OpenAI and Anthropic outearn every Chinese AI model combined by 10x

verifiedFinanceConsulting

Thursday, September 17, 2026

Confidence

High · — Rhodium primary + CNBC, Quartz corroborating

Evidence

named research group + Tier-1 reporting

Rhodium Group estimated that all of China's AI models combined generate only about 10% of the revenue reported for OpenAI and Anthropic.

The details

OpenAI ARR $40B, Anthropic $65B; ByteDance leads China at $4B.
Rhodium put Moonshot at 50x, DeepSeek 163x revenue, versus OpenAI's 34x.

What it means

The "China is catching up" frame breaks in dollars: parity without monetization is a financing problem, not a competitive one. Price China-origin models as scaling risk, not discount leverage.

Rhodium Group estimated that all of China's AI models combined generate only about 10% of the revenue reported for OpenAI and Anthropic.

OpenAI ARR $40B, Anthropic $65B; ByteDance leads China at $4B.
Rhodium put Moonshot at 50x, DeepSeek 163x revenue, versus OpenAI's 34x.

Sources

Moonshot at 50x, DeepSeek 163x ↗

Apply this today

The hands-on layer the brief points to: a workflow you can run, a tool to test, and today’s 60-second video.

Workflow of the day

Draft a One-Third Workforce Transition Brief: Map AI-Displaced Roles to 2029 Rehire Pathways

Human Resources · tested · saves 3-4 hours per role review cycle

Full walkthrough →

Tool watch

Gemini 3.8 Live and 3.8 Live Extended Thinking

Gemini 3.8 Live for the agent loop: does interleaved reasoning actually help?

Read the breakdown →

▶ Watch today’s 60-second brief →

In this issue

← Back to brief

Get the next one in your inbox

One daily brief. Every story gets a hype verdict.

No spam. Unsubscribe anytime.

On this page

The details
What it means

In this issue

← Back to brief

Stat of the Day

$105B▲

combined ARR: OpenAI + Anthropic

Rhodium's tally: $40B OpenAI plus $65B Anthropic, versus ~$10B across every measured Chinese AI lab combined. Source

Today’s Take

Three of today's five stories are the same story: the AI industry is trying to lock in the terms of its own oversight while it still has the capital and the revenue lead to set them. OpenAI's framework, Apollo's willingness to lend against Vision Fund 2 marks, and Rhodium's 10x revenue gap all point the same direction — the US frontier labs are self-regulating from a position of strength, not weakness. The commercial layer (HubSpot, Mastercard) is following: agents get invoice-chasing rights and payment rails before regulators finish debating whether they should have them. The bet worth tracking is whether voluntary disclosure survives its first non-voluntary incident. Considered and passed: the King Charles safety summit (symbolic, no binding output) and the Suleyman-Anthropic model-rights argument (podcast-clip sourcing, no peer review).

— Agentic desk

Never miss a signal

Join professionals getting the daily AI brief every weekday morning.

No spam · Unsubscribe anytime

No sponsored verdicts
Sources linked
Corrections public