Mark Zuckerberg promises free AI for billions — then buries a bidding war in the fine print

But this manifesto wasn’t just about Zuckerberg’s AI business objectives — at least not overtly. Instead, he used this text as an opportunity to spell out his philosophy on AI and why he favors an open infrastructure to bring “personal superintelligence” to “billions of people.”

To bring about his AI-enriched future, Zuckerberg claims Meta will offer everyone a “personal agent” that works “24/7 on your behalf to improve your relationships, health, career, finances, home management, hobbies and more.”

But as you read deeper into his post, you may notice these services aren’t entirely free. Ironically, in the bullet point titled “Everyone will have free or affordable access to these tools,” Zuckerberg says, “For those who want to pay to use more compute, there will be a dynamic auction mechanism that will guarantee that everyone gets the lowest price possible for the intelligence and compute they’re using while also ensuring the capacity is used for whatever people collectively find most valuable.”

This bidding war highlights the harsh reality of hardware scarcity. Even though an open source model makes intelligence accessible, it doesn’t magically create more resources. People who are willing to pay the highest price for tokens will still have access to more superintelligence.

Zuckerberg admitted this later in his manifesto when he wrote, “No matter how intelligent AI becomes, there will always be a finite amount of compute and therefore an opportunity cost for how we use it.”

Moneywise reached out to Meta for further comment but didn’t hear back before publication.

Is Zuckerberg more practical than philosophical?

Although Zuckerberg’s manifesto talks a lot about big ideas like “individual empowerment as the source of prosperity,” the support for open-source AI models might make good business sense.

AI leaderboards like Artificial Analysis and LLM Stats consistently show Meta’s AI model Muse Spark simply isn’t as popular as those from closed-source competitors OpenAI and Anthropic. Meta is also struggling against rising open-source Chinese competitors like Moonshot AI’s Kimi 3.

With this backdrop in mind, Zuckerberg’s call for more openness might signal the area where Meta sees its opening in the AI race: A U.S.-based open-source model.

Coincidentally, Zuckerberg recently announced Meta will be open-sourcing its Muse Spark 1.2 model so anyone can download and modify it. Meta also rolled out its open-weight Muse Glimmer AI model, which has the advantage of running on devices like laptops rather than in the cloud.

Neil Shah, co-founder of the market intelligence firm Counterpoint Research, views this as a way for Meta to distinguish itself from other American AI labs. As Shah told CNBC, “Most of its competitors in [the] USA are proprietary and there is an insatiable demand for non-Chinese open models and weights, and Meta can fill in this void well.”

With reference to Glimmer, Shah saw yet another angle where Meta could compete with hyperscalers in terms of compute cost. As Shah reasoned, “Bringing small, agentic models like Muse Glimmer directly onto PC and mobile hardware bypasses cloud compute costs to outcompete Google, Microsoft and others on the end-user’s device.”

While it’s up for debate just how much Zuckerberg truly believes in his philosophical ideals, open sourcing is clearly playing a role in Meta’s AI strategy.