Key insights
- Alphabet (GOOG) is testing its proprietary Tensor Processing Units (TPUs) in orbit with Project Suncatcher, aiming for full control of the space-based AI compute stack. NVIDIA (NVDA) is the exclusive chip supplier for SpaceX's orbital data center ambitions, potentially involving a million compute satellites, and holds a stake in SpaceX. Both companies are positioned to benefit from the emerging orbital data center market, with NVIDIA seeing additional GPU demand and Google potentially reducing reliance on external chip suppliers.
Investing.com -- The orbital data center thesis just got its most credible test yet: on Sep 24, 2026, Alphabet/Google (GOOG) announced Project Suncatcher — launching a prototype satellite carrying its own Tensor Processing Units (TPUs) into Low Earth Orbit (LEO) via SpaceX’s Transporter-18 rideshare. Meanwhile, NVIDIA (NVDA) has been named the exclusive chip supplier for SpaceX’s orbital data center ambitions — a constellation potentially scaling to one million compute satellites. Read more
Two very different plays, two very different risk profiles. Here’s the decode:
The SpaceX plan is architecturally simple: swap Starlink comms hardware for NVIDIA’s computing hardware, add solar panels, and let orbital physics do the rest — continuous sunlight, vacuum cooling, no land or grid battles. NVIDIA isn’t just a supplier here; it holds a sizable stake in SpaceX, aligning shareholder interests more deeply than a typical vendor relationship. Any buildout flows directly into fresh GPU demand on top of an already insatiable earthbound order book. Read more
Google’s TPU angle is structurally different — and cleverly hedged. Rather than buying chips from a rival, Alphabet tests its proprietary AI accelerators in orbit, aiming to own the full compute stack if space-based AI takes off. Project Suncatcher’s 2027 laser-link satellite tests will determine whether high-bandwidth inter-satellite connections — the nervous system of any orbital compute cluster — are viable. Read more
Screener/FinQL values are snapshots and may lag live prices.
Bull case for NVDA: Exclusive chip designation for SpaceX’s orbital constellation + earthbound data center demand already printing $215.9B in FY2026 revenue. A 29.3% fair value upside at current prices means the space thesis is essentially a free option on top of an already-undervalued growth engine.
Bull case for GOOG: Proprietary TPUs in orbit = no dependency on external chip suppliers. If Project Suncatcher succeeds, Google controls both the orbital compute infrastructure AND the AI workloads running on it. At 16.9x P/E, it’s the cheaper entry.
Bear reality check: IEEE Spectrum estimates a 1-gigawatt orbital data center would cost $50B+ across ~4,300 satellites. Radiation hardening, vacuum cooling, and repair impossibility are formidable walls. Most analysts don’t model meaningful orbital revenue before 2029 at the earliest. Neither stock’s near-term price is driven by this thesis — it’s a multi-year narrative catalyst, not a quarterly earnings lever.
NVIDIA is the clearest near-term beneficiary — exclusive chip status, a deep SpaceX relationship, a 29.3% fair value upside, and 92.9% ROIC that proves it converts every dollar of space-or-earth demand into strong returns. Google is the more intriguing long-term wildcard if its vertically-integrated TPU stack proves space-worthy.
The real dark horse to watch: SEALSQ (LAES), which is positioning quantum-resistant chips and secure microcontrollers specifically for orbital platforms — a niche play on the security layer of space computing. Read more
This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.
ProPicks AI re-scores NVDA every month against thousands of alternatives on 100+ financial metrics, then builds the strategies that came out ahead.Energy Elite is beating its benchmark by 37% year to date. Mid Cap Movers by 21%. Tech Titans, our flagship, has beaten the S&P 500 by 112% since launch three years ago.See whether NVDA made the cut — and which stocks did.