SpaceX's Next Big Business Isn't Rockets. It's Computer Chips.

SpaceX’s IPO filings show connectivity (mostly Starlink) contributed about $11.4B of 2025 revenue (~61%), but its new Terafab semiconductor plan targets AI data-center chips in orbit on a prototype in Austin with a large Texas build-out. Investment is pegged at ~$55B initially and up to ~$119B total, yet the article stresses the effort is early and not revenue-generating, creating heavy capital risk versus already-valuation-heavy expectations (stock valued at >$2T). Net: Terafab is framed as upside optionality rather than a near-term buy catalyst, implying a cautious stance for shareholders.

Analysis

The market is likely to over-assign near-term value to a project whose economic payoff is gated by manufacturing yield, launch cadence, and orbital reliability, not vision. The real incremental winner is INTC: it gets to monetize process know-how and regain strategic relevance without bearing the full capex burden, but the financial contribution is still more reputational than P&L accretive for at least 12-24 months.

SPCX is the most exposed to disappointment because the story invites investors to capitalize a 6-10 year option as if it were a 1-2 year product cycle. That can support a higher multiple in a momentum tape, but it also raises the chance of a sharp de-rate if management is forced to choose between Starlink/Starship reinvestment and a $100B+ science project. TSLA is a subtler loser: the more Musk’s ecosystem looks like a sprawling capital-allocation machine, the more the market can haircut governance quality and execution focus, even if no cash leaves Tesla directly.

The contrarian miss is that the first-order trade is not about orbital AI at all; it is about who gets paid during the experimentation phase. Equipment, packaging, launch, and connectivity are the monetization layers that matter first, while the end-market revenue is years away. The thesis breaks if SpaceX publishes binding customer commitments, credible fab milestones, or capex that is materially funded off-balance-sheet rather than from the core business.

Key Decisions for Investors

• Fade strength in SPCX on any headline-driven pop: initiate a small short or put-spread structure only if borrow/liquidity are workable, targeting a 1-3 month window where valuation can decouple from fundamentals; risk is a momentum squeeze if management announces a financing partner or anchor customer.
• Pair long INTC / short TSLA over 1-3 months as a cleaner expression of 'realized industrial optionality vs. empire-construction discount'; INTC benefits from process validation and strategic relevance, while TSLA remains vulnerable to conglomerate-overhang sentiment. Exit if TSLA demonstrates no capital or governance spillover from the venture.
• Do not chase NVDA here; keep it on a watchlist rather than a trade. If orbital AI remains speculative, capital is more likely to stay with terrestrial inference demand than to shift meaningfully to a new compute platform over the next 6-12 months.
• Set an alert on SPCX for any disclosed capex funding structure or customer prepayments. That is the key falsifier: if external financing or binding demand appears, the short thesis loses asymmetry and the trade should be covered quickly.