Intel joins Musk’s Terafab AI chip project to power humanoid, data centre goals

Intel will join Elon Musk’s Terafab project with SpaceX and Tesla to manufacture AI processors, supporting Terafab’s target of producing 1 terawatt/year of compute; Intel shares rose nearly 3% on the announcement. The partnership reinforces Intel’s turnaround narrative as processor demand improves and follows CEO Lip-Bu Tan’s aggressive restructuring and outside investments (including Nvidia and U.S. government stakes). SpaceX has confidentially filed for a U.S. IPO targeting a market launch later this year, which could be a very large listing if completed.

Analysis

The market is treating a high-profile endorsement as signal that a long-running execution risk for legacy foundries has dropped; that sentiment re-prices optionality around large, bespoke compute programs faster than fundamentals will evolve. If sustained, this can compress discount rates on capital-intensive fabs and packaging lines, moving multiples by mid-single to low-double digits within 6–12 months even before incremental revenue shows up.

A material second-order effect will be upstream — EUV/etch/implant tool queues, advanced packaging capacity and HBM supply become the binding constraints for any multi-exawatt program. Tool OEMs and OSATs operate on 12–36 month procurement cycles, so incremental demand will show up first as order-book strength, then as input-cost and lead-time pressure across the supply chain. This creates a multi-stage alpha opportunity: equipment/packagers first, memory suppliers second, foundries last.

Tail risks are conventional but amplified: timeline slips, yield problems and capex overruns can reverse sentiment rapidly; an IPO or financing hiccup elsewhere in the constellation could cause headline-driven volatility in days. Watch three horizon triggers — order-book re-rates (weeks–months), initial wafer starts (6–12 months), and meaningful product ship/robotics demos (12–36 months) — as stop/go milestones that will validate or undo the rerating. The consensus is buying credibility; the contrarian view is that credibility does not substitute for yield curves, software stacks, or years-long fabs economics.

Key Decisions for Investors

• Long INTC (6–18 month): accumulate on <10% pullbacks from current levels. Rationale: multiple expansion trade tied to order-book visibility and CHIPS-backed capex recognition. Target +30% if quarterly bookings/CMGR data confirm; downside -20% on execution/yield misses. Position size: tactical 3–5% net long of equity exposure, hedge with 6–9 month 5–10% OTM puts for earnings-risk protection.
• Long semiconductor equipment/packaging suppliers (LRCX or AMAT, KLAC, ASE) (6–24 month): buy name(s) that capture pre-revenue order flows. Entry: on next >5% sector-wide dip or XAU/semiconductor ETF divergence; target +35–50% if tool orders accelerate; expected drawdown -25% if program delays. Consider 9–15 month call spreads to cap cost and amplify directional exposure.
• Long memory suppliers exposed to HBM (MU or 9–18 month call spreads): directional exposure to elevated HBM demand as advanced compute programs scale. Trade: buy 9–15 month call spreads pegged to production ramp windows; R/R ~2:1 if HBM ASPs firm; downside limited to premium paid if lead-times slip.
• Tactical TSLA asymmetric options (12 month): buy a modest call spread (e.g., 1x long-dated call / sell nearer-term call) to capture robotics/autonomy upside while financing cost with near-term premium decay. Keep allocation small (1–2% net) due to binary execution risk from product timelines and regulatory outcomes.