Better Offline — "The AI Money Trap, Part Two"
Host: Ed Zitron
Date: August 21, 2025
Podcast by Cool Zone Media and iHeartPodcasts
Overview
In this hard-hitting episode, Ed Zitron dives deep into the systemic financial fragility of the generative AI boom, arguing that much of the hype, investment, and expansion of AI infrastructure is not only unsustainable but potentially dangerous for the broader economy. Zitron meticulously unpacks the intricacies behind the eye-watering cash burn, convoluted company structures, and the lack of viable exit strategies for AI startups and data center developers—ultimately questioning whether generative AI is actually profitable or merely propped up by a feedback loop of unsound investment from tech’s biggest players.
Key Discussion Points & Insights
1. The AI Bubble’s Dangerous Expansion
Timestamp: 02:46-10:00
- Ed opens by summarizing the central thesis: The massive investments in generative AI (startups, infrastructure, capital spending) are unsustainable and poised to end badly for the economy and public markets.
- Reviews the case study of Cursor, an AI coding startup valued at $10 billion with seemingly no real path to sustainability. Cursor’s failure would directly impact Anthropic (its customer), setting off a chain reaction of revenue loss and potential collapse.
- Highlights OpenAI’s financial structure, burn rate, and dependence on continued capital raises—which are only possible if it converts to a for-profit entity and goes public.
- Quote: "OpenAI appears to have burned at least $10 billion in the last two months... That is an unbelievable cash burn, one dwarfing any startup in history, rivaled only by Xai, makers of Grok the Racist LLM, which loses over a billion dollars a month." (06:53)
- Raises questions about potential Microsoft strategies—would Microsoft actually want OpenAI to succeed independently, given it owns the infrastructure and IP?
- Notes the dubiousness of OpenAI/Anthropic revenue disclosures and doubts about the true customer base and path to profitability.
2. The Myth of AI Bailouts & Essentialness
Timestamp: 12:30-16:35
- Zitron forcefully rebuts the notion that government bailouts will save the AI sector, arguing there is nothing essential enough to justify it.
- Quote: "You can't bail this out. Even if this were gonna happen, and it will not, who would they give the money to and for how long?... If a bank fails, people lose actual money. If this stuff fails, what do people lose? What is the service that people lose access to? ChatGPT." (14:22)
- He interrogates the idea that generative AI products are irreplaceably "essential," observing that even with substantial user numbers (e.g., ChatGPT), it's not clear these products are crucial for the majority or justify their valuations.
- Discusses the flawed economics of selling LLM access below cost, training users to expect unlimited access, and setting up startups for ongoing loses.
- Predicts a sectoral explosion as unsustainability becomes evident: "I'm kind of hopeful because I think it's time to ask a basic question, which is what if generative AI just is not profitable?" (16:00)
3. Unsound Data Center Expansion and CapEx Risks
Timestamp: 20:08-32:27
- Cites Wall Street Journal data showing AI infrastructure capex is now outpacing the dot-com boom—a "private sector stimulus program" forming a dominant component of recent US GDP growth.
- Quote (Paul Kudrowski): "As a percentage of gross domestic product, spending on AI infrastructure has already exceeded spending on telecom and Internet infrastructure from the dot com boom and it’s still growing." (20:15)
- Lays out the scale: “The Magnificent Seven” (Big Tech giants) may spend up to $400–520B in 2025 on AI-related capital expenditures—much of it driven by only a handful of companies.
- Points to structural risk: Big Tech offloads data center construction to third parties like QTS (owned by Blackstone), raising existential concerns about who will ultimately “hold the bag” if the AI sector falters.
- Exposes private equity’s lack of an exit strategy—companies like Blackstone don’t want forever assets, they want to sell to someone, but at the current valuations, there are no buyers on the horizon.
- Reviews the shaky finances and questionable business models of data center players like CoreWeave and Crusoe.
- CoreWeave—relying on Microsoft as the main customer—faces enormous debt and burn, and is highly exposed if Microsoft pulls back (27:50).
- Quote: "Every time I think about CoreWeave, I feel crazy… If CoreWeave runs out of money, by the way, and they collapse, which I think is very possible, I’m taking some fucking names on this one." (28:07)
- Stresses that AI capex represents a systemic weakness, as the “boom” is not based on sustainable, profitable demand, but on faith in future value.
4. The Economy’s Growing Reliance on AI CapEx
Timestamp: 36:00-43:50
- Asserts that US GDP growth is now “held up” by five companies' continued investments in AI infrastructure.
- Quote: "If AI CapEx represents such a large part of our GDP and economic growth, our economy does on some level rest on the back of Microsoft, Google, Meta and Amazon and their continued investment in AI." (39:14)
- Observes that free cash flow (the “purest” metric of business health) has been decoupling from net earnings at the tech giants due to relentless spending—an ominous sign.
