Inside the $10B Infrastructure Talks Between Meta and Anthropic as Compute Scarcity Bites
In the ultra-competitive artificial intelligence race, bitter boardroom competitors are being forced to come together in an age of interdependence never seen before. In a move that showcases the physical hardware limits of the sector, Meta Platforms Inc, according to the reports, is looking forward to leasing the computing infrastructure to none other than its competitor, Anthropic.
The deal, first presented in June by Anthropic, is said to be worth up to $10 billion in total within a flexible two-year period. It comes as a tactical shift in the technology combination of Mark Zukerberg’s. Meta is forecasted to pour a staggering $145 billion in expenditures into its data centers and infrastructure in 2023, raising continuous concerns on Wall Street about potential gains.
Meta’s leasing of the otherwise idle server racks as an income generator within a presumed “Meta Compute” strategy would help counteract infrastructure expenses as its consumer software lags behind. To guarantee the success of the deal, the firm even hired long-time Amazon Web Services expert Dave Brown to build its cloud architecture growth.
It is essential for Claude-maker to guarantee access to hardware at all costs because of the existing chip shortage, which has already affected the ability of the firm to deploy its new flagship products due to usage limitations. Anthropic has already been snapping up the available compute power in Silicon Valley.
Anthropic has already inked a deal worth $45 billion to use SpaceX’s Colossus supercomputers, and now Meta becomes the newest addition to this diversification play. The initial talks allow both sides to exit within 90 days, but the fact that the negotiations take place at all indicates one thing clearly.
The ownership of compute engines is infinitely more profitable today than writing software for them. This massive infrastructure partnership demonstrates that owning hardware capacity is the most powerful strategy in the global artificial intelligence world.
