SpaceX’s near-term AI payoff seen tethered to Earth, not outer space

Reuters highlights analysts at Barclays and others arguing SpaceX’s next stock-driving leg is earnings from terrestrial AI compute, not near-term orbital AI disruption. Reuters calculations estimate its latest Colossus compute contracts (e.g., with Anthropic, Google, and Reflection AI) could generate $28B+ in annual revenue versus ~ $3.2B in 2025 AI revenue, with J.P. Morgan expecting ~9 GW of terrestrial capacity by 2029 (about 4x Hoover Dam power). Analysts caution contracts may include termination provisions, but spending is heavily weighted to AI—nearly $18B in 2025 (about $12.7B capex and $5.1B R&D)—supporting a bullish near-to-mid-term fundamentals narrative.

Analysis

This is best read as a duration extension for the AI buildout, not a new demand shock. The equity implication is that the bottleneck is still physical infrastructure and power, so the incremental winners are the names that monetize watts, racks, cooling, and chips—not the ones trying to price in far-future science projects. That keeps the AI capex cycle alive for another 12-24 months, but it also means the market should focus on utilization and renewal rates rather than headline revenue run-rate claims.

The quality of the economics matters more than the scale. If customer contracts can be terminated or repriced, the market will eventually haircut the implied revenue stream, so this is more of an earnings-quality and guidance story than a pure growth story. In the next 1-3 months, watch whether large cloud/platform names keep lifting capex without compressing operating leverage; if they do, the current re-rating in AI infrastructure has room to continue.

For GOOGL, the read-through is constructive but nuanced: stronger enterprise AI demand supports cloud and custom silicon demand, yet the market will penalize any sign that AI investment is outrunning monetization. The contrarian point is that orbital compute is probably too far out to matter for trading, while terrestrial power scarcity remains underappreciated and should be the real factor driving relative performance among AI beneficiaries.

Key Decisions for Investors

• Long GOOGL on 3-5% weakness into earnings/capex updates; 1-3 month horizon. Thesis is that AI demand is still elastic enough to support cloud and TPU investment. Falsifier: cloud growth decelerates while capex steps up again.
• Prefer long XLU/NEE/CEG-style power and grid beneficiaries over high-multiple AI software for the next 3-6 months. The second-order winner is the electricity stack, not the speculative application layer; risk/reward favors infrastructure with visible backlog over names trading on 2030 narratives.
• Avoid forcing a trade in BAC, BCS, ETST, or WWRL on this headline alone. The linkage is too indirect until there is evidence of financing, project capital markets, or direct exposure to AI infrastructure spend.
• Set an alert on GOOGL if operating margin or free-cash-flow margin inflects lower despite higher AI revenue commentary. That would signal the market is overpaying for AI optionality and should reverse the near-term bullish read-through.