Speaking to shareholders, Elon Musk says Tesla is “probably going to have to build a gigantic chip fab” for AI chips and “maybe, we'll do something with Intel”
- Musk says ‘worth having discussions’ with Intel to make chips — Massive fab needed to make the volume of chips needed
Context & Ripple Effects
Tesla had already moved from buying AI hardware toward designing its own: Musk previously projected major Nvidia spending alongside internal AI investment, while the next-generation AI5 chip was assigned to Samsung in Texas and TSMC in Arizona in Tesla's stated AI5 manufacturing plan.
This makes the fab discussion an escalation from sourcing capacity across established foundries to contemplating dedicated production. Subsequent coverage tied Intel to a Terafab effort, giving the earlier outreach a clearer strategic through-line.
First-order effects
- Tesla signals that its anticipated AI-chip volumes may exceed what it is comfortable securing solely through outside foundry allocations, putting dedicated manufacturing on its planning agenda.
- Intel becomes a prospective manufacturing partner rather than only an external industry peer, though Musk described discussions—not a finalized agreement.
Second-order effects
- Samsung and TSMC face a customer evaluating greater vertical control even as they remain central to Tesla's near-term AI5 supply plan; Tesla gains more leverage and optionality in future capacity negotiations.
- An Intel partnership would test whether a leading AI-chip customer can help validate Intel's advanced manufacturing offer—an arc later reflected in Tesla's plan to use Intel's 14A process at Terafab.
Third-order effects
- If large AI users increasingly pursue captive or tightly aligned fabrication capacity, the industry could shift from a clean split between chip designers and contract manufacturers toward hybrid supply models.
- The proposal underscores that AI competition is becoming constrained by manufacturing capacity as well as chip design; whether this becomes structural depends on execution, capital requirements, and the ability of external foundries to meet demand.
The trend: AI-chip buyers are extending vertical integration from custom silicon design into long-term control of manufacturing capacity.