$2 trillion. That is the rough size of the AI financing ecosystem that analysts at ZeroHedge argue is now built on a single shaky assumption: that Anthropic and OpenAI will honour commitments totalling more than $1.5 trillion in cloud spending and infrastructure deals. If either firm stumbles, the circular logic holding the whole structure together starts to unravel fast.
How the loop works, and where it breaks
The mechanics are straightforward enough. Hyperscalers pour capital into Anthropic and OpenAI at sky-high valuations. Those companies then funnel much of that cash back to the same hyperscalers as cloud compute bills. Investors price the entire chain on the assumption that both firms will keep growing revenue fast enough to justify the cycle. Pull one thread and the rest follows.
The thread being pulled right now is competitive pressure from Chinese large language models. On platforms like OpenRouter, Chinese-origin models have already consumed a significant slice of total token traffic, a metric that directly reflects real-world adoption rather than benchmark scores or press releases. That kind of market-share erosion chips away at the monetisation story that underpins Anthropic's and OpenAI's valuations in the first place.
What the prediction markets are actually saying
Markets still assign a strong probability to Anthropic reaching a $1.25 trillion valuation by the end of this year, but the risk premium attached to that outcome has visibly widened. The gap between the base case and the tail risk is what traders are now pricing in, and it reflects a genuine shift in how the competitive landscape is being read.
Chinese open-weight models carry a structural advantage that is easy to underestimate. Because they are open, companies can self-host them, cutting out the API fees that feed Anthropic's and OpenAI's revenue lines. An enterprise that routes even 30 percent of its workload through a self-hosted Chinese model is, by definition, not paying for Claude or GPT-4o tokens. Multiply that across thousands of enterprise clients and the revenue assumptions baked into those $1.5 trillion commitments start looking optimistic.
The broader AI investment frenzy has attracted enormous capital on the premise that American frontier labs would dominate commercial deployment globally. The rise of capable, freely available alternatives does not kill that thesis overnight, but it does compress the timeline in which Anthropic and OpenAI need to prove their monetisation models actually work at scale. Neither company has yet demonstrated the kind of revenue run-rate that would make a $1.25 trillion valuation look conservative rather than aspirational.
Strategic responses from both firms, whether through new funding rounds, exclusive enterprise partnerships, or accelerated product launches, will be the key signals to watch in the months ahead. So far, the public positioning from both camps has been confident. The prediction markets, quietly, are a little less sure.
This article is for informational purposes only and does not constitute financial or investment advice.
