Meta launched an enterprise platform with no pricing

Meta just told businesses it wants to be their AI vendor, and did not say what it costs. On September 28 Meta announced the Meta Enterprise Platform, pulling Meta Business Agent, Muse API and Muse Code into one commercial line and hiring a Chief Enterprise Platform Officer to run it. There is no pricing page, no availability date and no waitlist in the post. For anyone weighing Meta Enterprise Platform as a stack decision, that absence is the most useful fact in it.

What actually happened

The announcement is a leadership hire plus a strategy statement. Chirantan "CJ" Desai joins as Chief Enterprise Platform Officer reporting directly to Mark Zuckerberg. He comes from MongoDB, where he was CEO and President; before that he led product and engineering at Cloudflare and spent nearly eight years at ServiceNow. That is a genuinely serious enterprise résumé, and it is the clearest signal in the post — you do not hire that person to run an experiment.

The products named are Meta Business Agent, Muse API, Muse Code, "and more." Zuckerberg's framing is that superintelligence will create new opportunities for businesses, and that Meta already serves billions of people and helps hundreds of millions of businesses reach customers.

What the post does not contain: pricing, tiers, rate limits, availability timing, regions, a waitlist, SLAs, or which businesses can use any of it today. We checked. It is an org announcement with a product list attached, not a launch.

Worth holding next to it: Meta shut down its hosted Llama API public preview on July 6, 2026, and its developer-facing surfaces have moved more than once. That is not a prediction about this platform. It is the base rate you should price in for any vendor whose enterprise business is being stood up this quarter.

Why a new enterprise AI vendor matters for your business

A strategy post is not a product you can plan around. You cannot forecast a line item with no price, and you cannot commit a roadmap to an API with no availability date. Treat this as a signal to watch, not a decision to make. The decision arrives when pricing does.

The distribution is the real pitch, and it cuts both ways. Meta's advantage is that your customers are already inside Instagram, WhatsApp and Messenger. An agent that sells where people already are beats a better agent nobody opens. The catch is the same one every channel carries: the reach belongs to Meta, the terms change on Meta's schedule, and your customer relationship is mediated by someone else's product decisions.

Write down the exit before you write the integration. Before any of this ships, decide what you would need to keep if Meta repriced or retired it: your product catalog, your conversation logs, your customer identifiers, your order records. If those live only inside a vendor's agent runtime, you do not have a channel — you have a dependency.

New enterprise platforms are staffed thin at the start. A brand-new business unit means new docs, new support paths, new bugs and no institutional memory. That is fine for an experiment and bad for anything touching checkout. Pilot it on something reversible.

Compare it on portability, not on demos. The question to ask any agent platform vendor is not "how good is the model." It is: can I export my data, can I run the same logic somewhere else, and what happens to my integration when you ship v2. Ask it before the pricing page appears, because you will get a straighter answer.

Sources: Launching Meta Enterprise Platform, Meta Newsroom.