NVDA因基本面强劲而被低估（目标价>$300），但$5T市值及隐性风险（影子担保/循环融资）限制了上行空间，导致更偏好其他标的。

And one of the straws that will eventually one day break the camel's back is GPU growth going negative for a company like Nvidia, which we don't think that time is anywhere close. Uh if I had to guess, uh my guess, as every other person waiting for Nvidia earnings is there's no doubt Nvidia is going to beat, but it's 100% a gamble as to whether it's going to go up or down after earnings. Nobody freaking knows.

They're running those NVIDIA chips. I I personally think that Nvidia is really undervalued. However, I am a bigger fan of other plays for upside potential mostly because I think that that weight of a $5 trillion company, a company is really really heavy to the upside and really heavy to the downside.

But I I think you could reasonably justify a $300 price for this. And if growth continues like this, you know, the way it has been with the margins they have, this is an $800 or $900 stock.

This is a fantastic freaking company. Jensen is the best salesperson in the world for his products. And not just honestly for his products, but I almost think of him as the best salesperson in the world, period. He's really good.

the cash flow statement is also getting interesting for Nvidia because you do have that circular financing that you can find under the investing uh section of the cash flow statement, which honestly a lot of people don't pay attention to and I don't blame it.

We got Nvidia sitting at about 210. It's been pretty stable in this range, you know, off the 195 range. We've been bobbing up and down around this line, you know, big deal.

it's trading for like a 69 peg. It's trading so freaking cheap. We know that Michael Burry, who is short a lot in the AI stack, is actually now buying calls on Nvidia uh just to protect that uh eventuality that Nvidia ends up skyrocketing here.

You add 10% to Nvidia, you're adding half a trillion dollars with a snap of a finger.

I believe it is under the Nvidia tab when we were prepping for the Nvidia earnings.

It might align more frankly with Nvidia because I think robotics is is definitely something we're going to be seeing a lot of compute investment go into and I'm sure Jensen's going to talk about it a lot today but it's still early.

uh for GPUs that's going to be a huge driver of GPU demand. And see that's the kind of GPU demand that I think is going to be harder to sort of in-house. Today we have a lot of enterprise AI where you know modern can run therapy uh uh mutation algorithms with AI using 5090 GPUs. You know people play World of Warcraft with that. You don't need the Vera Rubin. But for robot vision based training that might be different.

and then of course we'll get Nvidia. Now Nvidia is always the last one of the bunch.

Okay, Nvidia is coming out in about 13 minutes, by the way, in case I know everybody's excited for Nvidia. That's coming out in in 13 minutes.

Obviously, we're waiting for Nvidia. Nvidia is not out yet. Nvidia is still expected to be out in about 11 minutes.

Who who called for software bottom in Q3 Q4? What the What is this? You know what this is, right? I know what this is. Hold on. Where is it? It's a Need FOR THE NEGAN BAT. WHAT the hell? I hope you got your pooping pants on cuz we're about to poop when Nvidia comes out in six minutes.

And then as far as reminder, Nvidia's in 5 minutes.

Nvidia is expected at 20 minutes after the hour.

Still waiting for Nvidia, obviously.

Okay. All right. So now we're going to get ready for Invidier. All right, let's get Nvidia. There it is. Nvidia beats. We expected that 75% gross margin, but how MUCH, BABY? WHAT'S THE FORECAST? OH, THEY BEAT BY A GOOD AMOUNT. 96.2 is how much they saw come in.

Oh, Q3 that was raised 108. 108 is the forecast. Uh 108 billion forecast Q3. Uh bigger than uh bigger than expected. That's really good. So that's bigger than expected.

I've got uh data center revenue beat. Data center revenue beat at uh let's see 89 billion versus 85.86. That's like the majority of the company's revenues, right? But uh that's okay. 89 divided by 85.86. We know where the concentration is. That's a 3.65% beat. Like compared to all of these other beats, that's actually really good, right? Uh these other companies are beating by like 1%. Nvidia just beat by 3.65% on data centers. That's great. Nvidia just beat by 3.65% on data centers. That's great.

