Meta explores cloud infrastructure business

Meta is reportedly exploring a new way to make money from its initial investment into data infrastructure, which it had flagged for future artificial intelligence projects.

According to Bloomberg, Meta is developing plans for a cloud infrastructure business, which would sell access to AI computing power and models to other providers, including Meta’s competitors in the AI race.

That would provide another revenue pathway for Meta, which is facing an increasingly costly AI push and has committed hundreds of billions of dollars in AI development spending over the next three years alone.

However, this move may also signal Meta’s flagging prospects when it comes to generating significant income with its own AI initiative. The company may have overspent itself, by a significant margin, in its efforts to become the leader in the AI race.

Last month, xAI announced a similar initiative, with company owner Elon Musk’s AI project renting compute capacity to both Google and Anthropic. The obvious concern there is that xAI has also clearly overspent on its AI datacenter development.

According to SpaceX’s IPO documentation, xAI has committed more than $20 billion through 2026 to expand its massive Colossus data center projects, with much of that already invested into its initial infrastructure to power up its AI capacity. Musk’s xAI will also have ongoing costs attached to its AI projects, which means the company has to find ways to claw back that outlay in order to drive a profit.  

The fact that xAI is now renting out server space to competitors, and looking to implement a range of third-party integrations, suggests that the business doesn’t have a clear path to boosting intake based on the merits of its own AI offerings.

Does that mean that xAI is losing the AI race overall? Well, right now, it does seem to be falling behind, and given its massive costs, the company could become an albatross that weighs on SpaceX’s market performance. That is, assuming xAI can’t find more ways to offset those costs.

Meta may now be in a similar boat. The company has committed more than $600 billion to AI infrastructure projects over the next three years, far more than xAI and other competitors, in its race towards AI dominance.

But with market sentiment slowly souring on both AI and the usefulness of AI tools, maybe Meta is now reassessing that outlook. Meta is also facing problems with its advanced superintelligence project, which it hopes will eventually uncover the next level of AI processing, and give the company a clear advantage in the market.

According to CNBC, the initial offerings from Meta’s advanced AI lab have failed to spark significant market interest. At the same time, statements from Meta’s star AI hire Alexandr Wang haven’t inspired much confidence in future opportunities, with Wang offering only tentative projections on advancement.

This could explain why Meta is now looking to rationalize and commoditize whatever aspects of its AI tools that it can. The company is also looking to charge consumers for advanced AI tools within its apps.

Will that be enough? Will Meta be able to reduce its financial liabilities through subscription offerings and data center deals, in order to derive value from its AI push, even if the technology doesn’t catch on like Meta had initially hoped?

There are a lot of sunk capital costs to recover, and Meta will likely still need to drive significant take-up of its paid AI tools to eventually turn a profit from this.