AI & US: Too Big Not to Fail?
AI towers over us. It is not just an economic force. It is social and political too. The public debate has two endings. In the good one, AI makes us so productive that work becomes optional and prosperity spreads. In the bad one, it destroys jobs, concentrates wealth, and hands more power to those who own the systems.
This essay asks how we get to either future. The answer is uncomfortable. US AI may be too big not to fail. That does not mean AI will fade. It means the path from here to adoption could run through a financial and economic crisis first.
Start with scale. Since 2013, more than US$3.1 trillion has been invested in AI in the US alone. That is bigger than the combined cost of the Vietnam War, the interstate highway system, Apollo, the Marshall Plan, and the eradication of polio. The positive societal dividends from successful Mega Projects were far-reaching. And the consequences of the failure – the Vietnam War – were also deep, prolonged, and disruptive. By scale alone, AI has to be one or the other. It is one vast bet on one technology. The larger the bet, the larger the return required. AI must now satisfy investors, creditors, customers, and society at once. That is a high bar.
The first test is cash flow. Can this investment generate enough revenue to service the capital behind it? The answer depends on price. Here the US faces China. China has built models with roughly 10% of the US capital spend while reaching about 90% or more of US performance. The cost of using a Chinese model is often only 10% to 20% of an American one.
That matters. Most users will not pay ten times more for a small quality gain. Frontier researchers may need the best systems. Most firms do not. They need cheap, reliable productivity. On that basis, the Chinese model is hard to beat.
It also travels well. Many Chinese models are open source, so users can adapt them. They are built for practical use across the economy, not only for frontier display. The US approach is more capital-intensive. It is more exposed if prices fall.
The second test is financing. China has less capital at risk and lower AI equity valuations. A correction there would hurt, but it would be less systemic. The US is different. Its Hyperscalers and labs have raced ahead of their own cash flow. They now need debt to fund compute, power, and data centers.
By my estimates, AI Hyperscalers and labs carry US$356 billion of long-term debt and US$ 248 billion of lease liabilities. Those are the visible obligations. Off-balance-sheet sit about US$900 billion of lease commitments and US$1.5 trillion of purchase commitments for data centers and advanced chips. The real exposure is therefore much larger than reported debt suggests.
This structure is fragile. Circular shareholdings increase borrowing capacity, but they also spread losses. If one participant weakens, others’ balance sheets are damaged. The risk appears dispersed, but it is concentrated in substance.
The borrowing structure adds another weakness. A Hyperscaler (think Meta!) may guarantee the borrowing of a special purpose vehicle. The vehicle buys chips and builds a data center. A smaller lab (think Anthropic!) then signs a 15- or 25-year lease to use it. The debt is priced on the guarantor’s strength. But the cash to service it comes from the lab’s product. A large liability is therefore pinned to a narrow revenue stream. If the product misses, the structure breaks.
That mismatch becomes lethal if US providers must cut prices by 80% to match China. At those prices, the capital stack cannot earn enough. Debt service becomes doubtful. Equity values would fall first. Credit would follow.
A US AI collapse would then become an economic winter for four reasons:
AI is already embedded in capital spending, corporate strategy, and market optimism.
A break would hit credit markets, not just equity multiples.
Policy would lag. Denial would come before repair.
Households now own the AI trade through portfolios and pensions. Losses would cut spending and lift precautionary saving.
David Roche is a veteran global macro strategist with more than 40 years at the intersection of geopolitics and markets. He is the founder of Quantum Strategy & Geonomics, a Singapore-based macro-geopolitical research firm advising hedge funds, family offices, sovereign wealth funds and governments. Roche was Head of Research and Global Strategist at J.P. Morgan and later at Morgan Stanley. In 1994 he founded Independent Strategy in London, building it into a leading independent macro research house, before later founding Quantum Strategy & Geonomics.