Q2 revenue came in at 96.2 billion. Q2 versus 92.38 billion. That is 96.2 divided by 92.38. That's about a 4.1% beat. That's also really good. Uh I mean it's not like what we used to see. We used to see like the you know five to 10% beats Uh I mean it's not like what we used to see. We used to see like the you know five to 10% beats but uh Q3 revenue uh range. They also gave a range. So this is the forecast that I gave. It looks like the the expectation uh expectation was 10515. So we're beating that expectation, right? And then if I look at the range, range Q3 guide 10584 to 11016 billion. So that's the range of the guide.

but uh Q3 revenue uh range. They also gave a range. So this is the forecast that I gave. It looks like the the expectation uh expectation was 10515. So we're beating that expectation, right? And then if I look at the range, range Q3 guide 10584 to 11016 billion. So that's the range of the guide. Uh looks like based on um uh the alpha wire that I'm getting feeding through here. Uh you can get this in the meet Kevin app by the way. Alpha wire is indicating Nvidia shares down about 1.4%.

Alpha wire is indicating Nvidia shares down about 1.4%. Nvidia Corporation will pay the dividend. Nvidia Corporation is not assuming any data center revenue compute from China in its outlook. That's fine. Like neither is any investor right now. Like nobody cares.

Nvidia Corporation will pay the dividend. Nvidia Corporation is not assuming any data center revenue compute from China in its outlook. That's fine. Like neither is any investor right now. Like nobody cares.

Uh so gross margin forecast Q3 is expected down actually. So we've got Q2 gross margin of 75%. And uh Q3 we're looking at uh where is it? Where is it? Where is it? Where is it? Q3 gross margin 73.5 to 74.5% margin guide. So that's obviously lower than expected.

gross margin forecast Q3 is expected down actually. So we've got Q2 gross margin of 75%. And uh Q3 we're looking at uh where is it? Where is it? Where is it? Where is it? Q3 gross margin 73.5 to 74.5% margin guide. So that's obviously lower than expected.

Uh they took $7.8 billion in gains from securities. Wow. 7.8 billion securities equity gain in Q2. Okay. Uh, commitments increased from 119 last quarter to 279, $279 billion in commitments. Yeah, probably cuz Elon got a boatload of cash. Oops, that's I wrote that wrong. Uh, that is$ 279 279 billion in commitments if I can spell correctly.

Yeah, probably cuz Elon got a boatload of cash. Oops, that's I wrote that wrong. Uh, that is$ 279 279 billion in commitments if I can spell correctly.

Gross margins uh rose aided by Blackwell Ultra. Net gains. Okay, we talked about that. operating expenses about 9 billion uh versus the 8.97 expected. We'll go look at the numbers in a moment.

We'll go look at the numbers in a moment. Edge compute, edge compute actually came in nice and hyperscalers beat, but that's where the concentration is, right?

Edge compute, edge compute actually came in nice and hyperscalers beat, but that's where the concentration is, right? Like that's what Dell doesn't have.

Hyperscalers came in at 48.71 bill versus 43.55 billion expected.

commitments uh primarily due to memory procurement, right? In other words, like the dollar amount is up substantially. Vera Rubin is now in full production. Vera Rubin in full production.

But anyway, let's uh let's go actually look at sign data center lease packs with terms of 15. Nvidia is signing lease packs for 15year contracts. No. Uh, Nvidia signing lease packs for Oh, the very next line. Listen to this. Nvidia signing lease packs for 15 years. Uh, I'm like, WTF? Very next update. Very next line. Expecting to reassign data center leases to third parties. Yeah, we're going to offload this risk to private credit. What a joke. Okay, Nvidia investor relations. Come on, puppy.

Uh, let us go to the financial information for Nvidia. Man, I think we've earned a beer today. Uh, okay. Let's go. Hold on. We're actually going to go through the financials here. If I could just actually print the freaking document. It's a problem with Google Chrome. Oh, there we go. Download press release. Ironically, Safari makes it a little easier to save PDF web pages. Uh, Nvidia earnings press release 8 26 27. Okay, here we go. All right, let's pop this puppy in. Let's see what we got. Boom. Skitty booms. Boom. Record revenue up 85% from a year ago. Yeah, that is pretty impressive. 92% up on data center with this number should be higher because I see data center revenue higher than that uh in the notes or at least I thought I did. Okay, whatever. We'll come back to that in just a moment.

