Meta Platforms is in talks to lease its advanced artificial intelligence computing power to rival developer Anthropic in a landmark deal valued at up to $10 billion, according to sources familiar with the discussions. The potential partnership, emerging this week from Silicon Valley, underscores the critical global shortage of specialized microchips necessary to train and run next-generation AI models. By leveraging its massive hardware stockpile, Meta could pioneer a lucrative new business model while helping Anthropic scale its competitive Claude AI models.
The Global Scramble for Compute Power
The generative artificial intelligence boom has triggered an unprecedented race for infrastructure. Anthropic, a prominent AI safety and research company founded by former OpenAI executives, requires vast amounts of computational power to develop its flagship Claude models. While Anthropic has secured billions in funding from tech giants like Amazon and Google, the sheer demand for high-performance graphics processing units (GPUs) continues to outstrip supply.
Meanwhile, Meta has spent the past two years aggressively acquiring Nvidia’s H100 Tensor Core GPUs. Meta Chief Executive Officer Mark Zuckerberg previously announced plans to secure roughly 350,000 of these highly coveted chips by the end of 2024. This massive infrastructure investment has positioned Meta as one of the world’s largest holders of AI computing capacity, surpassing many traditional cloud service providers in sheer hardware volume.
Meta’s Pivot to Infrastructure Provider
If finalized, the $10 billion agreement would mark a significant strategic pivot for Meta. Historically, the social media giant has built global data centers exclusively to power its own platforms, including Facebook, Instagram, and its open-source Llama AI models. Leasing excess computing capacity to external developers would transform Meta into an unexpected competitor in the cloud infrastructure market, directly challenging Amazon Web Services (AWS), Microsoft Azure, and Google Cloud.
For Anthropic, diversifying its infrastructure providers mitigates the risk of relying solely on its primary backers. Although Amazon remains Anthropic’s primary cloud provider under a multi-billion-dollar partnership, securing additional compute from Meta provides the startup with the raw hardware needed to train its next-generation models without bottleneck delays. This multi-cloud approach ensures that Anthropic’s development timeline remains unhindered by hardware allocation limits.
The Economics of the $10 Billion Deal
Industry analysts view the potential deal as a validation of Meta’s heavy capital expenditure strategy, which has occasionally drawn skepticism from Wall Street investors. Meta recently raised its capital expenditure guidance for 2024 to a range of $37 billion to $40 billion, citing accelerated investments in AI servers and data centers.
“Leasing compute power allows Meta to offset its astronomical capital expenditures immediately,” says Daniel Newman, principal analyst at Futurum Group. “Rather than waiting for long-term monetization of consumer AI features, Meta can generate immediate, high-margin revenue from its hardware stockpile, turning a cost center into a profit engine.”
Data from research firm Synergy Research Group indicates that enterprise spending on cloud infrastructure services reached $79 billion in the second quarter of 2024 alone. A $10 billion compute lease would instantly position Meta as a major player in this rapidly expanding market, proving that physical infrastructure is just as valuable as software in the modern tech ecosystem.
Shifting Alliances and Regulatory Hurdles
The potential partnership could complicate the intricate web of alliances defining the tech industry. Anthropic’s close ties to Amazon and Google—both of which have invested billions in the startup—could face friction if Anthropic begins running its workloads on Meta’s infrastructure. Additionally, the deal highlights how computing power, rather than proprietary algorithms, has become the ultimate currency in the AI race.
Regulators are also likely to scrutinize the transaction. Antitrust watchdogs in the United States and Europe have already expressed concern over the concentration of AI resources among a handful of dominant tech firms. A massive infrastructure deal between two industry leaders will undoubtedly draw attention from the Federal Trade Commission (FTC), which is already investigating existing partnerships between big tech companies and AI startups.
Looking ahead, observers will watch whether this negotiation triggers similar compute-sharing agreements across the tech sector as smaller AI startups struggle to secure hardware. The finalization of the deal could also accelerate the development of Anthropic’s next-generation models, potentially altering the competitive balance between Anthropic, OpenAI, and Google in the race for artificial general intelligence. Furthermore, if Meta successfully monetizes its idle compute, other tech giants with massive capital reserves may follow suit, fundamentally restructuring the global cloud computing market.
