Elon Musk has revised what he said about chipmaking. In a new post, he admits he has “a tremendous amount to learn” and says he doesn’t know how to make chips. That’s a meaningful change, because only a few weeks ago he told Chinese media that China could get around lithography and chipmaking limits in “two or three years,” even with export controls in place and no access to the best equipment.
He says Terafab, the Texas chip project, would begin with standard tools, then modify those machines to push more throughput, and later move on to building custom equipment. The plan also pulls mask-making, fabrication, packaging, and process work into one operation, rather than relying on the usual supplier setup where those pieces are spread across different companies. Some people have compared the Texas effort to a Tesla-SpaceX-Intel mashup. Reports put the starting cost at about $17 billion. And Musk’s position seems to be that the team can still solve it, even if he’s not a chip industry veteran.
For people who follow semiconductors, the scale of the challenge is pretty clear. By most industry estimates, Taiwan produces more than 60% of the world’s chips and more than 90% of the most advanced ones, while Taiwan Semiconductor Manufacturing Co. (TSMC) still dominates the foundry business. The obstacles aren’t just money. They’re equipment, talent, process complexity, and, above all, getting high yields at volume. That’s why some observers think mass production at strong yields could turn out to be tougher than some of Musk’s aerospace targets. And all of this feeds back into the larger US-China argument over export controls: do they actually slow China down, or do they just give China more reason to build faster toward self-sufficiency?
