings to each other. Right. The sum total of this training is I don't know what the heck that word is. That's crazy. I think it like to me my my takeaway generally in this that LLMs are really good at replicating us because they're trained on us.

Okay, so let's see here. LLMs over time, what do they say? Biological traits they score lower. All right. Not so useful. What is this? Don't care about the seafood shack.

Local business group sues mom Donnie over plan to open municipal uh grocery stores. That's interesting.

All right. Let's see.

Excuse me. Oh, yeah. Okay. Here we go. local business group sued Manny Coalition blah blah blah will put us out of business basically. Well, yeah, I mean that makes sense. The government's basically going to loss lead this escalated US Canada threatened diesel prices man getting a little boring.

What's going on? Okay, where's the entertainment? Go look at them sticks again. We'll have to start talking about mold again. Still down 80 bips on the cues. Okay, so Bloomberg, what is this?

Musk sees SpaceX's orbital data center launch near end of 27. I mean, his time frames are always a little funny, but I'd like to see what they're talking about.

Broadcom credit risk soarses on mega debt financing. Yeah, it's always debt financing. Uh, let's see here. Okay, this is interesting. Decreases in home values due to general economic conditions can lead to an increase in meritus and non-mmeritus construction defect claims.

Interesting. I mean, that's not a surprise. And I also got some trends here on the warranty reserves and how they've sort of been moving. And pop those up really quick. So, okay, let's throw this up on screen and then we'll go to the SpaceX one.

So, that was here housing. It gonna work. Oh, not yet. Come on. There we go. Wall Street Journal on construction defects. Okay, perfect. Now I can put the picture in.

Okay, perfect. So, basically, Wall Street Journal has a story on increasing construction defects related to mold. Uh I pulled some data and found uh warranty reserves and uh acred liabilities including litigation reserves litigation reserves uh rising.

Uh there is there is a link between uh values going down and uh more lawsuits. And then let's just throw in my my thesis. Why is this happening? Okay.

Okay, then I want to see the SpaceX store and then we can always summarize the housing one. Cool. Just put my notes somewhere. Okay. So, what's going on here with SpaceX?

So this is the Groke announcement and then Musk, you know, wouldn't shock me if Jensen is like, "Hey man, can you give us a shout out before earnings?"

Public approval of Iran war falls to 31% and Trump's approval sits at 33%. Oh, that's interesting. Pretty low. So, let's go to let's just go see what else is going on out there.

Times Trump the tariffs. We saw this. Go see what some of the institutions are saying in a moment, too.

Okay, see what we got over here. >> Very viciously. >> Is there a level in the 30-year yield that would cause the Fed to move in? Because you're saying the market's in control. How does the Fed react then?

>> Well, first I think the 30-year while it's valuable for a lot of long-term investors to model their portfolios, the 10-year is the really the most important rate to look at. That's if you are a CFO today, that's what you're looking to model your business projections. If you are an investor, that's what you should be looking for. So the 30 year, I think, is is not the right rate. I would argue the 10 year is the right rate to look at to figure out where the economy is really going in the longer term. When you think about kind of what these long-end rates are saying, what are bond investors kind of telling everyone about inflation and about the macro environment right now? >> So, first of all, bonds investors should be really satisfied today.

>> They have been chasing spread. They've been really effectively trying to figure out if spreads widen or contract to earn their their outcome or yield. And the rates resetting give them some comfort. Now the downside of that for bond investors is they have very little to work with. Now yields are so com the spread is so compressed that if you pick the wrong credit you will have losses on the credit side and that makes it very tricky to be a bond investor right now. So the end investor is much better off. The fixed income managers have to be smarter in this environment to pick the right credit. So it's a different different situation now in the marketplace.

>> Does that create maybe a little bit of stasis? I mean, we had the head of PGM's business, fixed income business on last week and he said something similar. He got a quip that, you know, he's a credit investor who hates being a credit investor right now.

>> Yes. Yeah. >> You are so right. And then I would I would expand that Roma to say the credit market is not just public fixed income. There's a very very large and growing private credit market. And so from my standpoint, I would look at credit across both public and privates. And I know for for the longest time people have to choose between one or the other. I would say this is not the time to choose between one or the other. You pick the liquidity you want, you pick the risk level you want, and then you can you can find value everywhere.

