Elon Musk's Terafab is a WFE Windfall, Not an ASML Killer

Claim 1 · mechanism
Lithography tools swallow only 20 to 25% of a modern fab's budget; the remaining 75 to 80% goes into non-lithography equipment (chemical vapor deposition, atomic layer deposition, high-aspect-ratio etching, optical metrology), so Terafab's capital flows mainly to the broader WFE vendor ecosystem such as Applied Materials, Lam Research, and KLA.

What would break it: Terafab's disclosed capex allocation shows lithography taking a majority share of the fab budget, or the non-lithography share falling far below the stated 75 to 80%.

Claim 2 · mechanism
A centralized linear accelerator is a single-point-of-failure architecture: one vacuum leak or magnet failure along the main beamline halts 100% of the fab's throughput instantly, unlike modular per-scanner plasma sources.

What would break it: Terafab's disclosed architecture shows redundant beamlines or sources such that a main-line failure degrades throughput partially instead of halting the whole fab.

Claim 3 · thesis
Terafab does not break ASML's monopoly: a Free-Electron Laser solves only the illumination problem, and Terafab still has to buy TWINSCAN EUV and High-NA scanner frames directly from ASML, so any accelerator light source simply plugs into an ASML chassis.

What would break it: Terafab produces commercial wafers on a non-ASML scanner chassis, or its ASML scanner orders are cancelled because an alternative chassis replaces ASML frames.

Claim 4 · empirical
Initial filings outline a $16.8 billion Phase 1 investment for Terafab, with total commitments expanding to $119 billion over time, toward a planned 100-million-square-foot site in Grimes County, Texas with 100,000+ monthly wafer starts.

What would break it: The filings or subsequent authoritative disclosures show materially different Phase 1 or total commitment figures, or the project is formally scaled back below the stated plan.

Claim 5 · mechanism
Co-location is Terafab's structural bet, not light physics: stacking logic fabrication, high-bandwidth memory, photomask creation, and 3D packaging on one Texas site internalizes leading foundries' gross margins above 50% and compresses defect feedback loops that usually take six weeks across three oceans down to days.

What would break it: Terafab's built design keeps major pipeline stages (logic, memory, packaging, assembly) off-site on other continents, or its operating data show feedback-loop times and unit costs no better than the fragmented supply chain.

Claim 6 · empirical-forecast
In the 0-3 year window, Tesla's free cash flow margins take a hit as CapEx spikes ahead of yield during the initial $16.8 billion deployment.

What would break it: Tesla's reported free cash flow margins hold or expand through the Phase 1 deployment window despite the capex spike.

Claim 7 · mechanism
ASML's actual monopoly sits in thermal management and stage metrology, not photon creation: at 13.5 nanometers the beam must bounce off Molybdenum/Silicon multilayer mirrors built exclusively by Carl Zeiss SMT, each reflecting roughly 70% of the light, so barely three percent of the original photons reach the wafer across a ten-mirror relay.

What would break it: A production EUV lithography system ships that does not use Carl Zeiss SMT multilayer mirrors, or an end-to-end mirror relay demonstrates photon transmission far above the stated roughly three percent.

Claim 8 · empirical-forecast
In the 3-5 year window, pure-play foundries like TSMC face valuation pressure as captive megafabs absorb their own internal demand and external foundry pricing power weakens.

What would break it: While Terafab ramps captive production, TSMC's leading-edge pricing power, gross margins above 50%, and valuation hold or expand through the 3-5 year window.

Claim 9 · empirical
Standard EUV scanners' Laser-Produced Plasma sources hit a hard thermal ceiling around one kilowatt before destroying themselves, capping how fast wafers move through the line.

What would break it: A commercial Laser-Produced Plasma EUV source sustains multi-kilowatt output (for example 2 kW or more) in production scanners without the stated thermal self-destruction.

Claim 10 · empirical
ASML CEO Christophe Fouquet confirmed direct talks with Musk about Terafab.

What would break it: ASML or Musk publicly denies that direct talks about Terafab took place.

Claim 11 · thesis
The equipment vendors collect regardless of Terafab's outcome: purchase orders to Applied Materials, Lam Research, KLA, and ASML are cut today, years before a single commercial wafer clears the line, and toolmakers still cash the checks whether Terafab scales cleanly or bogs down in a multi-year yield nightmare.

What would break it: Terafab is cancelled or indefinitely stalled before substantial WFE purchase orders to the named toolmakers are placed and paid.

Claim 12 · empirical-forecast
Through 2026 and 2028, greenfield yield drag will stall Terafab: launching Intel's 14A process inside an unproven Texas facility carries high execution risk, and high wafer scrap rates will burn billions in cash before usable chips emerge.

What would break it: Terafab reaches usable 14A production yields in the 2026-2028 window with scrap rates at mature-fab levels and no multi-billion scrap burn.