Nov 17, 2025 · 58m · big-technology
AI Bubble Special Report: Debt, Depreciation, and Losses — With Gil Luria
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this special report, host Alex Kantrowitz and technology research analyst Gil Luria conduct a comprehensive financial reality check on the artificial intelligence boom. They evaluate mounting corporate debt, controversial GPU depreciation schedules, off-balance-sheet financing, and unsustainable startup losses to determine whether the AI market is heading toward a severe financial correction.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 33.7% of the talking time here. How this is scored →
Alex as informed peer 5.7 Guest teaching 6.2 Guest disagreement 3.9 Alex pushing back 2.9
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 25:14 Dismissing hardware longevity defense as sleight of hand
Gil forcefully rejects SemiAnalysis's thesis on hardware durability, calling it an intellectual sleight of hand that ignores economic compute value per dollar.
Hardest push from Alex ▶ 24:10 Host counterattacks with SemiAnalysis reportAlex directly confronts Gil's rapid-depreciation argument by citing Jordan Nanos's hardware warranty data and hyperscaler upgrade cycle constraints.
Biggest teaching moment ▶ 7:57 Foundational lesson on debt versus equity financingGil systematically educates the host on core corporate finance principles, explaining why predictable cash-flow assets justify debt while speculative AI capacity requires equity.
Alex holds their own ▶ 18:10 Alex cites precise Michael Burry depreciation figuresAlex brings high-level quantitative analysis into the discussion, quoting Burry's specific projections of $176B understated depreciation and earnings overstatements across hyperscalers.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Market Opportunity Versus Speculative Bubble Behaviors | 6 | 5 | 3 | 4 | Alex opens with a substantive bull-case argument from Semaphore's Reed Albergati and quotes Lisa Su. Gil validates parts of the premise while clearly separating healthy balance-sheet spending from unhealthy debt-fueled startup behavior like CoreWeave. | |
| Debt Financing Risks and Potential Systemic Fallouts | 5 | 7 | 4 | 3 | Alex asks about Oracle's debt exposure and whether failure poses localized or systemic risk. Gil gives a structured 'Finance 101' lesson on why debt should back predictable cash flows rather than speculative multi-billion dollar promises from unprofitable startups. | |
| Institutional Lenders, Flawed Projections, and Perverse Incentives | 4 | 7 | 5 | 2 | Alex questions why top institutional lenders like JP Morgan would finance these deals. Gil explains the agency problems and bonus structures in banking where underwriters capture upside bonuses while leaving downstream risk to the firm. | |
| Michael Burry's Warning: The GPU Depreciation Debate | 7 | 6 | 3 | 2 | Alex demonstrates strong research by citing Michael Burry's detailed mathematical tweet on $176B in understated depreciation. Gil affirms Burry's thesis, breaking down the accounting distortions of 5-to-6 year schedules against 2-to-3 year technology obsolescence. | |
| Hardware Lifespans vs. Economic Viability: Rebutting SemiAnalysis | 7 | 8 | 6 | 6 | Alex challenges the depreciation bear case using SemiAnalysis and Jordan Nanos's hardware longevity arguments. Gil counters decisively, dismissing the hardware survival argument as sleight of hand and explaining economic value via dollar-per-flop pricing. | |
| AI Trajectory Scenarios and OpenAI's Strategic Direction | 4 | 6 | 4 | 2 | Alex asks if OpenAI executing on Sam Altman's revenue vision would resolve the bubble risk. Gil outlines a three-part mental framework (pessimist, optimist, maximalist) and advises OpenAI to stick strictly to ChatGPT rather than overextending into infrastructure. | |
| Mid-Roll Sponsor Break: Indeed Hiring Platform | 6 | 5 | 3 | 3 | Following the mid-roll ad read, Alex brings in a Wall Street Journal investigation revealing how hyperscaler profit growth is heavily circular and funded by OpenAI's massive losses. Gil contextualizes Microsoft's risk profile versus Amazon and Google, using an Uber customer-acquisition analogy. | |
| Meta's Aggressive CapEx and Special Purpose Vehicles | 5 | 7 | 4 | 2 | Alex explores Meta's aggressive CapEx increases despite lack of a third-party cloud business. Gil explains Mark Zuckerberg's maximalist positioning and breaks down the balance-sheet offloading mechanics of the Blue Owl Special Purpose Vehicle. | |
| Game Theory, Price Wars, and Collateral Crunch Risks | 7 | 6 | 4 | 3 | Alex cites Bloomberg and Odd Lots on the prisoner's dilemma in inference flat-rate pricing and Raymond James on collateral crunch risks. Gil confirms the game theory dynamic, asserting hyperscalers are intentionally positioning to buy bankrupt AI startups' data centers at pennies on the dollar. | |
| Power Constraints and Behind-the-Meter Energy Solutions | 6 | 5 | 3 | 2 | Alex closes by raising electrical grid bottlenecks, citing quotes from Satya Nadella, Mustafa Suleyman, and ZeroHedge. Gil explains how behind-the-meter energy solutions like mobile generators and diesel setups bridge short-term power constraints. |
Statements from this episode (23)
Luria: Hyperscaler AI capex is sound because balance sheets back it
“The healthy behavior are reasonable, thoughtful business leaders like the ones at Microsoft, Amazon, and Google that are making sound investments in growing the capacity to deliver AI. And the reason they can make sound investments is that they have all the cu…”
Luria: Palantir is the best company in the world right now
“Palantir is the best company in the world right now. I'm not gonna even say the best software company. They're the best company in the world right now, because what they're able to do is go into a company and ask them, what's your biggest need right now? What …”
Luria: CoreWeave is the poster child for unhealthy AI debt financing
“CoreWeave, which is the poster child for the bad behavior that I'm talking about. We're talking about a startup that is borrowing money to build data centers for another startup. They're both losing tremendous amounts of cash, and yet they're somehow being abl…”
Luria: OpenAI promised $1.4 trillion across Oracle, Microsoft, Amazon, CoreWeave
“OpenAI promised me three hundred billion. They also promised Microsoft two hundred billion, Amazon thirty eight billion, CoreWeave twenty five billion. In total, 1.4 trillion dollars.”
