Reuters reported on Thursday that Meta plans to begin manufacturing its AI chip starting in September, lifting ancillary firms’ shares, but the rally fizzled.
- Lam Research and KLA Corp stocks declined 2.4% each, while Applied Digital shares fell nearly 1%.
- Citi estimates that the wafer fab market will grow from $145 billion currently to $200 billion in 2027 and $250 billion in 2028.
- Retail sentiment was ‘bearish’ for LRCX and KLA, and ‘neutral’ for APLD.
Shares of leading chip equipment makers pulled back in early premarket trading on Friday, giving back a portion of the previous session’s gains after reports that Meta Platforms Inc. was accelerating its custom chip program sparked a rally that appears to have run its course.
Lam Research and KLA each declined 2.4%, while Applied Digital fell nearly 1% — retracing part of the 6%, 3.8%, and 2.7% gains they posted respectively on Thursday.
Reuters reported that Meta plans to begin manufacturing a new AI chip, codenamed “Iris,” from September, as part of a broader plan to boost its total computing capacity to 14 gigawatts next year, according to an internal memo reviewed by the agency.
Investors initially read the news as a bullish signal for chip equipment makers — the companies that supply the wafer fabrication tools used to turn raw silicon into microchips — and piled into Lam Research, KLA, and Applied Materials. By Friday morning, however, that momentum had faded.
Citi estimates the wafer fab equipment market will grow from $145 billion today to $200 billion in 2027 and $250 billion in 2028, a second-order effect of the rapid expansion in chip production. Year-to-date, LRCX, KLA, and APLD have each risen more than 90%.
On Stocktwits, retail sentiment was ‘bearish’ for LRCX and KLA, and ‘neutral’ for APLD.
Hyperscales Targeting Custom Chips
Meta is designing custom chips in partnership with Broadcom, with production handled by Taiwan Semiconductor Manufacturing.
The forthcoming Iris chip is part of a four-generation project for Meta Training and Inference Accelerators (MTIA) that Meta will design in-house.
Major cloud players such as Alphabet and Amazon have pursued a similar strategy for years, using custom chips to lower costs and reduce their dependence on suppliers like Nvidia.
For updates and corrections, email newsroom[at]stocktwits[dot]com.<