History rarely repeats itself, especially speculative bubbles. As it becomes increasingly obvious that today’s AI bubble will dramatically burst, the real question is not when but how.

What makes this boom profoundly different from the DotCom crash of the nineties is OpenAI’s attempt to create an AI private monopoly by positioning itself at the center of trillions of dollars worth of self-serving “deals”. Sam Altman wants to simultaneously be the gambler, the slot machine owner, and the house. It’s a gamble that is, of course, brazenly rigged: he’s trying to simultaneously make OpenAI too important to fail and too well-financed to go public.

That Was The Week’s Keith Teare cutely describes this imperial play as “Come To Daddy.” But it’s more complicated—and more dangerous. By weaving OpenAI into the heart of America’s AI economy, Altman isn’t just building a company; he’s constructing a systemic chokepoint not just for Silicon Valley and Wall Street, but possibly for an entire global economy dependent on AI exuberance for growth.

If there’s a historical analogy, it’s the banking crisis of 2008. The US government bailed out the banks because they were supposedly too big to fail. The same will likely happen with the coming AI crash, especially given bipartisan American hysteria over the China threat —only this time, the crisis will center on OpenAI as both the dominant cause and the primary casualty of the crash. Here history might, indeed repeat itself: privatized gains during the boom, socialized losses during the bust.

Sam is dealing. Heads he wins, tails we all lose. Yes, the house always wins, especially when it is powered by OpenAI chips and wearing a ChatGPT hoodie.

1. OpenAI’s Platform Play Is Eliminating Startups

OpenAI’s developer day introduced an agent development platform, embedded ChatGPT applications, and Sora video generation—directly competing with dozens of startups. Keith Teare observed that over half of the 58 AI companies showcased at Andreessen Horowitz the next day had lost their competitive positioning overnight. OpenAI is no longer just a product company; it’s becoming a comprehensive platform that absorbs innovation opportunities across the AI landscape.

2. Potential Market Dominance Raises Competition Questions

Statistics from SQ Magazine claim OpenAI controls 88% of global AI interactions, with Anthropic at 8% and Google under 3%. While these figures require verification, such concentration would represent one of technology’s most rapid consolidations and raise fundamental questions about competition and innovation in the AI sector.

3. “Industrial Policy by Private Contract” Signals New State-Corporate Partnership

OpenAI’s relationship with the Trump administration suggests an emerging model of state capitalism without direct government funding. The state facilitates deals between major players and benefits through future taxation and ownership stakes in certain projects. OpenAI has become strategically essential for U.S. economic competitiveness against China—suggesting that no future administration, Republican or Democrat, could allow the company to fail. This creates an implicit government backstop without traditional public investment.

4. Infrastructure Funding Remains the Critical Challenge

AI requires approximately 10 gigawatts of power annually for the next decade—translating to trillions in data centers, chips, and energy costs. Recent deals involving Nvidia, AMD, and Oracle’s $500 billion Stargate project are down payments, not solutions. Energy costs remain a key constraint, with nuclear and solar options still expensive relative to demand.

5. The Speculative Age Concentrates Wealth

Andreessen Horowitz’s Alec Danco describes our current “speculative age” as defined by timing and short-term positioning. Unlike previous tech booms where retail investors could buy stock, OpenAI equity remains inaccessible to most, concentrating wealth among institutional investors and insiders while speculative energy redirects into prediction markets and gambling.

Keen On America is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit keenon.substack.com/subscribe

Transforming Data at Enterprise Scale with Kent Graziano

The Data T

Demo: Column Awareness in Action

The AI-powered Podcast Player