Meta Platforms saw its stock fall 2.6% on Thursday morning after the company revealed ambitious targets to expand its computing footprint, highlighting the enormous capital outlays needed to compete in artificial intelligence.
In a development that balances technological progress with heavy near-term costs, Meta said it will begin producing its first self-designed AI chip this September. The processor, built under the internal code name "Iris," is the opening salvo in a four-generation effort called Meta Training and Inference Accelerators - MTIA - developed entirely in-house. Meta is collaborating with Broadcom on chip design and has contracted Taiwan Semiconductor Manufacturing Co. for fabrication, with the stated aim of reducing reliance on expensive third-party suppliers such as Nvidia and Advanced Micro Devices.
Yet the financial scale required to reach that independence is vast. An internal memo reviewed by Reuters, which outlined Meta's deployment plans, said the company intends to bring seven gigawatts of computing capacity online this year and to double that level by 2027. To support that buildout, Meta expects to spend up to $145 billion on AI infrastructure this year alone - a sum that represents a substantial portion of an estimated $700 billion in industry-wide spending across Big Tech. It was this aggressive capital expenditure profile that weighed on the stock, as investors typically react negatively when large infrastructure bills threaten short-term margins before producing clear revenue benefits.
Adding to the cost dynamics, Meta is securing long-term supply arrangements with suppliers including Samsung Electronics, Sandisk, and Sumitomo Electric to lock in the components needed for the expansion. The simultaneous data center buildout across major technology companies has driven up prices for memory and AI chips. Analysts at Morgan Stanley have flagged this trend as "chipflation," suggesting rising component costs may amplify the financial pressure on firms expanding AI infrastructure and could further compress margins for companies like Meta in upcoming quarters.
From an operations and supply-chain vantage point, the company is attempting to convert an extensive backloaded infrastructure plan into future cost savings by vertically integrating chip design and securing supply lines. However, the immediate effect is a material increase in capital deployed and longer-term contractual commitments for components. Investors responded to that trade-off by marking the stock down, reflecting uncertainty about when, and if, the infrastructure investments will translate into improved profitability.
Context limitations: The information above reflects disclosures described in the internal memo and subsequent market reaction. It does not include forecasts beyond those stated, nor does it introduce developments not presented in the disclosed memo.