OpenAI shifts to enterprise
OpenAI is moving focus away from free consumer features toward paying business customers, a shift reported this week as the company rolls out a new pro model and deprioritises some consumer efforts. At the same time OpenAI has reportedly agreed to pay Cerebras more than $20 billion over three years for access to servers powered by Cerebras chips, a move that ties go‑to‑market strategy to large infrastructure commitments. (semissourian.com) (stratnewsglobal.com)
OpenAI is steering more of ChatGPT toward paying workplaces as it chases profit and signs up huge new computing capacity. (pressdemocrat.com) Chief financial officer Sarah Friar told the Associated Press that business customers now generate about 40% of OpenAI’s revenue, up from 20% when she joined in 2024. She said that share is on track to reach 50% by the end of 2026. (pressdemocrat.com) OpenAI said it will introduce a model aimed at “high-value professional work” as it competes with Anthropic for corporate customers. The company has also been shedding some consumer offerings while expanding business plans and admin tools. (fastcompany.com) The business push now reaches pricing, controls and product design. OpenAI’s Business plan starts at $20 per user a month with annual billing, while Enterprise adds custom retention policies, encryption, service-level agreements and sales-led contracts. (openai.com) On April 2, 2026, OpenAI added flexible pricing and a new Codex-only seat for ChatGPT Business and ChatGPT Enterprise. Those plans now let companies buy shared credits for heavier use of tools like deep research, image generation, advanced voice and coding features. (help.openai.com) That shift lines up with how OpenAI describes deep research itself. In its product note, the company said the tool is built for finance, science, policy and engineering work, and on February 10, 2026 it added connectors and trusted-site controls for enterprise-style research workflows. (openai.com) The company is also locking in the infrastructure needed to serve bigger customers. Reuters reported on April 16 that OpenAI agreed to pay Cerebras more than $20 billion over three years to use servers powered by the startup’s chips, in a deal that could also give OpenAI an equity stake. (usnews.com) The Cerebras deal, first reported by The Information, would reduce OpenAI’s dependence on Nvidia-powered systems at a moment when demand for AI computing still exceeds supply. That makes the company’s sales push and its server strategy part of the same bet: more enterprise demand, backed by more dedicated capacity. (theinformation.com)
