Anthropic's $1.4 Trillion IPO Meets a 10-to-1 U.S.-China AI Revenue Gap

America's two leading AI labs collect roughly $105 billion in annualized revenue. China's six largest AI companies collect about $11 billion combined. That estimate, published by Epoch AI in its 2026 comparison of run rates, lands days before Anthropic's IPO on Nov. 19 at a proposed valuation of $1.4 trillion.

The figures are blunt. Epoch AI credits Anthropic with approximately $65 billion and OpenAI with $40 billion. The Chinese side counts ByteDance, Alibaba, Z.ai, Moonshot AI, DeepSeek and MiniMax. The U.S. pair runs about 9.5 times the Chinese group's pace, a gap of roughly $94 billion. Anthropic alone generates nearly six times China's combined total.

What does the gap mean for the valuation?

The comparison, if anything, understates Anthropic's position. The company's own IPO disclosures put annualized revenue at approximately $81 billion by October. The estimates draw on different reporting periods and cover only these companies, so they leave out much of the AI revenue earned in each country.

Even with that caveat, the numbers show how far ahead U.S. labs are at selling AI. Enterprises pay for American frontier models at scale, and the distribution channels keep widening. On Amazon's third-quarter call, CEO Andy Jassy said Bedrock token volume is growing nearly 17x year-over-year, and Fortune 500 companies are spending nearly 7x more on the service. He tied that demand directly to the labs: "We have both closed models and models like Anthropic and OpenAI's models." AWS backlog stands at $578 billion.

China's leaders are growing, but from a smaller base. Alibaba's AI cloud and compute revenue rose 45% year-over-year, and Omdia ranks it first in China's AI cloud market with a 38.1% share. The capital behind each side shows the same imbalance. Alibaba spent $9.97 billion on capex in its latest quarter. Amazon spent $77.26 billion in its quarter and repeated full-year 2026 cash capex of approximately $220 billion for AI infrastructure.

Why the revenue lead could keep widening

Anthropic heads to market with the largest revenue base in the comparison, and that base kept growing after Epoch AI compiled its figures. The jump from $65 billion to $81 billion by October shows sales compounding rather than flattening.

The infrastructure behind U.S. AI is scaling to match. Jassy said Amazon has "very good visibility into 2027 and 2028 and some of 2029 too" on large customer capacity commitments. Alphabet's Google Cloud revenue rose 103% year-over-year to $30.71 billion. Both figures point to broad demand from paying customers, and the expansion behind that demand — power, cooling, networking — is where a large share of the next wave of AI suppliers will emerge.

Chinese rivals, meanwhile, work under export controls that Anthropic's own filing lists among its regulatory risk factors. Recent chip-access concerns have pushed down Chinese AI stocks.

Can a $1.4 trillion price be earned?

A valuation of that size requires sustained growth, better unit economics and a durable edge. The costs remain steep. Amazon posted a trailing free cash outflow of $42.4 billion, and its CFO said AI margins are "still lower" than core AWS margins. Buyers in the IPO will also own a public benefit corporation supervised by a Long-Term Benefit Trust, a structure that requires the company to weigh shareholder returns against its safety mission.

The Nov. 19 debut will test how much investors will pay for America's AI revenue lead. The $105 billion to $11 billion gap gives that lead a hard number, and Anthropic's October run rate suggests the number is still moving in one direction. Whether the company can keep growing fast enough to justify a price built on a lead of more than nine times China's combined total is the question the market will answer.

Amazon and Alphabet shareholders already own a piece of the infrastructure behind it. The rest of the market gets its turn on Nov. 19.