So, uh build out of AI factories largest infrastructure expansion in human history is ext expanding at an extraordinary speed. Agentic AI has arrived. It's doing productive work. It's generating real real value. Whatever. Jensen, we know you're the greatest salesperson ever. We love you in your black jacket. We know.

Okay. Data centers. Dash dash. Uh edge computing, data centers, two subm markets, hypers scales, and the others like AI clouds and industrial and enterprise. Enterprise, baby. That's what we care about. I want to see enterprise. Edge computing highlights uh Agentic for AI, including PCs, game consoles, robotics. Wait, you're now lumping gaming? Gaming is now lumped into edge computing. Lol. Come on, man. What is that kind of garbage? Segment out gaming and show me the small PP. I want to see it. Don't hide it from me. Selling spree. >> All right. Uh, we got over here, we got uh revenue growth. Here's that gross margin growth. Uh, but that was rounded. Look at that. We actually got rounded gross margin on the reported numbers. So we got negative uh you know.1% right there. Uh operating expenses actually grew 12% uh revenue coming in from the quarter over quarter. Still you got 20% quarterover revenue growth. Very good. Those operating expenses are popping off a little bit. Uh operating income though is exploding much more. And you still have pricing power, right? I mean pricing power is really uh if PP is the difference between revenue growth and margin and and I should say like operating income growth uh then Nvidia has big PP right I mean there's there's no question here like look at it you grow revenues in the quarter 20% you grew operating expenses 12% right that's 8% more big pee pe >> big pee pe >> they grew revenues year-over-year 85% they grew operating expenses 52% That's big pee pee. >> Big pee pee. >> That's pretty good. Net income's exploding.

Do we have a cash flow statement yet? So I could come back to this. I just want to see that. Oh, dirty. Dirty. No cash flow statement yet. Oh, no. Here it is. Okay. All right. All right. I was going to say if you're going to hide the cash flow statement until the 10 Q comes out, I'm going vomit.

Okay. I want to see the investments. Three months. Wait a sec. Wait a sec. Did I do it again? Is this the old press release? I've done this like every freaking time on Nvidia. Like, no wonder the numbers seem small. Hold on. Why? Why do they make it so damn hard? Is it just cuz it's not up on their web page yet? Yeah, it wasn't. Like, literally, when we went to this, the Q2 bubble wasn't filled in yet. Freaking losers. I got fooled again. It's okay. It's okay. Uh, honestly when I do that it is kind of fun though because then we can kind of compare to like how is it in the last one, right? Uh, Nvidia Nvidia earnings report A26. I'm going to say um this quarter. Get that freaking right. Uh, that was last quarter. Okay. Downloads Nvidia 130. uh this quarter. Okay. All right. Let's see the PP, baby. Let's see it. I want to see some of that pricing power. All right. Here we go. Here we go. Data center revenue up at 89 billion. That looks a whole lot more like what we were talking about. AI has reached an inflection point, says Jensen. It's doing useful work. Now, compute is revenue. Vera Rubin is in full production. During the second quarter, we returned $26 billion to shareholders. We'll pay a dividend.

All right, here we go. Quarter over-arter revenue growth. So, that's a little bit of a del. And look at the operating income. Look at that spread. The spread has narrowed. So, the pricing power spread has narrowed. Uh last quarter, 20% revenue growth, uh rev growth over last quarter. And we had how much? Uh, we had 12% operating expense increase. Uh, oh, wait, no, that's Yeah, because that's operating income right there. Did they not show that on this one? Oh, no, they did. 21% increase in operating Okay. All right. 20% to 12%, 18% to 10%. So, the spread's actually exactly the same. Uh, last quarter, 20% revenue over quarter. uh this quarter 18% rev growth over last Q. So a little bit of a slowdown. 10% increase in expenses versus 12% increase in expenses last quarter. There we go. That cleans that up nicely. So we can see net income on the quarter is only up 2%. Could see that right here. What happened here? quarter over quarter net income probably because of that's probably because of the anthropic gains lower um you know uh investment gains probably you know on the net income side so that's okay still pricing power here still PP here

