Meta is preparing to sell computing power directly to enterprise customers, entering a cloud market where chip spending could hit $265.8 billion by 2029.
Quay lại
Meta enters $265B chip market with enterprise cloud push
Meta is preparing to sell computing power directly to enterprise customers, entering a cloud market where chip spending could hit $265.8 billion by 2029.
Meta's plan to offer computing power and related services to enterprise clients would pit the social media company against Amazon Web Services, Microsoft Azure and Google Cloud in a market expanding at 25.1% annually.
"We see an opportunity to offer computing power and related services directly to enterprise clients," Meta Chief Executive Officer Mark Zuckerberg said, without providing pricing details or a launch timeline.
The move comes as the global data center semiconductor market grows from $86.8 billion in 2024 to a projected $265.8 billion by 2029, according to MarketsandMarkets. Meta has been investing in custom silicon, including its Meta Training and Inference Accelerator chips, to reduce dependence on Nvidia GPUs for AI workloads.
For Meta, enterprise cloud services would create a new revenue stream beyond its advertising business. The company's open-weight Llama AI models could serve as a differentiator against rivals that offer proprietary models, though Meta would need to build enterprise sales, compliance and support infrastructure from scratch.
Meta's Infrastructure as a Product
Meta operates one of the world's largest computing fleets to power Facebook, Instagram and WhatsApp, along with its AI research division. The company has developed custom accelerators — the MTIA family of chips — designed specifically for AI inference, the process of running trained models to generate responses. That in-house silicon, combined with data centers spanning multiple continents, gives Meta a potential cost advantage that smaller cloud providers lack.
The company's Llama family of large language models, released as open-weight software, already runs on third-party cloud platforms including AWS and Google Cloud. Offering Meta's own compute capacity directly would let the company capture revenue that currently flows to those rivals.
A Crowded Field With an Open-Source Twist
The enterprise cloud market is dominated by three players. Amazon Web Services holds roughly 31% market share, Microsoft Azure about 24% and Google Cloud approximately 11%, according to Synergy Research Group. Meta would enter as a challenger, but with a differentiating strategy: its open-weight approach to AI models.
Nvidia Chief Executive Jensen Huang recently voiced support for open-weight and open-source AI models, adding his voice to a debate that has drawn attention in Washington as US officials consider restrictions on Chinese AI. Meta's open strategy could appeal to enterprises seeking to avoid vendor lock-in, particularly in regulated industries where data control is critical.
The broader infrastructure buildout shows no signs of slowing. Nvidia is in talks to provide roughly $250 billion in financial backing to help OpenAI lease and build a massive AI data center campus in Ohio, the Wall Street Journal reported. Core Scientific and AMD have announced a partnership giving AMD access to more than 500 megawatts of US data center capacity starting in 2027, with potential to scale to 2.5 gigawatts. Anthropic secured a $1.8 billion cloud deal with Akamai for AI expansion.
Meta shares have gained this year as investors weigh the company's AI investments against its advertising revenue. The enterprise cloud push, if executed, would put Meta in direct competition with Amazon and Microsoft — two companies with decades of enterprise relationships and compliance certifications. Meta has not disclosed pricing or a target launch date, making the revenue potential from this initiative difficult to estimate.
This article is for informational purposes only and does not constitute investment advice.
