Latest / The PhilStockWorld Investing Podcast / 🤡 The NVIDIA Illusion: AI, Data Black Holes, and Market Structure

Transcript

🤡 The NVIDIA Illusion: AI, Data Black Holes, and Market Structure · The PhilStockWorld Investing Podcast

• 0:00
  Penny: Welcome to the deep dive. Today, we're we are tackling a day that just completely defied all logic.
• 0:08
  Roy: It really did.
• 0:09
  Penny: Thursday, 11/20/2025. I mean, this was the moment Right. The market's internal contradictions just boiled over.
• 0:16
  Roy: A huge, violent reversal.
• 0:17
  Penny: It was genuine cognitive dissonance. Good news suddenly becomes bad news, and it triggers this massive liquidation. Okay, so let's unpack this. This deep dive is all about the fragility of the whole AI ecosystem, the massive blind spots the Fed is dealing with, and why fear is peaking when prices are near all time highs.
• 0:37
  Roy: Right, and you have to see this reversal as a failure of the structure, not a failure of performance. We're dealing with three core tensions here that finally just snapped.
• 0:45
  Penny: Okay.
• 0:46
  Roy: First, you've got what we're calling the NVIDIA illusion. It's this massive AI spending boom but it's looking like it's built on circular financing. Right. Second, the Fed. They're facing what is basically a permanent data black hole after the shutdown.
• 0:58
  So they're driving blind.
• 1:00
  Penny: Completely blind.
• 1:00
  Roy: And third, you have this extreme fear, technically and psychologically, from institutional investors even while the indexes look, you know, totally confident.
• 1:10
  Penny: And to really crystallize that you, I mean, look at the numbers. The S and P 500 closed at 6,642 after this brutal four day drop.
• 1:19
  Roy: Yeah.
• 1:19
  Penny: But the CNN fear and greed index, it was at nine. Nine out of a 100.
• 1:23
  Roy: Extreme fear.
• 1:24
  Penny: Extreme fear. And the perfect symbol for the day was Nvidia amazing earnings, up 5.5% pre market and then closes down 3.1%.
• 1:33
  Roy: Which proves definitively that when liquidity fears take over, even the best fundamentals in the world don't matter.
• 1:39
  Penny: Okay, so let's start where the good news began. Nvidia's Q3 results for fiscal twenty six.
• 1:44
  Roy: They were incredible.
• 1:45
  Penny: They absolutely crushed it. Revenue was what? $57,000,000,000 profits jumped 65%. And Jensen Huang comes out saying Blackwell sales are off the charts.
• 1:53
  Roy: By all accounts it was a total validation of the AI thesis on paper. On paper. But what's so fascinating is how fragile the mechanism funding that performance actually is. The market saw those numbers and immediately asked, okay but is this real? Is this sustainable if the ecosystem is just feeding itself?
• 2:12
  Penny: And that gets us to this idea of circular financing.
• 2:14
  Roy: Exactly. It's this really sophisticated web where the companies selling the chips are also financing or in some cases directly paying for the purchases from their biggest customers.
• 2:25
  Penny: So it's not just venture capital money anymore. The suppliers themselves are funding the demand they're then booking as revenue. Can you give us the clearest examples of this loop?
• 2:34
  Roy: Absolutely. I mean, just take NVIDIA itself. They committed a massive $100,000,000,000 investment directly into OpenAI data centers.
• 2:42
  Penny: A 100,000,000,000.
• 2:42
  Roy: And what does OpenAI do with that money?
• 2:44
  Penny: They buy Nvidia chips.
• 2:45
  Roy: They turn right around and buy Nvidia chips.
• 2:47
  Penny: Or look at CoreWeave.
• 2:48
  Roy: Nvidia has a $6,300,000,000 deal with them. Oh, by the way, NVIDIA also holds a 7% stake in CoreWeave.
• 2:54
  Penny: So again, the revenue growth is tied directly to an investment in the customer.
• 2:58
  Roy: It's all connected.
• 2:59
  Penny: And that structure seems complex but the AMD deal with OpenAI feels even more, I don't know, abstract. How is equity financing chip sales there?
• 3:08
  Roy: It's highly creative and you could argue it's pretty risky for shareholders. So AMD is selling $90,000,000,000 of GPU capacity to OpenAI.
• 3:18
  Penny: But
• 3:18
  Roy: in exchange, OpenAI gets warrants. Basically the right to buy 160,000,000 AMD shares for a penny each.
• 3:28
  Penny: A penny.