let's go look quickly at their cash flow statement cash flow statement oh yeah look at this So that would be purchases of equity securities. That's this quarter, right? Yeah. Three months ending here. Why don't they show the last three months? They're not showing you the last three months. Oh, these numbers. Well, this would this would show you the total. Okay, that's fine. That that'll work. Yeah. Yeah. Yeah. Look at that. Purchases. So, that's this is the circular financing right here. and purchases of debt securities, maybe proceeds from sales, proceeds from sales. You know, that almost looks like more of sort of like a money market refinance. They sell a little bit of equity over here. And but but this right here, purchases of equity securities, $42 billion, 42.4 billion in uh equity purchases. uh basically the circular financing part uh 15.8 billion in this quarter. So that shows you this if you call the debt securities more of sort of a treasuries refinance right here. Uh in fairness they did sell seven uh minus 7.2 sold this quarter. So, you're still up $35 billion in equity securities that you bought.

How much revenue did they make or income did they get from other income this time? So, other income this quarter, oh yeah, see this quarter was a fraction. Last quarter they made 24140 minus 7773. Last Q 16.3 billion in equity in other gains. uh this Q 7.7. So that's why it looks like the net income isn't growing as much, right? That's reasonable.

R&D up seven versus that's versus last year. That's a little bit of an acceleration on the six month. So R&D is accelerating a bit. Some of that could be because of open weight accelerating a bit on R&D. open weight spend, uh, Grock spend, right?

Revenue, these are the year-over-year numbers, which are fantastic. Uh, but let's just do the gross profit comparison here. 72142 divided by 33853, uh, 2.13x on gross profit. And then revenue up here is 96 221 divided by 46743 2.05. 5. So that actually, you know, that's that's pee pe, right? They generated more revenue than they did gross profit. Uh, sorry, they generated more gross profit than they did in revenue growth. That's the pricing power portion.

Okay. If I look at the balance sheet, just a quick look over here, these are bills to pay of about $43 billion of bills. And we've got about 57 billion in cash versus 43 billion in bills. So, you know, the cash cash tightening a little bit. Cash is tightening, but we have 63 billion in receivables coming. So, not a big deal. You know, plenty of money available, frankly, here. Uh, and if I look at their long-term debts, I've got about 537 37.3 + 11 is uh 48.3 billionish longterm and their balance sheet is mostly fine.

So, let's see what other commentary there is. If there is anything else really exciting, DJX station, supercomputer, whatever. This is interesting here. Launched local AI initiative. This is the game changer right here. The problem is Deepseek doesn't make anywhere near as much money as uh as like Open AI, right? So, we're looking at maybe, you know, per the info, maybe 500 mil ARR at Deep Seek versus 65 billion at Anthropic. Like, it's just not even fair how much money anthropic makes versus these.

Uh, we've always talked about the robo taxi ecosystem that's coming. Groot for humanoid robots. This is huge. My VC invested into a robotic startup as well. Really excited for this. To me, this is 2040, but very exciting. 2040, but very exciting.

Okay. Uh, let's see. Japanese government Vera Rubin blah blah blah. Edge compute. Yeah, but you guys are build you're hiding gaming now in there. Whatever. Okay. Agent toolkit for agentic life sciences. SpaceX AI. Yeah. Here's the SpaceX mention which they just announced regarding by the end of uh you know this is where uh end of Q 427 they want to be launch ready. We'll see. We shall see. Uh this is huge that this is now in full production. It doesn't say scaled, you know, uh not necessarily scaled, but did um slam dunk Cerebrris. Uh IMO Cerebrris now fully uh derised from IPO concerns. You know, the 370ish priceish price way too expensive. Um, SRAMM has a future IMO and uh, you know, I I think it's it's nicely derised on its selloff.

Okay, we'll make that note for here. That's here. Okay, Vera Rubin ramping into production. Spectrum 6. You know, we talk about Spectrum, we're really talking about Ethernet, right? It's like, all right, that's cool, but got to put the names on it. It's got to look good, you know.

