OpenAI generated about $5.7 billion in revenue during the first quarter of 2026, according to a report from The Information out Wednesday. That is roughly $1 billion ahead of Anthropic's $4.8 billion in the same period, and OpenAI's strongest quarter on record. The gap is the first time OpenAI has clearly outpaced Anthropic on quarterly revenue in eighteen months — and the line items that drove it tell a sharper story than the topline.
What carried the quarter
Three buckets did most of the work: ChatGPT subscription growth, the enterprise API line, and Codex. OpenAI's coding product line has scaled rapidly through Q1, both as a standalone offering and as the embedded engine inside agentic IDE features at customer partners. The Codex contribution to API revenue is what the prior two quarters were missing — it is the line item that closed the gap with Anthropic on developer workloads, where Claude had been the default frontier choice through most of 2025.
Why Anthropic is not closing it
Anthropic's $4.8 billion is a great quarter on every absolute measure — and the company is reportedly on track for its first profitable quarter, with Q2 revenue projected to more than double. The bottleneck is not demand. The bottleneck is compute. Anthropic's enterprise contracts have outrun the inference capacity it has under contract, and the rate-limit cycles on Claude (raised in May after the SpaceX deal cleared) are the symptom. Enterprise customers do not want to be the one holding the bag when their model provider runs out of GPUs in the middle of a quarter. OpenAI's compute story, anchored by the Stargate buildout and the Microsoft envelope, has been steadier through Q1 — and that steadiness shows up in the topline.
The supplier scramble
The other thing the gap reveals is how aggressively the chip layer is restructuring around inference demand. Microsoft is in talks to sell Anthropic Maia 200 silicon on top of the existing $30 billion Azure compute deal. Amazon and Google are already in the mix with Trainium and TPU capacity. Anthropic is now actively running a multi-vendor chip strategy because the single-vendor version of the strategy did not produce enough capacity fast enough. Every additional chip supplier expands the inference ceiling — and pulls Anthropic's gross margin in a different direction depending on which one wins the workload at any given moment.
The IPO read
The Q1 numbers land directly into OpenAI's IPO push, where Kalshi prediction markets currently price an 83% chance the company beats Anthropic to public markets. A $5.7 billion quarter — implying a $22.8 billion annualized run rate — is the kind of number that drops cleanly into an S-1. Anthropic is unlikely to be far behind in absolute terms, but the visible compute story has to clear up before its prospectus reads as cleanly. The race is no longer about who has the best model. It is about who has the most disciplined balance sheet and the most diversified compute supplier mix when public markets get to read the financials.
Comments