SpaceX: Super Intelligence Segment Turns It Into An AI Infrastructure Heavyweight

SpaceX was rated Strong Buy with a $337-per-share price target, citing plans to reach at least 10 GW of compute capacity by eFY27 and strong GPU rental rates. The article says Super Intelligence could add margin-accretive growth as capacity scales, with potential further support from Terafab in coming years.

Analysis

The upside case hinges less on headline compute capacity than on whether SpaceX can secure power, GPUs, and customer utilization at attractive returns on invested capital. Scaling capacity faster than grid connections or demand would turn a growth narrative into a capital-intensity and execution discount; falling rental rates would amplify that risk. Near term, the optimistic framing may support sentiment, but the supplied information does not establish funding, contracted power, committed customers, or independently verified unit economics. Over 1–3 months, look for evidence on those items and any revised deployment schedule. Over 6–18 months, successful execution could create a distribution advantage if connectivity and compute are sold together; telco ambitions also bring regulatory, spectrum, and incumbent-response risks. Terafab should be treated as optionality, not base-case value, until scope, funding, and timing are disclosed. A contrarian risk is that investors capitalize a distant capacity target before demand and power infrastructure are secured. The thesis weakens materially if deployment slips, utilization disappoints, GPU rental rates fall, or funding needs rise without corresponding contracted revenue.

Key Decisions for Investors

• Do not trade the price target alone: verify the instrument, liquidity, current valuation, and whether the compute build has secured power, GPUs, financing, and customer commitments before sizing a position.
• For an existing SPCX exposure, keep any incremental risk small until deployment milestones and utilization economics are independently observable; reassess on schedule delays, weaker rental rates, or materially higher funding requirements.
• Watch Nvidia and established cloud providers as potential competitive and demand benchmarks, not automatic shorts: sustained rental-rate declines would challenge the economics of new capacity across the sector, while resilient rates and contracted demand would support the buildout thesis.
• Treat connectivity-compute bundling and Terafab as longer-dated catalysts rather than near-term earnings drivers. Upgrade conviction only with regulatory progress and disclosed evidence that customer demand and infrastructure are scaling together.