Enterprise AI Is Now Real
Enterprise spending on AI has moved beyond pilots into meaningful revenue streams, with vendors landing large business customers and budget line-items shifting. OpenAI says enterprise now accounts for about 40% of its revenue, and investors and analysts note adoption is concentrating on coding, legal, support and healthcare workflows where ROI is immediate. That shift matters because procurement now evaluates inference economics, deployment options and vendor resilience — not just proof-of-concept demos. (cnbc.com) (x.com)
The giveaway is not the chatbot demos. It is that chief financial officers are now carving out permanent budget lines for artificial intelligence, and OpenAI said on April 8 that enterprise customers make up more than 40% of its revenue. (openai.com) (cnbc.com) That is a different phase from 2023 and 2024, when many companies were still running pilot projects out of innovation budgets. Andreessen Horowitz said in a June 2025 survey of 100 chief information officers that artificial intelligence spending had moved into recurring core information technology and business-unit budgets. (a16z.com) Once a tool gets a recurring budget, it stops being a science fair project and starts being judged like payroll software. The same survey said procurement now looks more like traditional software buying, with harder checks on hosting, benchmarks, and switching costs. (a16z.com) That is why the early winners are not random. Investors and operators keep pointing to coding, customer support, legal work, and healthcare administration because those jobs have clear before-and-after math on hours saved, tickets closed, or documents reviewed. (a16z.com 1) (a16z.com 2) Coding went first because the output is easy to measure. OpenAI said Codex reached 3 million weekly active users, which gives companies a straightforward way to compare subscription cost against engineer time and software delivery speed. (openai.com) The vendor mix is also getting narrower, not wider. Andreessen Horowitz said OpenAI, Google, and Anthropic took the dominant overall market share in its 2025 chief information officer survey, while its January 2026 enterprise ranking said OpenAI still led large companies and Anthropic and Google were gaining fast. (a16z.com 1) (a16z.com 2) That concentration changes what buyers ask in meetings. Instead of “can your model write a nice email,” companies now ask where the model runs, what it costs at scale, whether they can swap vendors later, and how the system connects to internal data without breaking permissions. (a16z.com) (cnbc.com) OpenAI’s answer is to sell more than a chatbot. Its Frontier platform, launched on February 5, is designed as an “intelligence layer” that connects internal systems, ticketing tools, and data warehouses so agents can work across a company instead of inside one isolated window. (cnbc.com) The consulting firms circling this market are another clue that the money is real. On February 23, OpenAI announced multiyear partnerships with Accenture, Boston Consulting Group, Capgemini, and McKinsey to push those systems into production workflows, which is the kind of rollout companies use for enterprise resource planning software, not experiments. (cnbc.com) Even the revenue split tells the story. At Davos in January, CNBC reported that enterprise was roughly 40% of OpenAI’s business and 80% of Anthropic’s, which means the race is shifting from consumer buzz to long contracts with companies that care about reliability and procurement discipline. (cnbc.com) The next fight is not over who can make the flashiest demo. It is over who can survive a bank’s security review, fit inside a hospital’s workflow, price inference cheaply enough for daily use, and stay dependable after the pilot team hands the project to the rest of the company. (openai.com) (a16z.com)
