Meta launches Meta Enterprise Platform to sell AI

The new unit is Meta's first formal push to sell its AI models and agents directly to companies, rather than leaving chatbot and agent features inside its consumer products. Founder Mark Zuckerberg has argued that Meta holds advantages few rivals can match in advanced models, leading agents, and massive infrastructure — a case he will now have to prove with actual enterprise customers. The leader he picked, Chirantan Desai, brings a software-tools background from MongoDB to a unit that must compete with well-established coding assistants like Claude Code, Codex, and Cursor, as well as low-cost open-weight models from China.

The stakes are shaped by how much Meta has already committed to AI. Citing the Wall Street Journal, the article notes Meta is spending more than $100 billion on AI infrastructure this year alone, and it says investors want to see a return on that outlay. The enterprise push looks like a direct response to that pressure, though Meta has not yet said how the business will actually work or what it will charge for the services it plans to sell.

Much therefore still rests on details Meta has not disclosed. If the unit can turn models and agents into a priced, repeatable offering, it could give Meta a way to recover some of its infrastructure costs; if not, it may remain a side effort next to the consumer products that drive its business. For now, the launch is a signal of direction rather than a settled model, and the pricing and packaging choices due to come will show how seriously Meta intends to compete for enterprise budgets.