>> Well, can I push back on that because I mean, let's just fold in all of this debt issuance that we're seeing surrounding the AI trade because that definitely seems to be competing to a certain extent uh with uh basically any other sort of uh fixed income asset right now. uh is that something that will sort of uh rightsize itself in due time or is this going to be a longerterm story where as long as these issuers are I mean they're coming to market with with with issues that are bigger than some of these treasury auctions.

>> You you are absolutely right. We are seeing a complete rerating of how large our credit markets can get. Um, our public credit markets have always been large, but the private credit market got very large and now you see a lot of crossover credit happening. You could argue where it's going to be sourced from. At some point, you get to a natural capacity limit of how much credit could be issued, but right now they're all rerating. And they're rerating because you have these borrowers who have never borrowed before. So, they are not leveraged. They're not under stress. And that's allowing the market to expand today. They will hit their natural limits and like any credit cycle you will see stress but that is not for today.

>> Talk about those limits though and kind of when principal expects to hit exposure limits to single names to names that haven't been raising and now suddenly are raising billions.

>> Our view is in certain parts of the economy you are at that limits. You know, there's a lot of talk about AI and AI is is not just a credit issue or a stock issue. It's an economy issue and there are parts of this economy that are going to get resized because of the AI effect. And so from our seat, you have to look at credit in the context of where the economy for that particular sector is getting resized. What might have looked well 10 years ago doesn't make sense now from a limit. And we have to constantly think about the flywheel effect. Are we

>> I always think it's interesting when the suits start getting on saying, "Oh yeah, you know, we got to be a little more conservative with the debt. You know, we've been a little too aggressive with the debt 10 years ago. Bring on the debt today. There's still opportunity, but we're going to be a little more conservative." Yeah. Yeah. Yeah. Whatever. Clown show. >> Corporate debt issuance. like the bigger things that can really

>> Oh, they're talking about debt issuance, too. >> But that's where they're much more concerned about what's happening with things like the national debt right now. That's getting a lot harder to swallow. Those huge numbers that are coming out there and I think that's probably going to have more of a weighing. But I think

>> think at the end of the day, as long as earnings are hanging in there, and they are, and I think this week we have Nvidia on Wednesday and then we have the Fed speaking on Friday. I think Nvidia is going to be the bigger story. if they can prove that story continues and especially what they're saying memory chip and how that's affecting their demand and their customers affordability. I think that's going to be a lot more important than what the Fed says or doesn't say because people want clarity. I think the one thing we know from work to this point is he doesn't want to give that much information. People just need to get used to that.

>> Well, I just wonder if he's now in some respects forced to give more information. Don't forget it's not a news conference. It's no Q&A. It's a speech. And we remember when Chair Pal gave his speech right before the rate hike regime was 8 minutes. So who knows what we're going to get.

>> I've argued on the show AI is a much bigger story than the Federal Reserve and I've I've made that case. I'm certainly not alone, but I made that case for a year or two. I'll just add real quick with respect to Worsh everyone's focused on the forward guidance side of things. If there is going to be a surprise from him this weekend, it's going to be him, I don't want to say feeling the pressure, but choosing to respond to the demand, not for forward guidance, but for some call.

>> So, Jackson Hole starts on the 27th, but Wars speech is actually the morning of the 28th. How interesting. She is right. Uh, the agenda comes out on the 26th.

>> and how that's affecting their demand and their customers affordability. I think that's going to be a lot more important than what the Fed says or doesn't say because people want clarity. I think the one thing we know from work to this point is he doesn't want to give that much information. People just need to get used to that.

The reception is on the 27th. Wsh delivers the opening speech on Friday and then you get like the interviews and panels and all >> regime

>> guidance. I'm not telling you today, this is what he was saying. I'm not telling you today where I'm going to hike or cut, but I feel like I can give you the ideas in my mind on what the mechanisms are to get us from point A to point B. Rick Reer agrees with the reforms that Chair Wars is talking about making and even he last week when he was sitting here says the market does need a little bit more so maybe that it can settle down.

>> was 8 minutes. So who knows what we're going to get. Do you agree though that Nvidia is the the biggest story this week? Because you don't expect

>> They need better tools to figure out the accuracy of inflation, right? Because that's the other thing that they're focused on. It can't just be core PCE. He's looking at a lot of various different metrics. Looking maybe at the trailing three month. That's what Rick Reer also talked about. The trailing three month is 1.6% 6% in CPI. So, there's other ways to look at inflation and I think that's what they're going to try to say. And I just think it's too early, Scott.