• 3:29
  Roy: And those warrants fully vest if they buy all the capacity. So if AMD stock does well, those warrants become incredibly valuable and they could just completely offset the cost of the chips for OpenAI.
• 3:38
  Penny: So AMD shareholders are essentially footing the bill through massive dilution.
• 3:42
  Roy: They're financing the chips, yeah, and potentially giving OpenAI free hardware plus a huge equity stake.
• 3:48
  Penny: And what about that huge Oracle deal, the $300,000,000,000 cloud agreement with OpenAI? Did that add any stability?
• 3:54
  Roy: It just raised more questions. Analysts jumped on it, calling the $40,000,000,000 Nvidia chip commitment, a flexible arrangement.
• 4:01
  Penny: Meaning not binding.
• 4:03
  Roy: Not binding at all. And critically, the business itself runs at razor thin margins for Oracle, like 14%, when their usual profit margin is closer to 70%.
• 4:14
  Penny: So it's volume, but it's not necessarily valuable revenue.
• 4:18
  Roy: It's potentially value destroying revenue when you factor in the insane capital cost.
• 4:23
  Penny: And if we zoom out from all that, the math is just it's staggering. And I think that's really what drove that reversal. American consumers, the end users, are spending about 12,000,000,000 a year on AI stuff.
• 4:35
  Roy: Billion. Yeah.
• 4:36
  Penny: Meanwhile, tech companies are spending 400,000,000,000 building the infrastructure.
• 4:40
  Roy: That gap. That's the whole problem.
• 4:42
  Penny: That gap is the whole problem. Means the system is just running on fumes and future promises.
• 4:46
  Roy: Precisely. That's the $400,000,000,000 hole. It's why JPMorgan estimates that just to get a meager 10% return, the industry needs to generate $650,000,000,000 in revenue Into every single perpetuity. Into perpetuity. And then you have that MIT study that found 95% of organizations right now get zero measurable return on their AI investments.
• 5:08
  Penny: So, the shovel sellers are financing the miners, but the miners aren't finding any gold yet.
• 5:13
  Roy: And the market is starting to wonder how long the credit can last.
• 5:16
  Penny: Okay, so this whole AI financing problem, maybe it could have been contained if the macro picture was clearer, but it wasn't. The Fed is facing this deep, deep predicament driven by a total lack of data. Right? That forty four day government shutdown created a permanent data black hole.
• 5:35
  Roy: It really did. And for you, the listener, you have to understand this isn't just about a delay. We are talking about permanently missing data.
• 5:44
  Penny: So what's gone forever?
• 5:45
  Roy: We lost October CPI, the main inflation gauge. We lost October's PCE inflation, which is the fifth preferred measure. And for the first time since 1948, we will not have an official October unemployment rate.
• 5:57
  Penny: So they are making their December rate decision.
• 5:59
  Roy: Completely blindfolded, their models are compromised.
• 6:02
  Penny: And the irony is, the data that did finally show up, late and out of context, was the very thing that killed the rally.
• 6:09
  Roy: The September jobs report.
• 6:10
  Penny: The long delayed September non farm payrolls, it finally dropped and showed, what was it, a 119,000 jobs added.
• 6:19
  Roy: Yeah. Which easily beat the 50,000 consensus.
• 6:21
  Penny: And that was the rally killer.
• 6:22
  Roy: The ultimate rally killer. Normally, strong jobs, strong economy, that's good. But in this super leveraged market, everyone just heard one thing.
• 6:31
  Penny: The Fed has no reason to save us.
• 6:33
  Roy: Exactly. This labor market is too strong. No rate cuts for you. The market was desperate for weakness and it got strength instead.
• 6:40
  Penny: And unemployment did tick up a little, right? From 4.3 to 4.4%.
• 6:45
  Roy: A tiny bit. But it wasn't nearly weak enough to matter, not enough to force the Fed's hand.
• 6:50
  Penny: So how did that translate so quickly into just a wall of selling?
• 6:55
  Roy: It connected directly back to the Fed minutes that had just been released. Now the vote was 10 to two for a 25 basis point cut. Sounds fine.
• 7:02
  Penny: Sounds united.
• 7:04
  Roy: But the minutes revealed deep fractures. They literally said, participants expressed strongly differing views about what policy decision would be appropriate in December.
• 7:13
  Penny: And the market sees that division, sees the strong jobs data, and just puts two and two together.
• 7:19
  Roy: And realizes the consensus is probably to do nothing, to just hold steady.
• 7:23
  Penny: And the dissenters kind of confirmed that chaos. You had Stephen Marron saying they're cutting too slowly. He wanted a 50 basis point cut.
