Musk Demands "Light Speed" Response from Chipmakers for Terafab Project

Gate News message, April 16 — Elon Musk’s team has issued urgent directives to major semiconductor equipment manufacturers including Applied Materials (AMAT), Tokyo Electron, and Lam Research (LRCX), demanding “light speed” responsiveness to prepare for the Terafab project, according to sources. The initiative represents Musk’s ambitious attempt to enter advanced chip manufacturing, with the project targeting annual computing capacity of 1 terawatt, backed by investment of $20-25 billion.

Over recent weeks, employees working for the Tesla and SpaceX joint venture have been inquiring about pricing and delivery timelines for photomasks, substrates, etch equipment, deposition systems, and testing tools. Musk’s representatives requested rapid price estimates while providing minimal information about production specifications. In one instance, suppliers received requests for estimates on a Friday before a holiday, with delivery expected by the following Monday. The initiative has already boosted stock prices of Tokyo Electron (up 6% on the news), Advantest, Screen Holdings, and Disco.

The Terafab project plans to establish a pilot production line in Austin capable of processing 3,000 wafers monthly, with commercial silicon chip manufacturing targeted for 2029. Musk has emphasized that the semiconductors will support xAI operations, humanoid robots, and space data centers. The project addresses concerns about semiconductor supply constraints amid massive AI infrastructure investment—major data center operators are expected to spend approximately $650 billion on infrastructure this year alone.

Intel CEO Chen-Fu Giannou has publicly confirmed deep involvement in Terafab, with both companies collaborating on next-generation processors for robots and hyperscale data centers. Meanwhile, Musk’s team is also negotiating with Samsung Electronics, though Samsung has indicated preference for allocating dedicated capacity at its new Texas facility rather than supporting fully independent manufacturing. Tesla’s AI5 chip successfully completed tape-out recently, validating the necessity for large-scale in-house production capacity to support future full self-driving and Dojo supercomputer demands.

However, investment analysts at Bernstein and other institutions have raised significant concerns. Bernstein estimates the project’s actual capital expenditure could reach $5-13 trillion to achieve the 1-terawatt computing target, far exceeding current budgets. Tammy Qiu, head of technology equity research at Berenberg, noted the firm has not incorporated Terafab into its financial models for ASML, stating “the intent is real” but substantial progress is unlikely within the next two years.