>> and she's going to be here in a second after you're finished, but go ahead, finish your phone. I'll bring her. >> Let's bring her in. U Christina is is waiting anyway to uh to talk to us. And look, the stock's been on a six-day losing streak. It's down again. And that was after a nice move higher. So, it maybe complicates the story for investors a bit going into Wednesday because investors right now are just hesitant what's going to be that next major catalyst to move the stock, which is maybe why they're not getting in. Uh Dan mentioned it, the numbers at this point really aren't the story. Beat assumed across the board. We know that demand surging, lead times have stretched from 6 weeks a quarter ago to nearly 40 now. Wall Street actually wants is specifically in the guide, the first real read on how fast the new Reuben chips are ramping the next iteration. Then there's also the margins. Nvidia told customers it's raising AI server prices as much as 17%. It's reported that uh driven by soaring memory costs. The question is whether that protects its mid70s gross margins. You can see on your screen which is estimated to decline in Q3 maybe even more or if memory still eats into it. Watch too also for the Vera CPU a new multi-billion dollar line as AI agents take off. And then to your last point all of you on the panel hanging over it is the circular financing debate. Nvidia is bankrolling the same customers who buys these chips. Like it's up to $15 billion backs stop for an OpenAI data center in Ohio. That's just one example.

>> That's where this um I'm going to pull it up. This uh there was a section of the all-in pod that drove me nuts because Sachs just laid out this data that I just thought was not right and I think it's worth talking about. and then you get like the interviews and panels and all

Uh let me go find it. Okay. Oh, I found it. Okay. That's Yeah. Where they also talked about the DSPs and all that. Okay. So, this is the like a week ago one.

Let me see here. This was Yeah. nine days ago. Okay, let me pull this because this is like the foundation of what's going on with hyperscaler spending and we need to find 40. Let's see where in it they specifically talked about 100

See if I can find that. That was finance plans. How the market can all fall apart. Yeah. Okay. Here we go. >> because investors right now are just hesitant what's going to be that next major catalyst to move the stock, which is maybe why they're not getting in. Uh Dan mentioned it, the numbers at this point really aren't the story. Beat assumed across the board. We know that demand surging, lead times have stretched from 6 weeks a quarter ago to nearly 40 now. Wall Street actually wants is specifically in the guide, the first real read on how fast the new Reuben chips are ramping the next iteration. Then there's also the margins. Nvidia told customers it's raising AI server prices as much as 17%. It's reported that uh driven by soaring memory costs. The question is whether that protects its mid70s gross margins. You can see on your screen which is estimated to decline in Q3 maybe even more or if memory still eats into it. Watch too also for the Vera CPU a new multi-billion dollar line as AI agents take off. And then to your last point all of you on the panel hanging over it is the circular financing debate. Nvidia is bankrolling the same customers who buys these chips. Like it's up to $15 billion backs stop for an OpenAI data center in Ohio. That's just one example.

>> value. laid out this data that I just thought was not right and I think it's worth talking about.

It was see. Come on. Overbuild. Okay, I will find that. Where they also talked about the DSPs and all that. Okay. So,

Oh, but I can't. No, it's impossible to find. Oh, whatever. I might just not reference it, but it was it was in this segment. just slipped together. That'd be a disaster for everyone. It's just that you get a glut of compute and you get an overbuild.

>> That's it. You get an overbuild. Yeah. nine days ago. Okay, let me pull this

>> And in the same way that we had dark fiber after the com crash, if you had dark GPUs, that'd be a because this is like the foundation of what's going on with

>> Okay, so that's at 103. But where does he mention? hyperscaler spending

So that's at 103. Call it 20. and

Uh where does he mention the 100 ARR we need to find 40. Let's see where in it they specifically talked about

>> outcome? Because it is so hard to gear. Did I shift gears for a second to tell you what's so brilliant about what Jensen did here? 100

you know, no >> but where's the finance plans. How the market can all fall apart. Yeah. Okay. Here we go. was it's impossible to find. Oh, whatever. I might just not reference it, but it was it was in this segment.

the 100 ARR Nvidia is bankrolling the same customers who buys these chips. Like it's up to $15 billion backs stop for an OpenAI data center in Ohio. That's just one example.

what Jensen did was so brilliant it's because the numbers are getting so billion dollars of capex the company just raised hundred billion and it can grow uh as big as the the the TAM actually is right it's removing that constraint so for just to take one example Elon wants to add somewhere around six to eight gigawatts next year. We know that that would cost3 to400 billion dollar of capex. The company just raised hundred billion in its equity and debt offerings. So obviously they would have to go out and finance that somehow. And as we talked about on our previous episode, the simplest way to finance it would be to get seller financing from Nvidia, especially given that the payback period could be as quick as one year. So now Jensen is creating the you could say the line of credit using these big banks using these big private equity shops and he's making that available and that's going to now benefit all of these downstream purchasers.