Okay, let's see if there's anything else in the CFO commentary. I kind of don't think so, but I will look Nvidia CFO commentary 8.26. And then we'll go look at how the sticks are reacting. Boom, baby. All right. Where'd it go? Oops. I put it in the wrong folder. Got it. User error again. The guy's a What are you going to do?

Data center revenue up 117%. Double blah blah blah. Strength of Blackwell Ultra. Enterprise AI cloud and enterprise uh edge computing revenues up 27% from a year ago, 13%. Uh driven by strong sales of Blackwell workstations partially offset by slower consumer PC sales. Yeah. Um Blackwell workstations exactly what we use at Reinvest. a big fan enterprise AI but lower margin for Nvidia than you know uh shilling uh the Frontier Labs the latest and greatest right I mean like let's just be real that's where their fortune is you know that's where they make the big dollars

Enterprise AI cloud and enterprise uh edge computing revenues up 27% from a year ago, 13%. Uh driven by strong sales of Blackwell workstations partially offset by slower consumer PC sales.

Blackwell workstations exactly what we use at Reinvest. a big fan enterprise AI but lower margin for Nvidia than you know uh shilling uh the Frontier Labs the latest and greatest right

I mean like let's just be real that's where their fortune is you know that's where they make the big dollars so structural land power shell for data centers next critical phase we partner with clouds we signed lease agreements we expect to reassign these to third parties this is some fugay right here.

so structural land power shell for data centers next critical phase we partner with clouds we signed lease agreements we expect to reassign these to third parties this is some fugay right here. Uh this is uh yeah, let's uh let's sign 15-year contracts. Uh yeah, boys, we signed a 15-year contract and we promise we'll uh shift the risk to private credit. That's basically what they're saying there. Whatever, man. It's all a bubble. We know it. We know it. We know it, but we love it. We It's our favorite bubble. And who doesn't like playing with bubbles?

Uh this is uh yeah, let's uh let's sign 15-year contracts. Uh yeah, boys, we signed a 15-year contract and we promise we'll uh shift the risk to private credit. That's basically what they're saying there.

Whatever, man. It's all a bubble. We know it. We know it. We know it, but we love it. We It's our favorite bubble. And who doesn't like playing with bubbles? All right.

All right. Sovereign AI Enterprise blahy blahy blahy. Second quarter infrastructure compensation benefits. Okay. We committed to our extensive network of suppliers, critical components. I don't care. Boring. Boring. Guarantees. In August, we entered into guarantees to provide credit support on the buildout of 4.5 gawatt. That's the OpenAI facility. I'm pretty sure I believe this is the OpenAI facility which will host exclusively Nvidia infrastructure under its 20-year lease with OpenAI subject to limited exceptions. Oh boy, I would love to know what those are. Our guarantee obligations are capped at a total of $105 billion guarantee. OH MY GOSH. HOLY MOLY. DAMN, BRO. You guys don't even got that on the balance sheet. Oh. Uh, our guarantee exposure declines as Open AI fulfills lease payments. Uh-huh. We have the option to provide credit support in phases for 3.8 additional gigawatts. That's almost twice as much. 3.8 divided by 4.25. That's 89% more. potentially 89% more suggests Nvidia basically guaranteed 105 billion plus 89% right 105 times 1.89 89 equals, bro. Come on, man. That's $198.5 billion of guarantees. That's $200 billion of guarantees. Uh represent about 1.5 million GPUs, approximately 150 to 200 billion of Nvidia revenue over the next 20 years. We expect to uh support multiple infrastructure upgrade cycles. Right, right, right, right. I mean, this is the

Sovereign AI Enterprise blahy blahy blahy. Second quarter infrastructure compensation benefits. Okay. We committed to our extensive network of suppliers, critical components. I don't care. Boring. Boring.

Guarantees. In August, we entered into guarantees to provide credit support on the buildout of 4.5 gawatt. That's the OpenAI facility. I'm pretty sure I believe this is the OpenAI facility which will host exclusively Nvidia infrastructure under its 20-year lease with OpenAI subject to limited exceptions.

Oh boy, I would love to know what those are. Our guarantee obligations are capped at a total of $105 billion guarantee.

