We're adding a new, high-profile AI play to our Bullpen stocks to watch list

SpaceX reported adjusted Q3 EPS of $0.35, above analyst forecasts of about $0.13, while SI revenue rose to $7.56 billion from $2.56 billion in Q2 and $813 million a year earlier. Analysts’ 2027 adjusted EPS estimate increased to $3.71 from $1.59 in September; the article says the shares trade at 38.5 times that estimate, versus 451 times 2026 estimates. The bullish valuation case depends on execution, including SI revenue rising from an estimated $26.7 billion in 2026 to $132 billion in 2027; management expects to add 2.5–3 GW of compute by end-2026 and another 10 GW in 2027.

Analysis

The key valuation risk is not whether AI demand exists, but whether SpaceX can convert announced capacity into billable, profitable compute fast enough to support the steep 2027 earnings step-up. “Up to” contract value is not recognized revenue: verify minimum commitments, delivery schedules, cancellation terms, utilization, power and accelerator availability, and incremental capex before treating Anthropic (ANT) as a de-risking anchor. A delay would hit both the earnings bridge and the multiple at once; the scale-up claim also creates execution and funding sensitivity not captured by headline TAMs.

Starlink’s integrated launch capability is a durable cost advantage, but competitive pressure on Comcast (CMCSA) and AT&T (T) is likely concentrated in underserved areas rather than a near-term nationwide broadband displacement. Carrier partnerships may complement wireless networks as much as replace them. Airline enthusiasm could help United Airlines (UAL) differentiate service, but one executive’s comment does not establish material pricing power or retention economics. Do not short incumbents solely on this evidence.

Near term, volatility and changing estimates can dominate fundamentals. Over 1–3 months, watch compute delivery, contract disclosure, and earnings revisions; over 6–18 months, power/capacity execution and Starship reusability determine whether growth converts to cash. The contrarian point: a forward multiple that appears reasonable depends on an unusually sharp forecast ramp, so estimate revisions may be more informative than the quoted multiple. The possible Tesla (TSLA) combination is speculation, not a catalyst to underwrite.

Key Decisions for Investors

• SPCX: no immediate full-size chase. Consider only a staged, high-risk position after verifying compute capacity coming online, contracted minimum spend, and cash-flow/capex disclosure; cap exposure to what a failed 2027 ramp can tolerate.
• Set an alert to reduce or avoid SPCX if capacity milestones slip, Anthropic commitments prove non-binding or back-loaded, or forward earnings estimates begin to fall. Reassess if utilization and cash generation validate the ramp rather than relying on total contract value.
• Do not initiate a short in CMCSA or T on this article alone: likely substitution is geographically bounded and Starlink can be complementary to terrestrial and mobile networks. Revisit only with evidence of subscriber losses, pricing pressure, or lower broadband guidance.
• Treat the TSLA merger possibility as unpriced optionality, not a standalone TSLA or SPCX trade catalyst; require formal terms before changing exposure.