But that's still not where my Elon comment is because he buys the 100 ARR uh hook line at sinker. run rate? We talked about this ax a bunch will end the year between 100 and 120 billion dollars. This is an extraordinary ramp up of revenue that we have never seen in Silicon Valley uh at a $2 trillion valuation. Obviously, you all can do the math, but it's 16 to 20 times sales, which is a segment of what space that might be.

thought as you know Elon said that they were expecting right so if all of a sudden there are too many people racing to provide this compute of compute and you get an overbuild and in the same way that we had dark fiber after the dotcom crash if you had dark GPUs that'd be a disaster for everyone especially if you built out your compute infrastructure expecting a spot price of $30 to $50 a watt as you know Elon said that they were expecting Right. So if all of a sudden there are too many people racing to provide this compute and now there's an over supply and the market crashes that'd be the risk factor. In a weird way all the political headwinds I think ensure against that outcome because it is so hard to build data centers for all the reason we said there's a whole

this is an argument that has been circulating as well. I personally think they're these companies hedge by building in multiple places on purpose to avoid that panic hysteria hoax going on that actually it's those political headwinds I think will almost guarantee that there's not an overupp relative to the exponentially growing demand. So

that's really helpful and he's comparing that to exponentially growing demand. Hopefully it does actually grow exponentially.

There will be a pileup of companies behind it because Anthropic they're making 100 billion per gigawatt. No, they're not. That's the line I've been looking for. Found it.

So they're not they're making you know that was a sort of fatal flaw uh because right now they're expected to have what 64 billion so call it ARR now uh annualizing uh one month after Claude right so we get this explosion in Claude in Q1 so we get a $64 billion uh ARR annualized ARR run rate.

Current gigawatts uh one to two disclosed. Okay. Capacity 5 gawatt. Oh, there are multiple 5 gawatt deals. I think there are two 5 gawatt deals. What deals are they doing? Anthropic secures. Google, Broadcom, and Amazon. Right. So, that's what it was.

Let's write that down. So, capacity, I've got 5 gigawatts coming from Amazon. And then I've got another 5 gawatts coming from Google and Broadcom who's now taking on debt to kind of sort of, you know, make that happen. Okay, so that's our annual run rate right now in billions, which sort of implies

Current gigawatts at 1 to two. So, best case rev per gigawatt 64. That's best case scenario. Um, worst case rev per gigawatt would be the lower end. So, that would just be, you know, 64 divided by two, right? And then so if you take the midpoint which is potentially more realistic. So midpoint rev per gigawatt is actually probably somewhere around 64 divided by 1.5. That's just taking the midpoint. That's probably closer to their ARR but after co-work explosion, right?

So, how sustainable is that? But markets as you just heard from Sachs, right? Sax buys uh Hookline and Sinker 100 billion ARR per gawatt. uh enough to pay Elon's 30 to 50 billion per gigab. Uh most expect the break even or I shouldn't say the break even. Uh most expect the stabilized revenue or I should say cost per gigawatt to be closer to 8 to$14 per. So hence there's a big multiple and it's like if this is right if Elon's if this is correct and this is correct uh very bullish uh very bullish this would then be neutral to bearish right and then obviously over supply would be even worse cuz that's not over supply pricing that's like neutral pricing. So this would be even worse uh returns. So that's the question. That's the danger you want to avoid. And based on Anthropic's current numbers, we're at 42 in revenue. Okay? That's not enough to pay for 30 to 50 in cost. That's why you lose money. That's why I actually agree that the anthropic IPO is so freaking critical.

So I think that's uh that's really useful uh information and and why like people are on pins and needles about this anthropic IPO as a catalyst.

Hyperscalers have lined up 2.6 billion of future spending. Yeah, that's crazy. And it's a lot of it is to service this data that we just went through. Samsung HBM was up 122% quarteron quarter, suggesting HBM revenue to rise 80% quarteron quarter. Nice.

