SpaceX's Out-of-This-World Plan Has Flaws. That Bodes Well for This Terrestial Stock.

Wood Mackenzie estimates a 1-gigawatt orbital data center would cost about $170 billion, more than three times an equivalent terrestrial facility, implying SpaceX’s space-based AI infrastructure is not economically competitive near term. The article argues terrestrial AI buildout will continue to drive electricity demand, with American Electric Power highlighted as a beneficiary given its 5.6 GW data center load contracts in Ohio and 2.8% dividend yield. Overall, the piece is a speculative long-term thesis on space-based computing versus a nearer-term bullish case for utility stocks.

Analysis

The key market takeaway is not the moonshot itself, but the duration of terrestrial capex intensity it implies. Even if orbital compute becomes technically feasible, the adoption curve is likely to be gated by launch economics, repairability, and insuranceability, which keeps AI workloads anchored to land-based power and fiber for most of the next decade. That extends the runway for regulated utilities with transmission bottlenecks, because the scarce asset is no longer just generation — it is interconnection and high-voltage delivery to dense load pockets.

AEP is interesting because it monetizes the bottleneck rather than the megawatt itself. Data-center tariff structures reduce volume risk, while its transmission footprint gives it pricing power in a capital-intensive buildout where new load often outruns new wires. The second-order winner is the supply chain around grid reinforcement — transformers, switchgear, conductors, and EPC capacity — which should stay tight if AI load forecasts continue to get pushed higher, even if some hyperscaler projects are delayed.

The contrarian risk is that this becomes a sentiment trade on AI power scarcity rather than a durable earnings surprise. If regulators push back on tariff pass-through, or if hyperscalers self-generate through behind-the-meter gas/nuclear deals, the utility-beta trade can fade quickly over 3-9 months. Also, if orbital compute claims catalyze a broader rethink of data-center siting, the market may rotate from utilities toward grid equipment and industrial enablers instead of staying concentrated in AEP.

For now, the setup argues for owning the bottleneck with defined downside rather than chasing speculative AI infrastructure names. The asymmetry is best expressed through regulated transmission and equipment suppliers, while avoiding any assumption that orbital data centers will matter to cash flows before the early 2030s at the soonest.

Key Decisions for Investors

• Long AEP on a 6-12 month horizon; use pullbacks to build. Risk/reward is favorable if data-center load bookings continue, with downside cushioned by dividend and tariff protection.
• Pair trade: long AEP / short high-multiple AI infrastructure proxies that depend on near-term orbital or novel power narratives. This isolates the bottleneck beneficiary from speculative execution risk.
• Add a basket of grid-capex beneficiaries on weakness (e.g., transformer/switchgear suppliers such as ETN, HUBB, PWR) for 12-24 months; the second-order thesis is transmission scarcity, not just electricity demand.
• If you want convexity, buy AEP call spreads 9-12 months out rather than outright shares; upside comes from accelerated load contracts, while premium paid limits regulatory or project-cancel risk.
• Avoid chasing space-linked AI compute concepts until the cost curve visibly compresses; the implied timeline is multi-year, and the nearer-term winner is the land-based power stack.