OH MY GOSH. HOLY MOLY. DAMN, BRO. You guys don't even got that on the balance sheet. Oh. Uh, our guarantee exposure declines as Open AI fulfills lease payments. Uh-huh. We have the option to provide credit support in phases for 3.8 additional gigawatts. That's almost twice as much. 3.8 divided by 4.25. That's 89% more.

potentially 89% more suggests Nvidia basically guaranteed 105 billion plus 89% right 105 times 1.89 89 equals, bro. Come on, man. That's $198.5 billion of guarantees. That's $200 billion of guarantees.

Uh represent about 1.5 million GPUs, approximately 150 to 200 billion of Nvidia revenue over the next 20 years. We expect to uh support multiple infrastructure upgrade cycles. Right, right, right, right. I mean, this is the sales pitch, right? The sales pitch is like, "Hey, aka our R Vera Rubin uh will be upgraded to balls deep GPU 67 uh in eight years, you know, whatever."

Uh it's it's great. It's all good. It's a bubble. It's fine. Okay, good. So, that's really entertaining.

So, um, let's 44% of first half revenue was from three customers. Are you serious? Holy smokes. I'm just getting this coming through right now. 44% of 1H revenue equals three customers. That's it. WHOA.

One customer was 16% of Q2 revenue. Bro, wait. Wait until SpaceX raises money, like more money. You know, if Elon got a check right now from Adam Jonas's daddy for 400 mil tomorrow, Jensen would be raising guidance by 400 bill. Not 400 mil, 400 bill.

there will be more scrutiny of the company's balance sheet on the call. Uh let's see down a little bit. Stock repurchases investments. Blackwell Ultra Vera Rubin is next. Blotty blahy blahy memory costs are pushing. Data centers are obviously outperforming. Uh investors were expecting 105 billion. We did beat that. Okay.

And then where's Nvidia? And uh Nvidia's stable. Honestly, like that's pretty good. You know, people were expecting like a 5% swing on this in either direction to be stable like this 7. Like who cares? Watch. It'll end up going green tomorrow.

Well, Nvidia just reported earnings and wow, there are some skeletons in the closet, but they're also actually pretty dang good. I just want to be crystal clear about Nvidia. I I think markets are discounting the concern that in the future we're going to see revenues really start kind of second derivative going negative, right? The growth rate going down. Uh and and I think that's why the stock is cheap.

It trades for a six uh like a 67 peg ratio, which suggests the stock justifies today a price target well over $300. You could even go all the way up to potentially at the margins they run like an $800 stock.

But the biggest thing weighing this down is not only the fact that a lot is based on that future growth rate for this company, but it's a $5 trillion company. It's really difficult to move a $5 trillion company when your revenues are concentrated in the hands of just three bastards.

First half revenues. 44% of this company's first half revenues came from just three customers. I wrote a little note here that just wait until Elon gets his money on his $400 billion that he's got to raise to get to his 10 gawatts of compute because Elon is going to go Adam Jonas and Morgan Stanley. Please, baby, send me a check for 400 bill. I got to send Jensen some money. And that's pretty much what'll end up leading to more beat and raises from Nvidia, which is what we got.

We got a $ 108 billion forecast for Q3, which is bigger than expected. As usual, things are bigger than expected when Nvidia looks at the numbers. But people are so used to it being bigger than expected that the stock doesn't even really move on anymore. This what happens when you're so big that you're 5 trillion big. The expectation was 105.15 billion for Q3. We got 108. So, that's good. But it's just like, is that enough to really get people really excited again? I don't know. Maybe the stock will go up tomorrow now that this catalyst is over. I actually think that's kind of likely. I We'll see.

But usually once you get the catalyst event, it's like a clearing event and people could be like, "Thank goodness. Now, let's focus on the software bottom." No, we'll focus on the software bottom later.

We have uh lower investment gains reducing the net income growth at Nvidia. This is understandable because they took massive investment gains uh in the last quarter. We could actually calculate that by just subtracting this quarter's investment gains from last uh the last 6 months. And we can see we had about 16.3 in gains last quarter, 7.7 this quarter from investments into companies like SpaceX or uh Open AI or Anthropic or wherever they're investing in.

