SpaceX claims it will put a Vera Rubin NVL72 rack-scale system into orbit next year

Elon Musk said SpaceX will launch at least one datacenter rack in 2025, with “significant scale” planned for 2028, tied to Nvidia’s space-optimized Vera Rubin NVL72 (72 GPUs, 36 CPUs, 120kW liquid-cooled racks). Nvidia’s design calls for a 210kW solar array and liquid radiators, but SpaceX has not yet built a full satellite, leaving delivery and scalability unproven. The news is a constructive technology signal, but timing and “space-optimized” scope are unclear, limiting near-term confidence.

Analysis

This reads as a narrative extension of NVDA’s platform story, not a near-term earnings driver. The market may give it a small multiple bump because it reinforces the idea that NVDA is the default architecture wherever compute migrates, but the commercial path is so far out that any P&L impact is effectively zero for the next several quarters. The real value is optionality: if the company can prove it owns the thermal/radiation/reliability stack, it can deepen switching costs and broaden the moat, even before revenue shows up.

The more important second-order effect is that the economic beneficiaries, if this ever scales, are not the chip vendor first; they are launch cadence, space power, and cooling ecosystems. That means the market should be cautious about extrapolating this into a meaningful NVDA revenue line item before there is a verified manifest, customer commitment, or prototype performance data. Absent that, this behaves like a high-beta headline that can support sentiment for a few sessions but is unlikely to move consensus estimates.

Contrarian view: investors may be underpricing execution risk and overpricing “infinite TAM” optics. A multi-year delay, launch failure, or inability to harden and cool the system at scale would collapse the story quickly. The falsifier is simple: no quantified orders, no launch schedule, and no revenue disclosure by the next major NVDA catalyst window; if that remains the case, the trade belongs in the no-trade bucket rather than the fundamental long book.

Key Decisions for Investors

• Do not initiate fresh NVDA longs purely on this headline; if already long, use any 1-3 day pop to trim or overwrite calls. Time horizon mismatch is the key risk: the story is 2028+, while the stock will re-anchor to FY26/FY27 estimates within days.
• For tactical traders, sell short-dated NVDA upside via call spreads if implied volatility lifts on the news. Risk/reward favors theta decay because there is no obvious estimate revision catalyst; invalidate if NVDA holds a new 20-day high on strong volume or management quantifies contracted volume.
• Set a watch item for the next NVDA earnings call / GTC: only upgrade the thesis if management provides customer commitments, launch cadence, or revenue language tied to non-terrestrial compute. Until then treat this as strategic optionality, not a tradeable fundamental.