We saw the Bloom Energy. Nancy Pelosi. Good old Nancy Pelosi. So, this is about Wy. This is JP Morgan. JP Morgan Fed chairs open remarks on Friday morning. Jackson Hole blah blah blah often not marking move in. We'll see about that.

Standard deviation of daily changes in the 2year is around five bips. Yeah. Uh that's in the daily changes. Okay. Daily. Well, whatever. Uh let's see. The Fed chair's opening speech typically addresses bigger pictures for monetary policy. That's fair.

It just it all comes down to a surprise. So, we did this analysis and you know our analysis indicated it all comes down to the surprise factor. If we go into Friday expecting limited action, just like JP Morgan says here, the the risk is uh a hawking. The risk is hawking uh and a surprise, you know, the pants get put on so to speak. Let me see where we made had the note the data on that. That would be fed in the meet Kevin app. August 20th, four days ago. Where did we talk about here? A history of surprise jholes. Dovish surprise in 2025, hawkish surprise in 2022. And so when you get the hawkish surprise, that's when the market really has a poopers. So, you know, I don't think we'll get it, but there's always a risk of that.

Specifically, the after effects of the pandemic forced the chair pow to communicate sizable shifts blah blah blah could resemble the more muted grain span years. What? Look at this. With chair war skewing forward guidance, this year's market response could resemble more of the muted green spare years. Right, right, right, right. So that's why nothing uh is the base case because uh everyone is convinced will get no forward guidance uh and then he'll move on. Uh makes sense because that's what Trump would want right now too. And it's kind of like a puppet game like you know see no evil uh speak no evil about like what actually is going on.

What about over here on WSH? This is Morgan Stanley's summer negative momentum in equities has continued uh pullback and treasury yields follow those uh sizable announcement yeah last week and the reversal ongo oil price rise 12 month Brent futures hit a two-month high on Friday, which isn't far off the peak from May. So, expectations of oil prices of those higher oil prices put upward pressure on yields as well. Right. Conflict is said to stay in headlines. more or if memory still eats into it.

not worried about this. Where is the dollar? A mortgage back security. Well, that could lead to a car crash. But actually, this is why I think it's the anthropic IPO is really important for the market is getting those quarterly earnings and be able to see their numbers every quarter. I think it'll become probably the most important signal that the entire industry has. He's so right. Why? Because you got all these people out there saying that AI is a bubble. This capex is not justified. There won't be an ROI. And everyone's going to be looking to Anthropic's quarterly earnings as the pace car. Not just them. It'll be them and OpenAI and SpaceX, but you'd have to say that Anthropic right now is the pace car. And they're going to be looking to them to see if the demand signal is there. And if there's a hiccup or a wobble, you know, it'll show up on the brakes. Everyone in the chain, everyone. And I think the anthropic S1 is kind of going to make that clear. I will find it because there's a part where he kind of hookline and sinkers sinker sort of absorbs. Oh, I found it. Oh, okay. It's right after the 103 part. Ah, okay. Okay. Okay. Okay. Okay. Okay. Oops. Where did I just put it? What did I just do with it? It's all right. No problem. Got it. Okay. So, it's about 104 in. So, I got to write the times down when I write these notes in the future. I don't have to refind it. Uh, tweet they get sent overseas. I'll tell you where all of a sudden there are too many people racing to provide this compute and now there's no and the market crashes. Sure. against that outcome hysteria/hoax te that there's not an overupp that Jensen did

So, where's the Elon par? Dang it. All right. My last look, but I'm pretty sure should be able to find it. Ah, is it here?

Okay. All right. I think uh 109. Let's try this.

Let me see really quickly. Google, Broadcom, and Amazon. Right. So, that's what it was.

Okay, how else are things going? Q's still a little bit of that bleed kind of expected before all these big catalysts. Although that, you know, Besson announcement wasn't nothing for brah.

All right, we go throw another coffee on and turn my tea on here and then we'll summarize some of these things and kind of keep looking. Okay, hold on a sec. So, what is this? Samsung HBM was up 122% quarteron quarter, suggesting HBM revenue to rise 80% quarteron quarter. Nice.

Okay. Yeah, we heard about Bessence's end of the week threat, which is really just to kick down the road again. That's their classic nonsense. Let's kind of see what else is going on here.

