Intel tipped as hidden winner of Musk's Terafab chip push

Terafab is Musk's attempt to do what even the largest technology companies have struggled to accomplish: manufacture advanced AI chips at an unprecedented scale. The project stitches chip design, advanced logic and memory manufacturing, and packaging into a vertically integrated operation, with Tesla and SpaceX putting in $16.8 billion for phase one. Its long-term goal is enough annual production to support 1 terawatt of computing power, split between processors for Tesla's autonomous vehicles and Optimus robots on Earth and for SpaceX's space-based data centers.

The reason Intel keeps surfacing in this story is the state of its foundry business. In the second quarter, Intel Foundry generated $5.8 billion in revenue, but most came from Intel's own products; external foundry revenue was just $293 million, and the division posted an operating loss of roughly $2.1 billion. That gap is what makes a customer like Musk matter. Intel has been developing its 18A and 14A processes to compete with Taiwan Semiconductor, and Terafab could serve as a public proving ground.

Terafab won't fix Intel by itself. The company still has to improve manufacturing yields, control capital expenditures, attract additional external customers, and eventually make its foundry operations profitable. Those are the conditions that make the stock, in the author's framing, a speculative buy rather than a sure thing.