Elon Musk has confirmed that he is in discussions with TSMC about a possible collaboration centered on Terafab chipmaking. The statement, first reported by Tom's Hardware, marks the first time the Tesla and SpaceX chief executive has publicly acknowledged the Taiwanese foundry by name in connection with the project.

TSMC is now the second semiconductor manufacturer linked to the initiative. Intel remains the only other company Musk has named as a partner in the Terafab effort. No other suppliers or foundries have been disclosed.

According to the report, the Terafab arrangement would make the resulting fabrication capacity an exclusive supply channel for three companies in Musk's portfolio: Tesla, SpaceX, and xAI. Under the stated terms, output from the Terafab collaboration would not serve external customers.

What the confirmation establishes

The disclosure narrows the picture of how Musk intends to secure advanced silicon for his companies' workloads. Three demand drivers sit behind the arrangement:

• Tesla, whose vehicles and autonomy programs require custom compute hardware at scale.
• SpaceX, which depends on radiation-tolerant and high-reliability components for satellites and the Starlink constellation.
• xAI, Musk's artificial intelligence venture, whose training and inference clusters place sustained pressure on accelerator supply.

By tying fabrication output exclusively to these three entities, the structure resembles a captive supply model rather than a merchant foundry relationship. That distinction matters for an industry in which leading-edge capacity at TSMC is typically allocated through long-term agreements with a broad customer base.

TSMC's position

TSMC has not publicly detailed the scope, timeline, or financial terms of any Terafab-related discussions. Musk's confirmation establishes that talks are underway, but it does not indicate whether an agreement has been signed, what process nodes would be involved, or where fabrication would take place.

The foundry operates the world's most advanced logic manufacturing network, spanning facilities in Taiwan, Arizona, and Japan. Any dedicated capacity arrangement for a single customer group would represent a notable departure from its standard allocation practices.

Intel as the other named partner

Intel's inclusion as the sole other named partner gives the effort a two-supplier geometry. The company has spent recent years rebuilding its foundry business under its IDM 2.0 strategy and has courted large-volume customers for its upcoming process nodes.

For Intel, a Musk-linked commitment would provide a high-profile anchor tenant for its contract manufacturing ambitions. For Musk's companies, a second source would reduce dependence on a single supplier — a consideration that has grown more pressing as AI accelerator demand has tightened availability across the industry.

Market implications

The exclusivity clause is the most consequential element of the reported structure. If TSMC and Intel commit dedicated Terafab capacity to Tesla, SpaceX, and xAI, that capacity sits outside the pool available to other chip buyers.

The timing intersects with sustained competition for advanced packaging and leading-edge wafer supply. AI developers, automakers, and satellite operators have all reported extended lead times and allocation constraints over the past two years. A reserved supply channel for three affiliated companies would add a new variable to that calculation.

No production volumes, capital commitments, or target dates have been disclosed. Musk's confirmation establishes intent and named counterparties, but the commercial mechanics of the Terafab collaboration remain undefined.

Tom's Hardware, which first reported the confirmation, did not publish statements from TSMC or Intel responding to Musk's remarks. Neither company has issued a public comment at the time of writing.