- Raises a red flag at Microsoft's slowing of new data center leases and construction: "What should worry everybody is that Microsoft… has signed basically no leases in the last 12 months… nothing they're doing suggests they're super into it. It's not good." (40:25)
- Warns of what happens when "capex reduction" arrives: not only will big tech stocks take a hit (given their disproportionate weight in the market), but economic activity and knock-on effects (layoffs, VC contraction, pension fund hits, impacts on support businesses and cities) could accelerate into recession.
5. Generative AI: A Gothic, Unnatural Industry
Timestamp: 44:00-48:30
- Ed concludes that generative AI resembles a "fever dream" rather than a viable tech revolution—a sector built on wasteful spending, with no underlying value or path to profit except via cartel power and obfuscation.
- Quote:
"The generative AI industry is, at its core, unnatural. It does not make significant revenue compared to its unbelievable cost, nor does it have much revenue potential itself. ...the inherently compute intensive nature of generative AI basically requires the construction of these facilities without actually representing whether they are contributing to the revenues of the companies that operate the model." (45:30)
- Quote:
- Critiques both media complicity and VC groupthink for fueling the bubble; questions why the market continues to reward abysmal fundamentals.
- Uses Paul Kudrowski’s summary:
"We live in a historically anomalous moment. Regardless of what one thinks about the merits of AI or explosive data center expansion, the scale and pace of capital deployment into a rapidly depreciating technology is remarkable. These are not railroads. We aren't building century-long infrastructure. AI data centers are short-lived, asset-intensive facilities riding declining cost technology curves, requiring frequent hardware replacement to preserve margins.” (47:55) - Ed's rejoinder: “To which I say, what margins, Paul? ...You can’t bail this out, because there is nothing to bail out. Microsoft, Meta, Amazon and Google have plenty of money and have proven they can spend it ...the popping of the AI bubble will have downstream ramifications, just like the dot com bubble.” (48:15)
6. What’s Next?
Timestamp: 48:30-50:34
- Ed forecasts a grim future if/when the AI capex-driven boom ends:
- Acceleration of layoffs, contraction in tech-centric cities, and VC pullback.
- Pension funds and retirement savings hit by slowing growth or tech stock collapses.
- The opportunity cost of this spending—“what if companies instead invested in actual productivity or gave everyone a raise”—is unknowable, but likely immense.
- Ed leaves listeners with a warning: "I keep asking that. The original title of this was called 'how does this end' and I really don't know how this ends. It feels as if everything is aligning for disaster and I really feel there's nothing that can be done to avert it." (49:37)
Notable Quotes
- On AI Valuations:
"No, really. I'll link to an article I've linked in the spreadsheet called the Strangest and Most alarming things in WeWork's IPO filing. It's a bad sign if your IPO filing is described as strange and alarming." (09:00) - On AI’s Dubious Value:
"The revenue you see today is what people are willing to pay for a product that loses money. And I cannot imagine they would pay as much if companies in question charged their costs." (16:15) - On Systemic Weakness:
"The majority of AI capex is from big tech, which is a massive systemic weakness for our economy." (32:10) - On Entrenched Waste:
"This is what happens when the tech media repeatedly fails to hold the powerful to account, catering to their narratives and making excuses for their abominable billion dollar losses and mediocre questionably useful products." (48:57) - On Big Tech Strategy:
"Hell, even if they just gave everyone a 10% raise, it would likely be better for the economy than this if we're factoring in things like consumer spending instead, it's just waste. Ugly, pointless waste." (49:56)
Tone and Style
- Ed Zitron delivers his analysis with a blend of sardonic humor, exasperation, and deep concern.
- Language is direct, often profane, and deliberately unsparing in its critiques of tech and financial elites.
- The episode is informed by detailed research (reports, estimates, direct interviews with analysts), but always colored by Ed’s sharp editorial perspective.
Suggested Listen Timestamps
- [02:46] — The economic and market consequences of the AI bubble begin.
- [06:53] — Unprecedented cash burn at OpenAI and Xai.
- [14:22] — Why an AI-sector government bailout is unrealistic.
- [20:15] — Wall Street Journal data on AI capex outpacing the dot-com bubble.
- [28:07] — The financial instability and future of CoreWeave.
- [39:14] — US economic growth now tied to handful of big tech companies.
- [45:30] — Generative AI as a fundamentally unnatural and unsustainable industry.
- [47:55] — Paul Kudrowski's warning on AI infrastructure as a "rapidly depreciating technology."
- [49:37] — Ed's final warning on the unavoidable disaster facing the economy.
Summary
This pivotal episode of Better Offline makes a forceful case that the generative AI gold rush is built on shaky financial ground, with Big Tech’s capital expenditure propping up not only the sector but—alarmingly—the wider US economy. Zitron eviscerates the notion that generative AI will ever become widely profitable, exposes unsustainable feedback loops of VC investment and capex, and ruthlessly questions the lack of any plausible exit or bailout path when things go south. The episode stands as a warning: not just of how the tech industry shapes our economic fate, but how unchecked hype and growth-at-all-costs can bend, and potentially break, the system itself.
