Meta Google AI Chips Under Consideration
Meta Google AI chips talks could shift data-center procurement and prompted Nvidia share weakness while pushing traders to reweight AI-compute exposure.

KEY TAKEAWAYS
- Meta reportedly in talks to buy Google TPUs for data centers starting 2027.
- May rent TPU capacity from Google Cloud in 2026 to test before wider deployment.
- Deal remains unfinalized with terms, volumes, and financial commitments undisclosed.
HIGH POTENTIAL TRADES SENT DIRECTLY TO YOUR INBOX
Add your email to receive our free daily newsletter. No spam, unsubscribe anytime.
Meta Platforms (META) is reportedly in talks with Google to invest billions in Google’s tensor processing units (TPUs), known as Meta Google AI chips, with plans to rent TPU capacity from Google Cloud in 2026 for testing before broader deployment in data centers starting in 2027, according to reports on Nov. 25, 2025.
Talks Timing and Deal Status
Meta is discussing the use of Google’s TPUs in its data centers beginning in 2027. The company may start renting TPU capacity from Google Cloud as early as 2026 to evaluate the technology in production environments. The reported deal remains in early stages and is not finalized. Terms, volumes, and financial commitments have not been disclosed. Neither Meta, Google, nor Nvidia has made any official announcement, regulatory filing, or public confirmation as of Nov. 25, 2025.
Meta’s Current AI Chip Use and Potential Shift
Meta currently runs most of its large-scale AI workloads on Nvidia chips. Adopting Google TPUs would mark a strategic shift in Meta’s AI hardware procurement, potentially moving some data-center AI tasks from Nvidia GPUs to Google’s TPUs. Renting TPU capacity would allow Meta to test performance before committing to longer-term purchases, altering vendor dynamics in large-scale AI compute.