Let's see. Anything in the wire services? Not really. After the best meeting, we really haven't gotten any kind of juice here. just approval numbers. That's boring.

I wonder how the JPY is doing. Let's go look. USD. Oh, yep. Deteriorating as well. 159. It's almost back to 160. Bessent couldn't do that either. Ohay. All right, let's listen over here. I'm going to go get some coffee and I'll listen in as well.

Connecting the two brands audiences with benefits like redeeming built cash for credits that can be used towards Equinox experiences. Pleased to say joining us here in studio 2 is the CEO of Built Anker Jane and Equinox Group executive chairman Harvey Spac. Great to have both of you back here on the program.

I mean this is great. It's our first time doing this together. This will be fun. It will be fun. Thanks for having us. Well, well, I know how uh great Equinox smells on the inside. What What does built smell like now? No, in all seriousness, uh talk to me.

By the way, in every movement at a built apartment building, we should probably give you an Equinox eucalyptus towel as a moving. Smell them on the street. It's walking by an Equinox. You know there's an Equinox. Leave now. I'm good. It's better than the Abberrami smell in high school. Wow. All right. Well, there you That's That's a All right. Let's start. Let's get serious now. I But I am curious though, Anchor. I mean I mean we talk about the partnerships you have obviously apartment buildings, airlines, hotels, restaurants. Uh why gems?

I mean look, yeah, built today represents one in four apartment buildings across the United States. And when you move into a new home, you're resetting some of the most important lifestyle decisions that you have. It's what's your new pharmacy, what's your favorite go-to restaurant and bar? Where do you park your car? And one of the biggest decisions everyone makes is what's my new gym? And so when we thought about what's the gym that we want our members to connect to and have our concierge be able to coordinate your training sessions, your spa, your workout classes, I mean, there's really no place better than Equinox.

I I I certainly understand what you get out of the partnership with Equinox. Harvey, I'm going to kind of pose the same question to you because I've always looked at uh Equinox.

All right, I'm back. How to get the coffee in a quick little bite. Okay, so where do we stand? Well, we need to talk about the sacks, the mold. What's this? Wapo. Um, seniors making six figures reach over a third of all social security benefits. Wow. Well, I mean, you paid into the system for your whole life. Um, all right. We saw the Bloom Energy. Nancy Pelosi. Good old Nancy Pelosi. Splash of Bailey's in your coffee. All right. What else here? Uh, let's go see what these suits are saying. Let's see here. So, this is about Wy. This is JP Morgan. JP Morgan Fed chairs open remarks on Friday morning. Jackson Hole blah blah blah often not marking move in. We'll see about that. Standard deviation of daily changes in the 2year is around five bips. Yeah. Uh that's in the daily changes. Okay. Daily. Well, whatever. Uh let's see. The Fed chair's opening speech typically addresses bigger pictures for monetary policy. That's fair. It just it all comes down to a surprise. So, we did this analysis and you know our analysis indicated it all comes down to the surprise factor. If we go into Friday expecting limited action, just like JP Morgan says here, the the risk is uh a hawking. The risk is hawking uh and a surprise, you know, the pants get put on so to speak. Let me see where we made had the note the data on that. That would be fed in the meet Kevin app. August 20th, four days ago. Where did we talk about here? A history of surprise jholes. Dovish surprise in 2025, hawkish surprise in 2022. And so when you get the hawkish surprise, that's when the market really has a poopers. So, you know, I don't think we'll get it, but there's always a risk of that. Okay. So, what else? So that's a nothing burger. Specifically, the after effects of the pandemic forced the chair pow to communicate sizable shifts blah blah blah could resemble the more muted grain span years. What? Look at this. With chair war skewing forward guidance, this year's market response could resemble more of the muted green spare years. Right, right, right, right. So that's why nothing uh is the base case because uh everyone is convinced will get no forward guidance uh and then he'll move on. Uh makes sense because that's what Trump would want right now too. And it's kind of like a puppet game like you know see no evil uh speak no evil about like what actually is going on. Okay. All right. What about over here on WSH? This is Morgan Stanley's summer negative momentum in equities has continued uh pullback and treasury yields follow those uh sizable announcement yeah last week and the reversal ongo oil price rise 12 month Brent futures hit a two-month high on Friday, which isn't far off the peak from May. So, expectations of oil prices of those higher oil prices put upward pressure on yields as well. Right. Conflict is said to stay in headlines.

