What exactly is Intel doing in Elon Musk's chip venture? Nobody seems to know yet.
The partnership between Musk's new semiconductor company Terafab and Intel has been announced, but the specifics - what Intel will manufacture, on what timeline, and at what volume - remain vague enough that Wired identified five major unanswered questions about the deal.
The context matters. AI chips are currently dominated by Nvidia, and the most advanced chip manufacturing is concentrated at TSMC in Taiwan. Intel has been trying to rebuild its contract manufacturing business (called a foundry operation, where it makes chips designed by other companies) and needs high-profile customers to justify the billions it has spent on new fabrication plants. Musk, through xAI and other ventures, has an obvious interest in reducing dependence on Nvidia's supply chain and pricing.
The problem is that Intel's foundry ambitions have run into repeated technical challenges. Advanced chip manufacturing is measured by process nodes - the distance between transistors on a chip, where smaller means denser and faster. Intel has struggled to reach competitive process nodes that match what TSMC routinely produces. A chip company betting its supply chain on Intel is making a bet that Intel's manufacturing catches up to where it needs to be.
For the AI industry, chip supply concentration is a genuine infrastructure risk. Everything from API pricing to model training costs depends on who manufactures the hardware and at what yield. A credible Intel-based alternative would shift those dynamics. Right now the uncertainty is whether Terafab and Intel can move from announcement to working silicon - and the answers to those five questions will determine whether this partnership represents a real alternative or just a headline.