Meta shares fell after the company laid out aggressive plans to double computing capacity and begin producing an in-house AI chip in September, highlighting the heavy price tag of its artificial-intelligence ambitions. An internal memo cited by Reuters said Meta aims to deploy seven gigawatts of compute this year and could spend up to $145 billion on AI infrastructure, a sizable slice of Big Tech’s projected $700 billion outlay. The chip effort—built with Broadcom and fabricated by TSMC—seeks to cut reliance on Nvidia and AMD, but near-term margins face pressure as Meta locks in long-term supply deals with Samsung, SanDisk and Sumitomo Electric. Analysts warn of “chipflation” as memory and AI component costs rise, adding another headwind to profitability.






