• 7:31
  Roy: While on the other side, Jeffrey Schmidt wanted no cut at all.
• 7:33
  Penny: So they're pulling in three different directions at once.
• 7:36
  Roy: Exactly. And the market just aggressively repriced that risk. I mean, the probability of a December rate cut just collapsed. It went from almost 99% a month ago down to somewhere in the thirties or forties after that data. The big realization on November 20 was that the Fed is divided, it's data blind, and the economy is just strong enough to prevent a rescue.
• 7:58
  Penny: So the high cost of money isn't going away?
• 8:00
  Roy: Not anytime soon. Right. And that makes every high valuation out there instantly vulnerable.
• 8:04
  Penny: So if Nvidia is the emperor, Coreave, the infrastructure provider, is the canary in the coal mine.
• 8:11
  Roy: And that canary has stopped singing.
• 8:12
  Penny: The stock completely collapsed. 50% in November alone. It went from what, 01/1987 down into the seventies?
• 8:20
  Roy: And CoreWeave's financials give you the clearest window into the weakness of this whole AI infrastructure thesis. This is a company with massive growth, dollars 55,000,000,000 in backlog, dollars 100 and thirty three percent revenue growth.
• 8:33
  Penny: And yet.
• 8:34
  Roy: And yet they posted a $110,000,000 net loss. They have an 8% negative margin.
• 8:40
  Penny: So they are losing money on every single dollar of revenue, even with that exclusive growth. How is that possible?
• 8:48
  Roy: It's the cost of capital and the hardware. That's the whole story. It's all about the debt structure. Core Weave has borrowed $14,500,000,000 at a blended rate of eleven-twelve percent. And that includes these GPU backed loans which are running at about 15% interest.
• 9:03
  Penny: 15%. And they're planning to spend another twenty-twenty 3,000,000,000 on CAPEX next year.
• 9:07
  Roy: Right. And CAPEX, for the listener is the money spent on physical assets like data centers and crucially those Nvidia chips.
• 9:15
  Penny: Okay. So they are borrowing money at extremely high rates.
• 9:18
  Roy: 15%.
• 9:19
  Penny: To buy assets that become obsolete incredibly quickly.
• 9:22
  Roy: That's the death spiral. We already saw Amazon AWS cut the rental price on the older H200 chips by 50%. The second a new chip comes out, the value of all their existing collateral just plunges.
• 9:35
  Penny: But the 15% debt payment doesn't.
• 9:38
  Roy: It doesn't. Which is why analyst Giluria has this severe underperform rating arguing the long term value could be less than $5 per share. So for us, the signpost to watch is that $70 level. If Core Weave breaks below $70 decisively, it's a five alarm fire. It signals the financing model for this whole thesis is unsustainable.
• 9:59
  Penny: Which brings us back to that psychological disconnect in the broader market, we are near all time highs.
• 10:04
  Roy: But the fear and greed is at nine.
• 10:06
  Penny: The smart money is terrified. Warren Buffett is sitting on a record $382,000,000,000 in cash.
• 10:12
  Roy: Stanley Druckenmiller got out of Nvidia, got out of Palantir.
• 10:16
  Penny: 60% of fund managers are saying equities are fundamentally over valued.
• 10:20
  Roy: It is the absolute textbook late cycle scenario. You have all this institutional caution, and at the same time, retail investors are just flooding in.
• 10:29
  Penny: Pouring a 160,000,000,000 into equity ETFs in two months.
• 10:34
  Roy: That divergence just can't last, and technically, the market was so fragile. I mean, the S and P hit its highs, it was hiding this massive weakness. It had the worst market breadth in recorded history.
• 10:45
  Penny: Meaning most stocks were actually going down.
• 10:47
  Roy: 80% of stocks were declining, while the index was being dragged higher by just a handful of mega caps. The reversal just showed how thin that support really was.
• 10:56
  Penny: Alright. So we've laid out the macro volatility, the micro fragility. Let's pivot now and look at some practical lessons this environment teaches us, starting with risk management.
• 11:06
  Roy: You have to be flawless.
• 11:07
  Penny: This environment demands it.
• 11:08
  Roy: It absolutely does. And the high value lesson we'll walk through is simple. Never ever leave short calls naked or without a stop. We saw this with the position in Barrick Gold where a trader had a long term short call say a $20.25, $22 strike.
• 11:24
  Penny: And the stop just blew right past it.
• 11:25
  Roy: Blew right past it.
• 11:26
  Penny: So for the listener just quickly why is that so dangerous?
