Musk pitches “Terafab,” an ultra-scale U.S. chip plant that could dwarf Tesla’s Texas factory

Tesla CEO Elon Musk on Tuesday reignited attention on a grand vertical-integration semiconductor plan he has discussed before, describing a facility called Terafab that he says would be built at a scale unlike anything currently in U.S. chip manufacturing. The renewed pitch is being attributed to an X post that circulated through investor and tech circles, and it has been picked up by business media outlets tracking Musk’s comments.

According to the reporting, Musk characterized Terafab as roughly a 100 million square foot plant, about 10 times larger than Tesla’s Giga Texas facility. He also linked Terafab to a long-term output target of one terawatt (1 TW) of compute hardware per year. A terawatt of compute hardware is not a standard metric used in public semiconductor capacity disclosures, and the reporting frames it as an ambitious way to express overall compute output rather than simply wafer starts or finished-gate production.

The same reports say Musk’s message was fundamentally about scale, echoing a theme of using manufacturing scale to drive cost and speed. The idea, as described, is that Terafab would function as a “closed-loop” operation, spanning multiple stages of the chip supply chain in one location, rather than relying on a patchwork of external suppliers.

In explaining what Terafab might include, the outlet further said the plan is tied to documentation associated with SpaceX. It reported that SpaceX’s filings describe a vertically integrated model that combines lithography mask design, fabrication of logic and memory chips, and advanced packaging within a single end-to-end plant. That characterization remains secondary, because the underlying filing text is not reproduced in the reporting, and Musk has not laid out full technical and commercial details in the public statements highlighted by the media.

The renewed discussion also places Terafab in the context of Tesla’s growing internal compute needs. The outlet referenced Tesla’s Q1 2026 update, including reported revenue of $22.39 billion and a jump in automotive gross margin to 21% from 16% year over year, and it attributed part of Tesla’s captive-silicon rationale to demand for AI and autonomy systems. It also pointed to Tesla’s broader chip ecosystem, including inference chips used for AI5/AI6 and other in-house silicon efforts, as beneficiaries of continued access to leading-edge manufacturing.

From an industry perspective, the prospect of a massive, integrated facility would matter most to suppliers involved at the “middle and end” of chip production, including companies that provide packaging, test, and other steps that can become bottlenecks as compute demand grows. The business media coverage also suggested that semiconductor giants that sell equipment and process technology could face both increased competitive pressure and new partnership opportunities, depending on how Terafab is ultimately built and staffed.