Meta Explores AI Infrastructure Deal With Anthropic Amid Data Centre Expansion
Meta is reportedly in early discussions with artificial intelligence startup Anthropic about leasing computing capacity, a move that could see the social media company expand into the cloud computing market alongside established players such as Amazon, Microsoft and Google.
A source familiar with the discussions confirmed that talks are ongoing but remain at an early stage. While reports have suggested the potential agreement could be worth up to $10 billion over two years, the source said figures circulating publicly are speculative.
Neither Meta nor Anthropic commented on the reported negotiations.
If completed, the arrangement would create a new revenue stream for Meta, which has invested heavily in expanding its AI infrastructure in recent years.
The company has projected capital expenditure of between $125 billion and $145 billion this year, with much of the spending directed towards building data centres and other infrastructure needed to support its artificial intelligence ambitions. The planned investment is roughly double what Meta spent the previous year.
To help offset those costs, Meta announced in April that it would reduce its workforce by about 10%, affecting approximately 8,000 employees.
Meta Chief Executive Officer Mark Zuckerberg has previously indicated that the company could lease unused computing capacity if its internal demand falls short of the infrastructure it is building.
“Almost every week there are different companies that come to us from outside asking us if we have compute that they could buy from us at some premium to what we’ve bought it at,” Zuckerberg said during Meta’s annual shareholder meeting in May.
He added that while the company currently expects to utilise most of its computing resources, leasing excess capacity remains an option if it eventually builds more infrastructure than it requires.
Demand for AI computing power has surged as businesses accelerate the adoption of artificial intelligence technologies and AI developers race to train increasingly sophisticated models.
Anthropic already maintains multibillion-dollar computing partnerships with companies including Google, Amazon, Microsoft and SpaceX.
The reported discussions also come as investors continue to scrutinise Meta’s substantial AI spending and its ability to generate meaningful financial returns. The company’s shares have fallen by more than 8% over the past year.
Last month, Meta unveiled an upgraded version of its Muse Spark AI model, saying it offers coding capabilities that can compete with products from OpenAI and Anthropic. The company also announced plans to introduce a paid version of the AI service, signalling a broader effort to monetise its investments in artificial intelligence.
