Assessments in the semiconductor industry are split over the "TeraFab" project that Tesla CEO Elon Musk has pledged. Analysts say attention should focus less on technical feasibility and more on the shock it could send through the global supply chain and its signal of a "power shift."
In the semiconductor industry, there is an assessment that, regarding the internalization of the 2-nanometer process, "the essence is securing mass-production Production yield rather than implementing the process itself." This is an ultrafine process that reduces circuit line width to the 2-nanometer (one billionth of a meter) level, a next-generation technology that can significantly boost performance and power efficiency. However, an industry official said, "Advanced processes at 2 nanometers and below are not solved simply by bringing in equipment; they require years of accumulated process data combined with precise equipment control capabilities," adding, "Even TSMC and Samsung Electronics took considerable time to secure stable Production yield." Accordingly, there is an assessment that TeraFab will act as a long-term threat factor rather than replacing the foundry (contract chip manufacturing) market in a short period.
Many also view this project more as a "strategic bargaining chip" than an actual production plan. Market research firm Jon Peddie Research said in a recent report that "simply mentioning the possibility of in-house production publicly gives Tesla greater leverage in price and volume negotiations with existing foundries." Beyond the trend of big tech corporations gaining control by internalizing design, the interpretation is that they are now signaling an intent to vertically integrate into manufacturing as well, shaking the industry's balance of power.
More complex changes are expected in memory. While the industry sees a low likelihood that TeraFab will internalize high bandwidth memory (HBM), it points to a shift in "spec leadership" as a key variable. According to semiconductor analyses by global investment banks (IB) including Morgan Stanley and Goldman Sachs, until now memory manufacturers have led technical standards, but going forward there is a possibility of a reshuffle in which the influence of large customers demanding "custom memory" optimized for their own AI systems—like Tesla—expands. In particular, for Korean memory makers that focus on general-purpose products, there is a cited risk that their bargaining position could weaken as they respond to customized customer demands.
The biggest constraint determining TeraFab's realism is securing EUV (extreme ultraviolet) lithography equipment. This equipment, essential for advanced processes, is effectively monopolized by ASML, and Intel, TSMC, and Samsung Electronics have already locked in future volumes. A semiconductor equipment industry official said, "Without securing equipment, mass production is impossible even if you have process technology," adding, "If Tesla jumps into large-scale competition to secure equipment, it could affect the facility investment schedules of existing corporations."
Some argue TeraFab should not be interpreted solely through the profitability logic of the existing semiconductor industry. Considering Musk's business structure spanning Tesla, SpaceX, and xAI, the analysis is that this project is not a simple expense reduction but part of a long-term "compute infrastructure acquisition strategy." From the Starlink satellite network to Autonomous Driving and Humanoid Robot development, the assessment is that it is closer to a declaration not to rely on the existing semiconductor supply chain for explosive compute demand.
In the end, the core of TeraFab is not whether Musk actually completes a factory. Analysts say the essence is how much the existing semiconductor ecosystem can respond to his bold demands and speed, and where supply chain leadership shifts in the process.