Luria: OpenAI will generate at most $15B revenue, lose over $20B
“A company that at best will have fifteen billion dollars of revenue this year and will be losing more than that. So we'll be losing probably more than twenty billion dollars this year.”
Luria: Falling GPU rental prices could trigger systemic AI debt defaults
“If we are here two years from now and there's hundreds of billions of dollars of debt, And the demand for AI stabilizes, or we built enough data centers to support the demand we have at that point in time, and the price for leasing pieces of AI for renting acc…”
Luria: U.S. Bank, J.P. Morgan, and Mitsubishi Bank lend to CoreWeave
“U.S. Bank, J.P. Morgan, Mitsubishi Bank, those are the companies lending to CoreWeave.”
Luria: OpenAI's $100 Billion 2027 Revenue Projection Is Completely Disingenuous
“So for them to say, oh, we're gonna have a hundred billion dollars of revenue by 2027, which Sam Altman just did, is completely disingenuous. He has no idea. He's competing against much bigger, more powerful companies that have technology that's at least as go…”
Luria: Three-year GPU depreciation would drastically cut Oracle and CoreWeave profits
“If you told Amazon and Microsoft, and certainly if you told companies like Oracle and CoreWeave, no, no, no, no. These chips will only generate meaningful revenue for three years. Their profitability would decline very dramatically. So again, from Microsoft, A…”
Luria: Claiming Five-Year-Old AI Chips Still Function Is Accounting Sleight of Hand
“Saying I have a five-year-old chip that works just fine doesn't mean that it can generate the same revenue that it did five years ago, which is the accounting question. So that's sleight of hand by these companies to tell you that the chip still works. If it's…”
Luria: Three-year-old AI chips will not generate sufficient revenue to run
“And that's where we think in a three year timeframe, three year old chips will just not be able to do enough computations to be worth keeping them on. So either we replace them or we use them for much less important things. That will generate much less revenue…”
Luria: Three-year GPU depreciation could slash AI company valuations by 40%
“If a company has to depreciate most of its assets, Over three years instead of five years, that means their profitability is going to go down proportionally. And we could have in, in, in, in a stylized case, the value of a company declined by 40% because an ac…”
Luria: ChatGPT has reached 800 million weekly active users
“And they came out with a great consumer product that we now know as ChadGPT and now has eight hundred million weekly active users and is driven the fastest growth of any startup ever.”
Luria: OpenAI won't make it if it builds datacenters and chips
“If they go down the current path, over committing, Deciding that they have to build their own data centers, they need their own hardware, they need their own chips, they won't make it.”
Luria: Roughly half of Microsoft's AI revenue comes from OpenAI
“Half of their AI revenue approximately which is again, a big piece of the Azure growth, which is the biggest piece of Microsoft's growth, is coming from OpenAI.”
Luria: OpenAI is a negative gross margin business
“Open AI is a negative gross margin business. They claim they're not, but they are right. Which is to say it costs them more to answer your chat GPT question than they make revenue from you.”
Luria: ChatGPT will become profitable as users accept higher prices
“So as we use ChatGPT a lot more broadly, we're going to be willing to go from a 20 dollar a month price point to a much higher price point. So at some point, this is a product that will be profitable. We just have to expand the usage so much that the people th…”
Luria: The AI chatbot market is not winner-take-all
“Unlike Uber and the rideshare market, which lent itself to winner takes most or winner takes all, chat is entirely not like that. Because I could have the same conversation with Gemini, and Meta is going to give me the tools to do this, and Grok is going to gi…”
Luria: Meta Lacks Enterprise Customers to Rent Out Excess Data Center Capacity
“Meta again is an unusual situation because they don't actually have business customers to rent this capacity to. It really is just Mr. Zuckerberg wanting a bigger and bigger toy.”
Luria: Meta used a Blue Owl SPV to finance data centers off-balance-sheet
“The other thing that happened with Meta that was interesting is that when they Went out to borrow. They didn't borrow the capital. They created a special purpose vehicle, right? They went out with Blue Owl and they said, you know, we'll put a couple of billion…”
Luria: Hyperscalers will buy bankrupt AI data centers for pennies
“So if I'm Microsoft, I know that in two or three years, I probably won't have to do any capex because I'll be able to buy data centers out of bankruptcy for pennies on the dollar.”
Luria: AI companies park diesel trucks behind the meter to power chips
“And then what these companies are doing is putting power, what they call behind the meter, which is to say generators and turbines and diesel trucks, because it's so lucrative that it's worth it for them to park 10 diesel trucks and run them so they can power …”