OpenAI Quarterly Revenue Hits $10 Billion as Enterprise Push Narrows Gap with Anthropic
The AI leader's third-quarter performance reflects sharper pricing, model improvements and a pivot toward business customers, though it still trails its rival in annualised sales.
Revenue Acceleration Marks Strategic Shift
OpenAI's quarterly revenue surpassed $10 billion for the first time in the third quarter of 2026, according to figures cited by people familiar with the matter. The milestone represents a 49 per cent rise from the $6.7 billion the company recorded in the second quarter, a marked acceleration from the 18 per cent growth it posted between the first and second quarters.
The jump reflects a combination of tighter pricing discipline, selective discounts for large customers, and what people close to the company describe as stronger enterprise adoption of its latest models. OpenAI's operating margin also narrowed, moving from negative 150 per cent in Q2 to negative 109 per cent in Q3, indicating that revenue growth is beginning to outpace the rate at which costs are climbing.
The Enterprise Lag and the Anthropic Challenge
For much of 2026, OpenAI's revenue trajectory had been robust in absolute terms but modest relative to Anthropic, which reported third-quarter revenue of $17 billion and has captured a disproportionate share of enterprise deals. Anthropic's Claude models have become the default choice for companies seeking to embed large language models into regulated workflows, particularly in financial services, healthcare and legal sectors where auditability and fine-grained control matter.
OpenAI's early strategy prioritised reach over segmentation. The company accumulated more than one billion users, but the bulk of that base consists of free-tier accounts and subscribers to its lower-priced consumer plans. That breadth delivered brand recognition and data scale, but it left OpenAI underweight in the higher-margin enterprise segment where Anthropic has been able to command premium pricing and longer contract terms.
Pricing, Models and the Agent Bet
OpenAI has responded by recalibrating its go-to-market approach. The company introduced tiered pricing structures that reward volume commitments, rolled out enterprise-grade features such as dedicated capacity and stricter data residency guarantees, and began offering time-limited discounts to anchor large customers. At the same time, improvements in model efficiency have allowed OpenAI to serve inference requests at lower unit cost, widening the margin between what it charges and what it spends.
The third element of the strategy centres on agents: software that can execute multi-step tasks with minimal human intervention. OpenAI has invested in building agent frameworks that integrate with enterprise software stacks, positioning itself not merely as a model vendor but as a platform for automating knowledge work. Early adopters in customer support, legal research and software development have reported productivity gains, and those use cases are now forming the basis of OpenAI's pitch to chief information officers.
Annualised Rates and the Accounting Nuance
On an annualised basis, OpenAI's revenue run rate now stands at $50 billion. Anthropic, which announced plans for an initial public offering at a $1.4 trillion valuation, reported an annualised run rate of $81 billion at the end of October, up $10 billion in a single month.
The two companies calculate annualised revenue differently. Anthropic includes sales made through its cloud distribution partners, such as Amazon Web Services and Google Cloud, in its reported figures. OpenAI, by contrast, counts only its share of revenue from partner channels, excluding the portion that flows directly to the cloud provider. The methodological difference makes direct comparison imperfect, but it does not alter the broader picture: Anthropic remains ahead on top-line growth, while OpenAI is closing the gap through a sharper enterprise focus.
A Crowded and Fragmenting Market
Both companies operate in an environment that has become significantly more competitive over the past year. Meta continues to release open-weight models that enterprises can deploy without per-token charges. Amazon, Microsoft and Google have integrated frontier models into their cloud platforms and are subsidising usage to lock in customers. Meanwhile, a wave of capable models from Chinese labs, many of which are open source and cost a fraction of Western alternatives to run, has begun to erode pricing power at the lower end of the market.
The result is a bifurcating landscape: high-value enterprise deals still command premium pricing, but commodity inference work is rapidly becoming a low-margin business. OpenAI's ability to sustain its current growth rate will depend on how effectively it can defend its position in the former category while managing cost pressure in the latter.
What the Margin Trajectory Signals
OpenAI's improving operating margin, while still deeply negative, suggests that the company is beginning to extract leverage from scale. The narrowing from negative 150 per cent to negative 109 per cent indicates that each incremental dollar of revenue now requires a smaller incremental dollar of spending, a pattern consistent with improving model efficiency, better capacity utilisation and the early benefits of enterprise pricing.
Still, the path to profitability remains long. Training runs for the next generation of models are expected to cost billions of dollars, and competition for talent in machine learning engineering continues to push compensation higher. OpenAI's ability to reach break-even will hinge not only on revenue growth but on whether it can stabilise the cost structure of inference and amortise training expenses across a larger revenue base.
The third-quarter numbers mark a turning point in OpenAI's commercial trajectory, but they also underscore the challenge of competing in a market where technical leadership, distribution scale and capital efficiency all matter, and where no single player yet dominates all three dimensions.
