Meta is reportedly preparing to launch a cloud computing business that would rent out its unused AI computing power to outside customers. The social media behemoth is developing a service that would allow enterprises, AI startups and developers to access Meta’s unused GPU infrastructure whenever it is not fully occupied by internal workloads, reports Bloomberg. The initiative could become one of Meta’s biggest efforts to diversify its revenue beyond its core social media and digital advertising businesses, while creating a new long-term revenue stream from the billions of dollars it has invested in AI infrastructure.
If launched, the business would put Meta in direct competition with established cloud providers like Amazon Web Services (AWS), Microsoft Azure and Google Cloud, all of which already offer AI computing services to enterprise customers. Notably, Meta has never operated a public cloud platform. Instead, it has built one of the world’s largest private AI computing networks to support its own products, including Facebook, Instagram, WhatsApp, Threads, Meta AI, its Llama family of AI models, recommendation systems, advertising technology and content moderation tools.
According to the report, under the proposed plan, the Mark Zuckerberg-led firm is considering multiple ways to commercialize its AI infrastructure. One option is to rent out excess GPU computing capacity to businesses, AI startups and developers that need high-performance computing for training and running AI models. Another option under consideration is offering customers access to AI models hosted directly on Meta’s infrastructure, similar to Amazon Web Services’ (AWS) Bedrock platform. Instead of requiring customers to manage their own infrastructure, this approach would allow them to access and use Meta-hosted AI models through a managed cloud service.
The move comes after Meta sharply increased its spending on AI infrastructure. CEO Mark Zuckerberg has made AI the company’s top strategic priority, leading to massive investments in advanced GPUs, networking equipment and hyperscale data centers. Meta has forecast capital expenditure of $125 billion to $145 billion in 2026, almost double the around $72 billion it spent in 2025, with most of the investment aimed at expanding AI computing capacity. Even earlier this year, Zuckerberg said the company could eventually enter the cloud computing business if it ‘overspends’ on data centers and ends up with excess computing capacity. At Meta’s annual shareholder meeting in May, he also said many companies had already expressed interest in buying access to Meta’s AI infrastructure and APIs, although the company is currently using most of its capacity for its own AI development.
Meta’s entry would further intensify competition in an AI cloud market currently dominated by AWS, Microsoft Azure and Google Cloud, while also challenging newer AI-focused infrastructure providers like CoreWeave and Nebius. At the same time, investors reacted positively to the report, with Meta shares rising about 6% in premarket trading, while shares of several AI cloud companies declined on expectations of increased competition. The timing is even more notable as several tech giants are estimated to spend more than $700 billion on AI infrastructure this year, compared with about $400 billion in 2025.
The Tech Portal is published by Blue Box Media Private Limited. Our investors have no influence over our reporting. Read our full Ownership and Funding Disclosure →
Ashutosh is a Senior Writer at The Tech Portal, largely reporting on new tech, and intersection of technology and business. Ashutosh’s career spans across nearly a decade of technology writing across multiple platforms and languages.