• 11:29
  Roy: Great question. I mean, if you sell a short call as part of a spread, your risk is capped. You sell the 22, you buy the 25, your max loss is known. Right. But if that short call is uncovered or if the stock runs so far that you basically have a naked short call, your risk flips from bounded to unbounded.
• 11:49
  Infinite. You have to assume a 20% move against you can happen at any time.
• 11:54
  Penny: So you need a stop loss.
• 11:55
  Roy: You must have one. And the fix for this particular trade was, it was really a master class in salvaging a bad position. The strategy was to roll the blown up $20.25 dollars $22 calls.
• 12:09
  Penny: Which had way too much delta risk.
• 12:10
  Roy: Way too much risk and rail them up in price and way out in time to a much higher strike maybe $20.26, 30 five's.
• 12:16
  Penny: So that reduces the immediate risk and buys you time.
• 12:19
  Roy: It buys you crucial time and then they added a totally new long term bull call spread something like the twenty twenty eight, thirty, thirty seven's.
• 12:26
  Penny: And the purpose of adding that new spread was what?
• 12:28
  Roy: To create fresh, protected upside. The potential gains from that new, healthy spread could then be used to pay for the paper loss on the original bad trade. It turns an infinitely risky position into a manageable long term plan.
• 12:41
  Penny: That's a great transition from defense and trading to defense and corporate strategy. Let's switch gears to corporate moats and talk about Adobe. I mean, they're facing a real threat from Canva, and now the totally free Affinity three Suite, is their moat finally breaking? Breaking?
• 12:56
  Roy: The threat is 100% real. You have to consider it. Yeah. But we see it more as a slow bleed than a mass exodus.
• 13:03
  Penny: Okay.
• 13:03
  Roy: I like to use the RC Cola analogy. Adobe is Coke. It's the premium industry standard default tool.
• 13:09
  Penny: Right.
• 13:10
  Roy: Affinity is now the free RC Cola. It's functional, the price is unbeatable, but it just doesn't have that institutional power.
• 13:16
  Penny: So Adobe's moat is really about the enterprise customer.
• 13:19
  Roy: It's about institutional workflow gravity. It's enterprise, it's IT departments, it's the entire education market. Schools teach Adobe. That creates generational walk in.
• 13:29
  Penny: And those customers care more about support and compliance than a free price tag?
• 13:32
  Roy: Exactly. They prioritize support, security, integration. They're incredibly sticky. And if you look at the valuation, Adobe is trading at a forward PE of 15.6.
• 13:44
  Penny: And for context, that's low for a company like Adobe, right?
• 13:47
  Roy: That is very low for a dominant tech leader. A growth stock like Adobe would usually be closer to twenty twenty five. That 15.6 tells you the market has already priced in these competitive fears. It's trading at like a 30% discount, even while the company keeps raising guidance.
• 14:03
  Penny: So the market thinks the moat is breaking, but the business results say otherwise.
• 14:07
  Roy: The enterprise customers are staying put for now. So after all that, what does this reversal on November 20 really mean? I think the duality is the key takeaway. Yeah. NVIDIA proved the micro AI engine is robust, it can generate incredible revenue.
• 14:21
  Penny: Yeah.
• 14:21
  Roy: But the macro environment, the divided fed, the data black hole, the indebted infrastructure players, It all proved that the high valuations are running on a parachute that is just way too thin.
• 14:32
  Penny: So for you, the listener, the key signposts to watch now are the support levels. The S and T 500 has to hold the 100 moving average and that October low. That's a shelf around $65.50.
• 14:44
  Roy: A very firm shelf.
• 14:46
  Penny: And if that breaks?
• 14:46
  Roy: If that breaks, the next major uh-oh line, where a lot of institutional stops are probably set, is down around $63.94.
• 14:54
  Penny: That will be the real test.
• 14:55
  Roy: That will be the decisive test of resilience.
• 14:58
  Penny: And here's a final provocative thought for you to explore as you process all this. Our sources noted that consumer loan delinquencies are spiking. Credit cards, auto loans, mortgages, all while GDP and job numbers still look positive.
• 15:12
  Roy: The definition of a late cycle inequality driven boom.
• 15:15
  Penny: Exactly. So the crucial question is this. Given that this whole circular AI spending scheme needs immense ongoing financing that already dwarfs consumer revenue, what happens to the entire AI infrastructure thesis when that consumer stress finally hits the job market and you remove that last pillar of economic stability?
• 15:34
  Roy: Thanks for joining us for this deep dive. We'll talk to you next time.