Vid earnings have been a big macro event in their own right with reactions on par with jobs reports and CPI. Wow, that's an interesting comparison to make. But in most recent quarters, the positive earning surprises haven't been as big as we saw in 2324 after each of the last four Nvidia earnings. Nvidia price actually fell the next day.

Salesforce and Nvidia in two days. Those will be really fun to see. This weighs heavily on the growth sector like these chip names, not Nvidia per se, but uh I think that has a lot to do with it.

This is by the way now a little bit of a a tangent a tangential discussion on Nvidia. We'll listen to part of it and we'll keep getting uh in about one minute we'll hit a little bit on the payback periods and then we'll get down to the actual signal for the artificial intelligence.

I'm going to fast forward this. Gavin basically suggests that Nvidia is giving a 25% residual value guarantee to some of their customers which has been publicized and that there's an opportunity for them to potentially take uh you know some revenue share if revenue is above a certain level. basically a way of continuing this circular financing to make sure that cloud companies like you know Neb Nebius or otherwise can can continue to expand their buildouts and in other words buy more Nvidia chips.

And as we talked about on our previous episode, the simplest way to finance it would be to get seller financing from Nvidia, especially given that the payback period could be as quick as one year.

That's why SpaceX is then able to pay 30 billion to Nvidia for chips for the buildout. That's why Nvidia is able to

You know, 30 billion might very well just be the GPUs, but realistically, that's probably going up to closer to $40 billion on the latest generations of GPUs.

Nvidia's gone down four out of four of their last earnings after earnings despite beating expectations. It's not Kevin Worsh. It's not Nvidia earnings.

That's why Nvidia is able to pay TSMC and Micron and SKH Heinix and all the way down. It's a little funny when we get into this circular financing, but let's listen to Sax's reply

that the TAM is getting constrained by the ability to finance this build out. And what he's doing is alleviating that finance constraint so that he can grow uh as big as the the TAM actually is, right? It's removing that constraint.

So for just to take one example, Elon wants to add somewhere around 6 to 8 gawatt next year. We know that that would cost3 to400 billion dollars of capex.

Could be. Could be. See a lot of the assumptions that have been layered in here. Okay. Politics will stop overbuilding. Token growth is exponential and the payback period could be as low as one year. That those are huge assumptions. And you know the payback period is related to that 30 to50 billion of of revenue that Elon wants to pick up from compute.

Maybe it'll happen but look at the assumptions that are built in. And when you see the assumptions then you can kind of make the determination like okay we getting closer to uh you know the toppydoodle law.

They actually well Sachs comes up with a very similar thing that I've been harping on for a while in terms of the potential top uh and how to measure it. And it comes after 3 minutes of yapping from Gavin and the others. And it is right at 109. I'll play from about here.

Yes. If Anthropic slams on the brakes, there will be a pileup of companies behind it because Enthropic, they're making 100 billion per gigawatt.

No, they're not. Okay. Again, this is where David Saxs reads the crap on X and I don't think he's putting his critical thinking hat on. I think he's scrolling on X too much. They're not making hundred billion of ARR right now. Annualizing out the last day of July, they're at 65 supposedly.

Okay, we don't know if that's just because of the co-work explosion. We don't know uh if that's going to last, if that's going to get whittleled away by openweight models. We don't know. But David Saxs here just literally told you they're making a hundred billion dollars. That's enough to pay Elon's 30 to 50. But they're not. They're making on one or two gigawatts of compute, they're making about 6465.

Now, what's the problem with that? Well, the problem with that is they have massive compute coming online. What I also wrote over here that I didn't talk about yet is they have 5 gawatt coming from Amazon and another 5 gawatt coming from Google and Broadcom.

Now, not all of those 10 gawatts will get built. But even if half of this gets built, that's another 5 gawatt, which is 2 and a half to five times as much power as they have now. So, is their revenue going to be able to scale to keep making what they're making now this many more times over? Well, according to David Sachs, demand is exponential. So, in that case, sure, why not? But that's not what we're seeing. We're not seeing exponential demand. We're seeing declining second derivatives on token usage. There's still really big numbers. But that's a concern.

Okay. If compute, that's why they're able to pay SpaceX say 50 billion a gigawatt. No, they're losing money. They're not able to. They're losing money.

