Enterprise is OpenAI’s center
OpenAI says enterprise customers now make up roughly 40% of its revenue, signalling a clear shift toward managed, business-focused offerings. At the same time, cloud providers like AWS are funding multiple model makers and encouraging customers to keep options open, which suggests enterprises will buy reliability, governance and integration rather than a single best model. That commercial shift gives buying organisations more leverage to demand features like permissioning, SLAs and deployment controls as part of model contracts. (cnbc.com) (techcrunch.com)
OpenAI’s sales mix is starting to look less like a consumer app and more like enterprise software: the company said business customers now generate more than 40% of revenue, and it expects that share to reach parity with consumer by the end of 2026. (cnbc.com) (openai.com) That is a sharp change for a company that became famous through ChatGPT subscriptions and viral demos. In January, OpenAI executives were already telling CNBC that enterprise was roughly 40% of the business and heading toward about 50% by year-end. (cnbc.com 1) (cnbc.com 2) OpenAI has been building around that shift with products aimed at large organizations, not just individual users. In February it launched Frontier, an enterprise platform CNBC described as a direct push to win more business customers. (cnbc.com) It also started recruiting the companies that already sell into boardrooms. CNBC reported in February that OpenAI signed deals with Accenture, Boston Consulting Group, Capgemini, and McKinsey to help big clients deploy its tools inside real workflows. (cnbc.com) OpenAI’s own numbers show why that market is attractive. The company said this week that it has more than 1 million business customers, that its application programming interfaces process more than 15 billion tokens per minute, and that customers now include Goldman Sachs, State Farm, Thermo Fisher, DoorDash, and LY Corporation. (openai.com 1) (openai.com 2) At the same time, Amazon Web Services is signaling that enterprises should not bet the company on one model vendor. TechCrunch reported on April 8 that Amazon Web Services chief executive Matt Garman defended backing both Anthropic and OpenAI, saying Amazon is used to partnering with companies it also competes with. (techcrunch.com) That matters because Amazon Web Services makes money when customers stay on Amazon Web Services, even if they switch models. Its Bedrock service is built around offering a menu of foundation models, and Amazon Web Services documentation says Bedrock provides access to models from Anthropic, Meta, and OpenAI through one managed layer. (aws.amazon.com) (docs.aws.amazon.com) Once buyers can compare models inside the same cloud console, the sale stops being only about which chatbot sounds smartest in a demo. It starts looking more like every other enterprise software purchase, where procurement teams ask about security controls, uptime commitments, administrator permissions, and how cleanly the tool plugs into existing systems. (openai.com) (aws.amazon.com) OpenAI has been leaning into exactly that language. Sarah Friar wrote in January that 2026 would be the company’s year of “practical adoption,” and the company’s enterprise materials frame artificial intelligence less as a novelty and more as core infrastructure inside large firms. (cnbc.com) (openai.com) So the balance of power is shifting a little toward the customer. If OpenAI wants enterprise revenue to catch consumer revenue, and Amazon Web Services is telling clients to keep options open, large buyers can push harder for contract terms that look like normal software deals instead of research-lab access: service-level agreements, deployment controls, audit trails, and clearer permissioning. (cnbc.com) (techcrunch.com) (openai.com)
