SpaceX's AI infrastructure plans are not yet fully priced in, analysts say

Bernstein raised its SpaceX price target to $268 from $248, implying about 82% upside from Friday's close, while Deutsche Bank lifted its target to $250 from $230, implying nearly 70% upside. The analysts cited SpaceX's terrestrial AI buildout and potential compute advantage; Deutsche Bank said the company aims to bring roughly 10 gigawatts of incremental computing capacity online by the end of next year and estimated its compute could be priced at about $35 billion per gigawatt versus an industry standard of $50 billion. LSEG data show 33 of 43 covering analysts rate the stock buy or strong buy; shares were roughly flat over the past month.

Analysis

The investable question is not whether AI demand is large, but whether SpaceX can turn planned capacity into energized, utilized compute with attractive returns on incremental capital. Analyst targets are not evidence of customer commitments or unit economics; in particular, verify what the cited per-gigawatt figures measure before using them in a valuation. Capacity announcements without power access, hardware delivery, networking, utilization and contracted pricing could create a capex story before they create cash flow.

Near term (days to weeks), broad buy-side analyst support raises the risk that a positive narrative is already crowded, even without a recent share-price surge. Over 1–3 months, the key catalysts are verifiable customer contracts, capacity actually brought online, and disclosure of capex and utilization. If execution is credible, terrestrial compute could add a nearer-term monetization path than orbital data centers and challenge conventional AI infrastructure providers. Conversely, delayed power or equipment, weak customer demand, or aggressive spending would make the growth narrative a balance-sheet and return-on-capital concern. Over 6–18 months, sustained low-cost capacity could pressure incumbent compute providers, but the durability of any cost edge will depend on power, networking and hardware economics—not headline capacity alone.

Contrarian view: investors may be valuing a gigawatt ambition as if it were productive, revenue-generating capacity. There is no basis here to underwrite the targets without validating the metric, funding needs and commercial commitments.

Key Decisions for Investors

• Avoid chasing SPCX solely on analyst target increases. Treat the setup as execution-dependent; verify the definition and source of the per-gigawatt estimate, committed customers, capex, power availability and delivery schedule.
• Set an alert for the next company update or earnings disclosure: upgrade the thesis only if capacity is energized on schedule and accompanied by credible utilization or contracted revenue. A capacity target without monetization is not confirmation.
• For investors seeking exposure, consider a staged SPCX position rather than a full entry ahead of evidence; define risk around a missed capacity milestone or guidance that indicates materially higher spending without customer commitments. No numeric target is supportable from the supplied information.
• Watch conventional AI infrastructure providers for relative weakness only if SpaceX demonstrates delivered, cost-competitive compute. Until then, a short against those providers is premature: the article supplies no verified operating-cost comparison or evidence of customer substitution.