Enterprise AI Turns Practical
Generative AI is moving from experimental pilots into budgeted enterprise procurement as firms buy capabilities, not just demos. Survey and market signals show widespread adoption in coding, legal and healthcare while OpenAI says enterprise revenue already accounts for roughly 40% of its sales — a shift that’s intensifying competition between vendors like OpenAI and Anthropic. (Where Enterprises are Actually Adopting AI) (OpenAI will allocate IPO shares to retail investors as it preps for debut, CFO says) (Anthropic may have closed the revenue gap on OpenAI. Here's what it means for their IPOs | Reuters)
Three years after ChatGPT launched in November 2022, large companies are no longer treating generative artificial intelligence like a science fair project. Andreessen Horowitz says 29% of the Fortune 500 and about 19% of the Global 2000 are already live, paying customers of a leading artificial intelligence startup, not just pilot users. (a16z.com) That changes the buying process. Andreessen Horowitz says its count only includes companies that signed a top-down contract, converted a pilot, and went live inside the organization, which is much closer to normal software procurement than to an innovation lab demo. (a16z.com) The first place companies are spending real money is coding. Andreessen Horowitz found that coding, customer support, and search make up the biggest enterprise use cases, with coding standing out by an order of magnitude over the rest. (a16z.com) The industry pattern is just as specific. Andreessen Horowitz says technology, legal, and healthcare are adopting fastest, which makes sense because those fields are full of expensive text work like drafting contracts, summarizing records, and answering internal questions. (a16z.com) OpenAI put a hard number on the shift this week. Chief Revenue Officer Denise Dresser said enterprise now makes up more than 40% of OpenAI’s revenue and is on track to reach parity with consumer revenue by the end of 2026. (openai.com) (cnbc.com) OpenAI is also describing what companies want next. In its April 8 post, the company said customers are tired of isolated artificial intelligence tools that do not connect to each other and are instead asking for a single operating layer tied to internal systems, outside data, permissions, and controls. (openai.com) That is why the fight is moving away from flashy chatbots and toward workflow control. OpenAI says it is pushing products like ChatGPT Enterprise, Codex, and its Frontier platform so companies can use one vendor for models, tools, and the employee interface. (openai.com) The pressure on rivals is rising because enterprise customers are the stickiest customers in software. CNBC reported on April 8 that OpenAI was valued at $852 billion in its latest financing round, and the company is already preparing itself to “look and feel” like a public company ahead of an eventual initial public offering. (cnbc.com) Anthropic is not sitting still. Reuters reporting summarized in multiple outlets says Anthropic has rapidly narrowed the revenue gap with OpenAI, and Bloomberg reported on April 8 that employee demand to hold shares ahead of a possible initial public offering stayed strong as Anthropic’s annualized revenue climbed. (bloomberg.com) (reuters.com) The practical result inside companies is simple: budgets are moving to the teams that can show work getting done faster. When coding tools write software, legal tools review documents, and healthcare tools summarize records well enough to survive procurement, generative artificial intelligence stops being a bet on the future and starts looking like ordinary enterprise software. (a16z.com) (openai.com)
