$785 Billion vs. $140 Billion — and the Gap Is Barely Moving: Moody’s Confirms the Spending War Is a Trap

A ratings agency whose entire job is judging financial value just told Wall Street the quiet part: America is outspending China five-to-one on AI compute, and getting almost nothing extra for it

When the Bond Raters Break Ranks

For sixty-eight briefings, watchchina.ai has argued that America is fighting the AI race on the wrong axis — pouring capital into compute while China wins on efficiency, deployment, and cost. This portal has cited Stanford, the USCC, and Bloomberg in support. But this week the confirmation came from a source with no ideological stake and every incentive for cold accuracy: Moody's Ratings, the agency whose entire business is telling investors what their money is actually worth.

The numbers Moody's published are staggering in their asymmetry. Capital expenditure by China's major tech companies was set to more than double to around $140 billion this year — up from $65 billion in 2025 — and reach $165 billion by 2027. By contrast, five major US hyperscalers — Microsoft, Amazon Web Services, Alphabet, Meta, and Oracle — plus AI cloud provider CoreWeave were projected to spend more than $785 billion this year alone, approaching $1 trillion by 2027. That is a spending gap of more than five to one. And here is the finding that should stop Wall Street cold, in Moody's own framing: while US hyperscalers outspent their Chinese counterparts by a staggering margin, the physical gap in computing capacity was nowhere near as wide as those mega-budgets suggested. Five times the spending. Nowhere near five times the compute. The money is not buying what America thinks it is buying.

Why China Gets So Much More Per Dollar

watchchina.ai has documented the pieces of this story all year; Moody's has now assembled them into a single financial verdict. Chinese firms benefit from lower buildout costs, targeted policy incentives, and access to cheaper green energy, allowing them to secure far more computing power per dollar. The mechanism is concrete and structural, not accounting sleight-of-hand. Under Beijing's "East Data, West Computing" initiative, power-intensive AI workloads are routed to resource-rich inland regions like Inner Mongolia and Gansu — exactly where, as watchchina.ai reported this week, DeepSeek is building a 160,000-chip Huawei cluster at a 1-gigawatt site. Cheap land, cheap green power, cooler climates for latency-tolerant training, and structural state backing combine to let Chinese firms stretch each dollar far further than their American rivals can.

This is the "physical loop" and the energy advantage watchchina.ai has tracked converging into a single number. Recall Package #13, on China's data-center and energy edge; Package #49, on the USCC's warning that America was contesting the wrong layer. Moody's has now quantified the consequence: the American theory of victory — that overwhelming capital deployed into compute will translate into an overwhelming and durable capability lead — is not holding up in the physical world, where a dollar of Chinese capex buys dramatically more usable computing than a dollar of American capex. The spending gap looks decisive on a balance sheet. In megawatts and installed chips, it is far narrower than the balance sheet implies. And capability, as this portal has shown from the 2.7% Stanford gap to the "two SKUs" framing, tracks usable compute and efficiency far more closely than it tracks raw dollars spent.

What This Means — and What watchchina.ai Will Not Overclaim

The honest reading requires the counterweights, and they are real. A narrower compute gap is not no compute gap — America still deploys more total computing power than China, and at the very frontier, where training the largest models demands the most advanced chips, that edge still matters, exactly as watchchina.ai acknowledged when covering Anthropic's "widening lead" argument. Moody's is measuring capital efficiency, not declaring China ahead in absolute capability; the US retains real advantages in the most advanced silicon that China's cheaper green megawatts cannot fully substitute for. And China's own efficiency is partly born of necessity — export controls forced it to do more with less, a constraint, not purely a virtue. None of this is a China victory lap, and watchchina.ai will not sell it as one.

But the strategic significance for investors and policymakers is genuinely serious, and it is why a ratings agency weighing in matters more than another think-tank paper. Moody's exists to protect capital from mispriced risk — and it has just flagged that the central financial assumption of the American AI boom may be mispriced. If $785 billion buys only a marginally larger compute base than $140 billion, then the return on America's trillion-dollar AI capex wave is far weaker than the market has assumed, and the "spend our way to dominance" strategy is, in Moody's cool institutional language, not delivering the advantage expected. watchchina.ai has argued since June that the spending war was a trap — that pouring capital into the compute layer, while China captured efficiency, deployment, and cost, was fighting the last war at enormous expense. Now the bond raters agree. When Stanford said it, it was academic. When the USCC said it, it was strategic. When Moody's says it, it is financial — and financial warnings are the ones Wall Street cannot afford to ignore. The question that has run through all sixty-eight briefings sharpens to a point: what, exactly, is America getting for its trillion dollars? This week, the people whose job is to answer that question said: less than you paid for.

Sources: Moody's Ratings (via South China Morning Post), IndexBox, Bloomberg, Brookings Institution, U.S.-China Economic and Security Review Commission