The big question to me is like how much more money are they throwing at other businesses? And they did show us that on the cash flow statement somewhere around here. Uh they ended up throwing Okay. Well, we'll find it in just a moment. Where the heck is it? This is their income statement. We'll come right back to this in a moment. This is their balance sheet. And here we go. 42.4 billion in equity purchases purchases. This is the circular financing part at Nvidia. 15 uh.8 billion last quarter. Uh sorry, 15.8 billion this quarter, 42.4 billion in total over the last two quarters. And they did sell 7.2. too, which means in total they made circular financing purchases of somewhere around $35 billion.

That wasn't really the news today that many people really wanted to pay attention to. The bigger news today was that in the CFO commentary, they actually talked about circular leasing deals. And this was quite quite interesting, but basically they signed these 15-year lease contracts where they're basically saying, "Hey, we're going to try to offload those leases to third parties." I translate this and I say, "Yeah, boys. We signed a 15-year lease so they can buy more of our GPUs, and we promise we'll subordinate this toxic debt to a third party private credit company." That's my opinion. Obviously, I I don't know, but that's kind of how I read between the lines here.

It's also worth noting in the CFO commentary here that yes, indeed, this is the OpenAI facility, but in August, Nvidia issued guarantees of $105 billion for the OpenAI facility. There are some limited exceptions, and those guarantees do decline as OpenAI starts paying uh Nvidia once this facility actually goes online. It is a 20-year lease to OpenAI, though. So that $15 billion commitment remains a sag like saggy balls. Uh it remains a sag on the stock, right?

But it's not just 105. You have to know that it's $105 billion plus the potential expanded capacity. The expanded capacity is right here. Listen to this. Uh over 20 years, we expect the site can support multiple infrastructure upgrade cycles. That doesn't just mean new chips, but it also means larger because they say right here, we also have the option to provide credit support in phases for approximately 3.8 additional gigawatts. Okay. Well, 3.8 additional gigawatts on top of the 4.25 you already have committed to. 4.25 was $ 105 billion of uh guarantees. If you add 89% to that, you basically have $198 billion of guarantees that you're giving open AAI, that's a lot of a guarantee that doesn't show up on your balance sheet. That's an offbalance sheet commitment.

And your balance sheet has weakened a little bit. It's still good. I got 57 billion in cash. I got $43 billion in bills. Okay, what does that get me? 57 minus 43 in bills. That means I got $14 billion in free cash for circular financing. Now, in fairness, I got $63 billion coming in. That brings me up to $77 billion. Then I can subtract out the $48.3 billion of long-term debt they have. Okay, so cash receivables less short-term and long-term debt leaves me at free cash of 28.7 billion. Okay, but if I'm potentially guaranteeing 105 billion to 198 billion to OpenAI plus some other losers 15-year leases, those debts don't show up here. uh at least as far as I can tell. You know, maybe some of the leases do, but I don't I certainly don't think the Open AI ones do, right? That's problematic. So, that's like shadow debt that does show up here.

Now, in fairness, again, they beat and raised. Okay, Q3 Guide came in with a larger range than we expected, 108 plus or minus 2% as usual. Data center revenue beat really nicely. We got $89 billion. Great. That's fantastic. That beat by 3.65 65% over the expectations. I got Q2 that beat by 4.1% over the expectations. I got gross margin that came in at 75%. This is still pricing power. They're still cranking money. They're still doing good. These people are printing money. They really are.

Uh their margin guide for the third quarter, not that great. Uh well, I should let me put it this way. It's still great. It's just not as great as Q2. Q2 75% we're now guiding 74%. So, a little bit of a decline. They argue some of this is because of memory prices and they raise prices a little bit, but not enough to offset all of that.

Honestly, this report overall, like we know what the risks are, but it's pretty good, right? Yes, we've got shadow circular financing going on with leases. We got equity investments being made to other companies. Yes, we know 3% of three customers made up 40 uh 4% of of your first half revenue. We know you guys are really exposed to the hyperscalers.

But part of it is probably just because Nvidia is such a big company. It's going to take some more effort probably during market open now that the catalysts are clear and people can digest. People could go, "This is still a really good company. It's still really freaking cheap. This should be a $300 stock."

