Anthropic and Google secure $35B AI data center financing push

This financing package is essentially Anthropic’s declaration that it wants to sit in the same infrastructure league as hyperscale cloud providers rather than merely rent from them. A $35 billion, Google-anchored data center buildout at 1+ GW scale gives Anthropic dedicated, long-term access to the scarce training and inference capacity that now defines the AI hierarchy. In the race to AGI, ownership and control of compute have become as strategic as model quality, because they determine how fast you can iterate, how cheaply you can serve, and how insulated you are from rivals’ pricing power. ([ainexusdaily.com](https://www.ainexusdaily.com/article/google-backs-anthropic-35-billion-data-center-expansion-ipo))

The deal also tightens the knot between Anthropic and Google. By guaranteeing leases and supplying TPUs, Google effectively becomes both Anthropic’s landlord and chip vendor, creating a vertically integrated financing model reminiscent of how the cloud majors built their own fleets. That raises real competitive questions for other labs that still depend heavily on third-party clouds, and for regulators worried about concentration of AI compute under a handful of tech conglomerates. If Anthropic can successfully multi-source between TPUs, Nvidia GPUs and other accelerators while owning its data center footprint, it will have one of the most resilient hardware stacks in the field—an asset that can accelerate AGI-scale experimentation while making future policy interventions more complicated.