Meta Platforms is negotiating a substantial computing infrastructure arrangement with artificial intelligence firm Anthropic valued at approximately $10 billion over a two-year period, according to sources briefed on the discussions. The proposed transaction would involve Anthropic, which creates the Claude Code AI assistant, acquiring access to Meta's computing capacity through monthly payments throughout the contract duration. While market observers reacted positively to the announcement, Meta's share price declined more than 2 percent on Friday alongside broader technology sector weakness, though the stock recovered slightly in after-hours trading.
The tentative deal represents a strategic diversification initiative for Meta, which has historically derived its primary revenues from digital advertising. By monetising its substantial computational infrastructure, the company seeks to establish a fresh income stream beyond its core business model. This move also positions Meta to compete directly with emerging cloud infrastructure providers such as CoreWeave and Nebius, which have capitalised on surging demand for computing resources driven by the rapid expansion of advanced artificial intelligence applications globally.
The arrangement carries considerable strategic significance for Anthropic, which is preparing for a public share offering. The company recognised the critical importance of securing reliable computing capacity to support its AI systems and approached Meta with the proposal in June. However, the negotiations have encountered complications stemming from Meta's lack of established infrastructure in the commercial cloud computing sector. Unlike traditional cloud providers with mature business operations, Meta must construct an entirely new division capable of managing client relationships, service delivery, and technical support for external customers seeking computing access.
Industry observers note that the potential agreement bears similarities to computational partnerships recently established by other technology enterprises. Notably, Anthropic concluded an agreement in May with SpaceX to harness the complete computational capabilities of Colossus 1, an advanced data centre facility located in Memphis, Tennessee. This pattern suggests that technology companies possessing significant computational resources increasingly view such arrangements as viable revenue opportunities in an environment characterised by intense competition for artificial intelligence infrastructure.
Meta's leadership has publicly signalled openness to expanding into cloud computing markets. During the company's May shareholder meeting, Chief Executive Mark Zuckerberg indicated that participating in cloud services represented a commercially viable prospect under active consideration. Zuckerberg noted that external parties approach Meta on a frequent basis, often weekly, requesting access to either the company's proprietary artificial intelligence models or surplus computing capacity. This steady stream of inquiries underscores the acute scarcity of available computational resources in the contemporary artificial intelligence landscape.
Recent reporting by Bloomberg News has corroborated Meta's intentions regarding cloud service expansion, documenting that the organisation is actively constructing a dedicated cloud division. This unit would function to commercialise excess computing infrastructure and provide hosting services for artificial intelligence models utilised by external software developers. Such infrastructure presents opportunities for Meta to capture revenue from its substantial capital investments in computational facilities while simultaneously building relationships with artificial intelligence developers and enterprises.
The current negotiations remain in preliminary phases, and sources close to the matter emphasise that discussions may ultimately prove unsuccessful in reaching a final agreement. The proposed arrangement explicitly permits both parties to terminate the contract prematurely, a provision that reflects the evolving nature of these negotiations and the complexities involved in structuring a relationship between two organisations operating in the rapidly changing artificial intelligence sector. The specific commercial terms remain subject to modification as talks progress.
Neither Meta nor Anthropic has offered official comment regarding the discussions. Meta's reticence to confirm or deny the arrangement reflects standard corporate practice when handling sensitive commercial negotiations, whilst Anthropic's silence may relate to its forthcoming initial public offering and associated regulatory requirements regarding disclosure of material information. Industry analysts anticipate that clarity regarding the negotiations' outcome will emerge as one or both companies approach critical business milestones.
For Southeast Asian technology enterprises and digital businesses, this development carries meaningful implications. The emergence of alternative computing infrastructure providers, particularly major technology firms leveraging existing computational assets, promises to expand access to artificial intelligence capabilities and reduce dependency on traditional cloud providers. Malaysian and regional companies pursuing artificial intelligence integration and digital transformation initiatives may benefit from increased competition among infrastructure suppliers, potentially translating into more competitive pricing and customised service offerings. Additionally, the growing willingness of established technology platforms to monetise computational resources signals the maturation of artificial intelligence as a fundamental utility within the technology ecosystem, comparable to electricity or telecommunications bandwidth.
