SpaceX Terafab Could Become a New Variable in Semiconductor Equipment Demand

Overview

This report evaluates SpaceX's proposed Terafab semiconductor factory plan from the perspective of the semiconductor supply chain and wafer-fab equipment demand, assessing its potential scale, build-out path, and impact on the semiconductor supply chain. UBS believes that while the market will still question project scale and execution feasibility, public hearing information from Grimes County, related executive hiring, process engineering positions, and equipment supplier feedback are all increasing confidence in the project.

Core views

The core view is that if Terafab progresses as planned, it could begin with a roughly $5 billion pilot line in C2027, rise to about $10 billion in C2028, and reach WFE spending levels above $50 billion per year in C2030/C2031, equivalent to adding a new device-equipment demand source near TSMC scale. UBS believes this would strengthen the view of global WFE nearing $300 billion by C2029 and become an important theme for the semiconductor equipment industry across several upcoming reporting seasons.

Analysis framework

The report combines a top-down capex split approach with a bottom-up capacity benchmarking approach: it first derives approximately $135 billion in WFE demand based on an assumption that SpaceX's AI business will spend about $1.1 trillion in capex over the next five years and that around 20% is for Terafab. It then references greenfield fab investment scales such as Intel Ohio and Micron Boise ID1 to infer that Terafab may include about 80k wsm of memory capacity, two advanced logic/foundry fabs of about 20k wsm each, a mask shop, and back-end packaging and test capability.

Methodology notes

• Industry Capex EstimationWFE capex split

  Share of wafer fab equipment spend in capex

  The report applies a typical assumption that 60% of capex is allocated to WFE and converts Terafab's potential capex into wafer-fab equipment demand.

• Comparative Project Benchmarkinggreenfield fab benchmarking

  Benchmarking against greenfield fab investment scale

  The report compares Terafab with advanced greenfield fab projects such as Intel Ohio and Micron Boise ID1 to infer potential capacity and capex intensity.

• Valuation approachP/E, EV/FCF, SOTP and multiples

  Multiple valuation frameworks

  The report discloses valuation techniques for related companies using P/E, EV/FCF, sum-of-the-parts, and multiples, but the main report text does not provide an explicit target price for this theme.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Semiconductor equipment suppliers

  Potential direct beneficiaries

  Strengths

  Terafab could generate large-scale, sustained WFE orders and raise the industry demand ceiling.

  Weaknesses

  Project pace, order conversion, and equipment delivery timing remain uncertain.

  Comparison

  If WFE spending reaches over $50 billion per year, the scale approaches TSMC's equipment spend level over the next few years.

  Risks

  Execution delays, technology roadmaps changing, capex cuts, or supply chain bottlenecks.

• Intel Corp.

  Potential technology knowledge provider or partner

  Strengths

  May provide process recipes, manufacturing IP, PDKs, design rules, equipment recipes, and operating experience, and may participate through licensing or a JV.

  Weaknesses

  Participation structure is not determined, and Intel itself faces competitive pressure from AMD and NVDA.

  Comparison

  The report compares potential cooperation to the historical IBM-AMD technology-transfer framework and says Ohio One could be one of the Terafab campus sites.

  Risks

  Pilot failure at Terafab, partnership structure not being implemented, or changes in Intel's execution capability or competitive landscape.

• Memory suppliers

  Potential coexistence of competition and investment catalyst

  Strengths

  If Terafab emphasizes memory, attention to U.S. domestic front-end memory manufacturing investment could increase.

  Weaknesses

  The source of SpaceX's memory IP is unclear, and existing memory suppliers have limited incentive to license their IP.

  Comparison

  The report says Korean peers may be prompted to increase U.S. front-end memory manufacturing investment, noting that Samsung has expansion land in Taylor, TX.

  Risks

  If new entrants successfully scale, the memory competitive landscape could shift; if storage IP cannot be obtained, execution of the project could be constrained.

• SpaceX

  Project sponsor and capex principal

  Strengths

  By vertically integrating semiconductor manufacturing, it may support exawatt-scale AI compute and shorten design/manufacturing/testing/redesign cycles.

  Weaknesses

  Capital intensity is extremely high, and commercialization and technology path are unproven.

  Comparison

  The report likens it to Tesla's Gigafactory vertical integration strategy in battery manufacturing.

  Risks

  Dependence on Elon Musk's key decisions, delays in technology milestones, AI monetization below expectations, regulatory and geopolitics pressure, and execution risks across the supply chain.

Key data

• SpaceX AI business capex forecast over next 5 yearsabout $1.1 trillionExcludes Connectivity and Launch businesses.
• Terafab-related capexabout $225 billionApproximately 20% of SpaceX AI business capex over the next 5 years.
• Terafab 5-year WFE estimateabout $135 billionEstimated using an assumption that 60% of capex flows to WFE.
• C2027 pilot line scaleabout $5 billionBased on UBS supply chain research and equipment order clues.
• C2028 WFE assumptionabout $10 billionIncluded in UBS's assumption for 2028 WFE of about $25 billion.
• C2030/C2031 annualized WFE potentialover $50 billion per yearComparable to the WFE spending scale of TSMC over the next several years.
• Long-term global WFE viewC2029 close to $300 billionThe partial execution of Terafab would also reinforce this view.
• Initial semiconductor factory capex plan$55 billion, potentially expanding to $119 billion after all stages are completedLocated in Grimes County, TX.

Impact & implications

If Terafab is implemented, the most direct beneficiaries are semiconductor equipment suppliers and related SPE supply chains. It could also affect potential Intel technology licensing, foundry, or joint venture opportunities, and shift expectations for onshoring investment in U.S. memory supply chains. The project may also increase capital expenditure optionality for future advanced-node, memory, packaging and testing, and mask capabilities.

Risks

• Terafab scale and timeline remain materially uncertain.
• The source of storage IP is unclear, and existing storage suppliers have limited motivation to grant licenses.
• Project capital intensity is high, and it may face financing, supply-chain, and execution risks.
• AI commercialization and the space-compute model are still not sufficiently proven.
• Macro downturn, international trade disruptions, technology substitution, and industry structure shifts could affect related company revenues and valuations.
• Intel-related upside depends on whether the cooperation structure, technology licensing, JV, or site arrangement is actually implemented.

What to watch

• Subsequent approval, hearing, and construction progress in Grimes County, TX.
• Whether the C2027 pilot-line order is confirmed at around $5 billion.
• Whether semiconductor equipment suppliers provide more visibility on Terafab orders and demand in company disclosures over the next few reporting seasons.
• Whether additional evidence emerges of SpaceX and Intel agreements on technology licensing, process support, equipment scheduling, or JV arrangements.
• Whether Terafab prioritizes building memory capacity and where its memory IP would come from.
• Whether Korean memory vendors increase U.S. front-end manufacturing investment, especially DRAM-related investment.