Stock Markets March 23, 2026

Musk’s 'Terafab' Ambitions Could Redirect Chip-Equipment Demand Toward Select Suppliers

Mizuho flags potential winners among WFE suppliers if Elon Musk’s proposed in-house semiconductor buildout proceeds

By Derek Hwang ASML KLAC AMAT LRCX
Musk’s 'Terafab' Ambitions Could Redirect Chip-Equipment Demand Toward Select Suppliers
ASML KLAC AMAT LRCX

Mizuho’s technology research team says Elon Musk’s proposed Terafab - a multistage effort to produce bespoke chips for Tesla, xAI and SpaceX - could reshape investor sentiment across wafer-fab-equipment suppliers. The bank highlights the project’s potentially vast energy and GPU requirements, an initial equipment spend penciled in for 2026, and a shortlist of suppliers that may stand to benefit if the plans are funded and move forward.

Key Points

  • Elon Musk’s proposed Terafab is described as a phased, large-scale semiconductor buildout for Tesla, xAI and SpaceX, with early equipment spending of about $20 billion in 2026.
  • Mizuho highlights extreme technical demands cited for the project - power needs up to a terawatt and roughly 1 billion NVIDIA GPUs annually - and suggests this could influence investor sentiment across wafer-fab-equipment suppliers.
  • If implemented and funded, ASML, KLA, Applied Materials, Lam Research, Tokyo Electron and Teradyne are identified as suppliers that could benefit from increased WFE demand.

Mizuho semiconductor coverage warns that Elon Musk’s outline for a large-scale in-house semiconductor complex - dubbed "Terafab" - could be a meaningful narrative for companies that supply wafer-fab-equipment (WFE).

In a note, Mizuho TMT analyst Jordan Klein described the initiative as having "big semi fab spending plans" to support custom chips for Tesla, xAI and SpaceX. The note characterises the effort as sizable in both capital expenditure and the volume of new production tools that would be required.

Mizuho highlights two headline-scale technical metrics cited in the plan: estimated power requirements that could approach a terawatt, and demand for around 1 billion NVIDIA GPUs per year. The bank says the program is expected to be phased, with an early tranche of approximately $20 billion earmarked for new equipment in 2026.

The analysis stresses that a key driver for Musk pursuing in-house manufacturing is capacity constraints with current foundry partners. Klein notes that Taiwan Semiconductor Manufacturing Co. (TSM) is effectively sold out, while Samsung is already producing Tesla’s next-generation chip. According to the note, neither foundry would commit to large-scale capacity expansion "for data centers in space projects this early," which is part of the rationale offered for bringing production in-house.

Mizuho argues that the immediate market reaction may be more about investor perception than about near-term WFE revenue forecasts. Nevertheless, the firm lists specific equipment suppliers that could see demand if Terafab advances through funding and final planning.

Among those named, ASML is identified as a primary potential beneficiary due to the expected reliance on extreme ultraviolet (EUV) lithography for sub-2nm node capability. KLA is cited as being well positioned, and the note says Musk’s stated needs for extensive etch, deposition and advanced packaging could also support Applied Materials (AMAT), Lam Research (LRCX), Tokyo Electron (TEL) and Teradyne (TER).

Klein frames the project as "another potential call option for upside to WFE in [the] coming 3+ years," with the caveat that outcomes depend on financing and the final structure of the plans.


Contextual note - Mizuho’s commentary links the Terafab concept to potential supplier demand but does not provide definitive commitments or firm orders from Musk’s companies. The analysis emphasises a combination of technical scale and uncertainty around funding and implementation.

Risks

  • Funding and finalisation risk - Mizuho notes the outcome hinges on whether the Terafab plans are financed and how final plans are structured, which affects the likelihood of supplier orders.
  • Timing and implementation uncertainty - the project is described as phased and the bank stresses investor perception may move faster than actual near-term WFE revenue impact.
  • Foundry capacity and partner commitment - existing foundry partners are cited as constrained (TSM sold out; Samsung producing Tesla’s next-gen chip) and unwilling to make very large early expansions, which is the stated reason for pursuing in-house production but also a source of execution complexity.

More from Stock Markets

Warsaw Stocks Finish Higher as WIG30 Advances 0.71% Mar 23, 2026 Copenhagen Market Edges Lower as Key Sectors Weigh on Index Mar 23, 2026 Barclays Sees AWS Revenue Surge as AI Spending Boosts Amazon Outlook Mar 23, 2026 Taco Bell and Dunkin’ Franchisee to Pay Over $1.5 Million in NYC Scheduling Dispute Mar 23, 2026 BofA Keeps Buy Rating on ASML After Asian Investor Meetings, Sees Extended Memory Strength Mar 23, 2026