Now, in fairness, I personally think other stocks are going to be able to outperform Nvidia from here. So, like just in full transparency, I've moved my bets from Nvidia. We've already talked. This is old news. We already know that. But that's how I can maintain this belief that like I still think it's a good company. I still think there's money to be made. I still think it's undervalued. I just think I can make more money somewhere else. Mostly because I think software is bottoming.

So as far as the rest of the Nvidia documents, it's pretty impressive. I want to be really clear like the SpaceX GPUs. Elon's going to blow all his money on these. Jensen is going to make so much money selling this stuff. It's brilliant.

the fact that they bought Grog uh or Grock, however you want to say it, and they're now doing the ultra fast token accelerators. Great. This is not necessarily scaled yet, even though it's in production, but it's a slam dunk. It did slam dunk Cerebrris.

this uh local AI initiative for these open AI or uh o open weight models which they're launching in partnership with Deepseek and Quen is really important to remember that per the information deepseek has maybe annual recurring revenue of somewhere around $500 million. You have to compare that to Anthropic. It basically means uh Anthropic is 130 times the size of Deep Seek, Deep Seek open, Anthropic closed weight. That's where the margin is. As soon as you take away that weight, the margin goes away. That sucks.

So now the question is, do we uh you know, is that a good thing for Nvidia? Well, Nvidia wants to sell more of their enterprise hardware to people. This makes sense. They're selling their humanoid uh uh uh models and hardware to humanoid robotics companies. My VC company has a lot of exposure into robotics company. Big fan of that. It's really a 2040 long-term play. If there's a recession or a depression between now and then, I'll get into robotics, but not now. I think it's a little overhyped right now, even though we're investors. Just being transparent.

Uh, and then of course they are making more money. They mentioned on their Blackwell systems for enterprise. I can't remember where they mentioned that, but somewhere they mentioned that. I'm very excited about their Blackwell systems. The basically workspace systems. I think that is where the

Blackwell workstations offsetting slower PC sales. We do this RTX 6000, RTX, you know, the 5090 gaming PC, doesn't matter. We can use those for local AI, uh, which then we maintain our proprietary data on. We don't have to send that to the cloud. And that's how we run the AI behind the homes AI, which remember the lifetime access for that is going away on Friday.

So, uh, going back to the Nvidia document here, these stations, they are now bundling together edge compute revenue and gaming. Gaming is usually a drag, a negative, and that's true across the whole sector because memory has gotten so expensive that, you know, people aren't really upgrading their computers or their servers unless they have to outside of artificial intelligence. Edge Research just did a big piece on that and suggested that's exactly what's happening. So, the upgrade cycle is really slowed for PCs. And I think that's why Nvidia is purposefully trying to hide from that a little bit.

I did also notice that their operating expenses on research and development did accelerate a little bit in this quarter compared to the last 6 months. Not much, but a little bit. I'm wondering if that's because of spending on grock, whatever, or uh open weight spending. I don't know.

They still do have pricing power though because we have a 2.13x gross profit increase from last year and uh pricing uh or revenue said uh what's what's it called? Um went up 2.05. So in other words, you made more money on uh on that revenue growth that you had than you previously were making, which is good. That's bullish.

Still got revenue gains in here from their equities. Obviously, that goes to a down cycle or in a down cycle that turns to the downside, which is not great. We've got those shadow guarantees. Not great. And uh and then, of course, we're spending more money on uh circular financing over here. We already know that. In fairness, almost $20 billion in buybacks. Uh however, they issued debt over here of about $25 billion. Although I think some of that I'm not sure if some of that was um a refinance because I see proceeds over here as well. So I'm not sure about that. Well, we have to compare the balance sheet. Doesn't matter so much. We we understand roughly the cash position of the company.

Uh and then if I look uh actually that's all I got. That's my take on Nvidia. Uh so we'll cover some of the other companies as well. But uh this is pretty exciting for Nvidia and honestly overall it's good. It's a cheap stock, but is it cheap because of concentration risk and the fear of how long can this giant company keep this going, especially with those shadow guarantees? Maybe. Okay, that is Nvidia.