- OpenAI’s revenue grew 18% sequentially from Q1 to Q2 2026, reaching $6.7 billion in the three months ended June — a solid absolute growth rate that nonetheless disappointed some investors who had expected faster acceleration given the scale of AI demand signals in the market and the company’s dominant brand position with ChatGPT.
- OpenAI’s operating margin worsened further into negative territory during Q2, pushing the company farther from profitability ahead of its much-anticipated IPO — a particularly unfavorable data point given that deep losses at scale, combined with slowing sequential growth, are exactly the combination that tends to trigger valuation compression in public market comparisons.
- The contrast with Anthropic is stark: while OpenAI grew 18% sequentially to $6.7 billion in Q2, Anthropic recently disclosed an annualized revenue run rate topping $65 billion — implying quarterly revenue well above OpenAI’s — making Anthropic’s revenue trajectory appear significantly more dynamic even as it remains private and pre-IPO itself.
- OpenAI told investors its growth accelerated in Q3, a forward-looking claim designed to reassure shareholders that the Q2 results understate the company’s current momentum — but the acceleration is yet to be verified by reported financials, and the company’s credibility on forward guidance will be closely scrutinized given that its Q2 performance already came in below some investor expectations.
What Happened?
OpenAI disclosed to investors that its second-quarter revenue reached $6.7 billion, an 18% increase from $5.7 billion in Q1 2026. While 18% sequential quarterly growth would be considered exceptional for most companies, the result disappointed some OpenAI investors who had hoped to see faster progress given the explosive demand for AI services and the company’s position as the category’s defining brand. More concerning was the trajectory of losses: OpenAI’s operating margin sank further into negative territory during Q2, meaning the company is burning more cash relative to revenue even as it scales — the opposite dynamic investors hope to see as a business matures. The company attempted to counter the bearish read by telling investors that growth had accelerated in Q3, though those figures have not been independently verified.
Why It Matters?
The OpenAI Q2 figures land in a context that makes them look worse than the headline numbers suggest. Anthropic — OpenAI’s most direct competitor, also pre-IPO — recently disclosed an annualized revenue run rate exceeding $65 billion, which implies current quarterly revenue at a pace that appears to substantially exceed OpenAI’s $6.7 billion Q2 result. The juxtaposition is damaging to the narrative that OpenAI, as the pioneer of the ChatGPT moment, will inevitably command the AI market’s largest revenue share. Investors heading into an OpenAI IPO will now need to weigh whether the company’s first-mover brand advantage translates into durable revenue leadership, or whether Anthropic’s enterprise-focused go-to-market is winning the institutional AI budget more effectively. The widening loss margin adds another complication: at a high valuation, an investor buying OpenAI at IPO is paying for future profitability that appears to be getting further away, not closer.
What’s Next?
All eyes will be on OpenAI’s Q3 results to validate the company’s claim of accelerating growth. If Q3 revenue comes in materially above $6.7 billion and the loss trajectory begins to flatten, the IPO narrative can be rebuilt around a re-acceleration story. If Q3 growth disappoints or losses continue to deepen, the company will face harder questions about its unit economics and competitive positioning ahead of what would be one of the most scrutinized public market debuts in years. The IPO timing is also a variable: in a market where 30-year Treasury yields are at 19-year highs and long-duration growth asset valuations are under pressure, OpenAI’s window for a favorable public debut is narrowing with each quarter of sub-expectation results.
Source: The Wall Street Journal