You know, a spot price for compute. That's why SpaceX is then able to pay 30 billion to Nvidia for chips for the buildout. That's why Nvidia is able to

the buildout is going to be way more than and this is the other problem. The numbers compress so much more than what David Sachs is saying and and is fine. Maybe maybe the exponential ramp will come but it's not 30 billion that you're putting into the buildout. That's the problem.

You know, he's telling you, oh, they're making a hundred and they're going to spend 50 to Elon and Elon's going to give 30 to Jensen and everybody's going to win. But the reality is Elon's probably spending, you know, 50 for the entire data center, right?

We forgot that there are other components to the data center. It's not just the Nvidia GPUs, it's the CPUs, it's the memory, it's uh the flash storage, the hard drives, uh it's the server racks, it's the building or whatever else. And this hundred billion dollar figure is not validated. It's not what reality is right now.

The put option of, you know, we'll call it the liberal put or put liberal. It's probably backwards. Uh, and then again the exponential. We can see there's a lot of sloppiness in this presentation from Sachs. And it's a little disappointing, but whatever.

pay TSMC and Micron and SKH Heinix and all the way down. It's the entire food chain. So if there's like a wobble, if somebody slams on the brakes,

if the pace car slows down, the whole is going to feel it. But it's still better for those numbers to come out on a quarterly basis. I think it's such in a vetted gap compliant way.

So Gavin, as we're sort of saying here, there's a kill switch, there's a break, whatever you want to call it. Is there enough space? Is there enough distance between the cars to

and the people behind it? Blah blah blah. Okay, so we've gone through that. The the whole point here is Sachs has just given a very clear heads up and we've talked about this as well, but it's nice to hear it from another point of view.

They're calling Anthropic the lead car and if they slam on the brakes as seen by quarterly earnings, everything else crashes in. Then of course there's this little bit of a debate of like, well, how much lead is there? You know, how much of a heads up or whatever.

This is why probably one of the biggest catalysts is not Kevin Worsh. It's not Nvidia earnings. Nvidia's gone down four out of four of their last earnings after earnings despite beating expectations.

It's the S1 that drops for Anthropic. The S1 for anthropic will be the most critical element of whether or not we are in a bubble and spotting the bubble and understanding some of these different numbers that I gave here will be really helpful for you to track what's going on.

I'm not saying anybody's right or wrong. Well, I guess I kind of am because we're not at a hundred billion dollar runway and and Saxs I'll say mistakenly said that they are.

Um maybe that's what you know finid is projecting by the end of the year. Okay, but that wasn't what was said. But whatever. Uh it's fine. It gives us the tools to look for identifying where we are.

Does Anthropic get to hundred billion dollars by the end of the year? Uh how much more compute is coming online relative to how much they have?

How are tokconomics improving or disimproving? How are the open weight models affecting Anthropic's ability to raise revenue? Honestly, almost all of this information is going to come to us from Anthropic public filings. It's even more critical.

Honestly, the SpaceX filings weren't really critical. We kind of knew that SpaceX was going to lose money handover fist. Like, it was just a matter of like how bad was it going to look. The anthropic ones are really going to be the first look at a leading frontier lab.

That's going to be insane. We'll be covering it a lot obviously in the course member live streams and the alpha report. Make sure to join us there at meetke.com. Use that coupon jhole before the Jackson Hole event and lock in uh lifetime access to our uh alpha membership uh courses on building your wealth, buy sell alerts. Check it all out over at mekevin.com.

Okay, cool. So that's sacked. Now, let's do a short little breakdown on the mold. The mold. The mold. The mold. The mold. I think the mold thing is very interesting. Maybe not other Maybe other people don't think it's very interesting, but I think there's some value here. all right, let's pop this up.

Home inspections on shoddy real estate and construction defects are taking off. And in this video, we're going to break down what this means for you. Should you avoid ever buying real estate? Should you be deathly afraid of mold? What about these Tik Tok videos of new construction homes getting torn apart by home inspectors? What is actually the bottom line? And how do we learn from this? Let's get into it. First, we have to, of course, start by watching at least one or two Tik Toks on this construction defect trend. Let's start with this one right here. This is America's best builder. This is the builder that makes $40 billion a year.

Nvidia reports Wednesday. Um, they'll beat the number. They'll guide up. I get that. But it's traded down the last four quarters and they've beaten, right?

part of it I think is a sentiment heading in to that. I think the competition is real in China not just on the chips potentially but obviously on