Should You Buy SpaceX Stock in September?

SpaceX shares have fallen 34% from their $225 peak to $147.95, despite Q2 2026 revenue rising 92% year over year to $7.8 billion, led by 247% growth in AI revenue to $2.56 billion. The company trades at a steep 86.9x trailing sales on a $2 trillion market capitalization, though its valuation falls to 19x 2027 revenue if Wall Street's $105 billion forecast is achieved. The article sees limited potential for near-term gains but significant long-term upside if SpaceX can execute Musk's forecast of $1 trillion in annual revenue by 2030.

Analysis

SPCX is being priced as a conglomerate with three separate terminal-value assumptions: launch leadership, Starlink monetization, and hyperscale AI. The key underwriting issue is not top-line growth but whether AI compute revenue carries software-like gross margins after depreciation, satellite replacement capex, spectrum costs, and the cost of serving latency-sensitive workloads from orbit. If the AI segment is principally capacity rental, the relevant public comparables are GPU clouds and data-center operators rather than NVDA; that would justify materially lower revenue multiples and makes consensus revenue forecasts insufficient for valuation.

Over the next 1-3 months, the stock is vulnerable to any disclosure that separates AI bookings from recognized revenue, reveals customer concentration, or raises capex intensity. A 20% miss to the implied 2027 revenue path, with the sales multiple merely compressing toward premium infrastructure peers, could create 35-50% downside even while revenue remains high-growth. Conversely, independently verifiable long-duration compute contracts, positive segment contribution margin, and evidence that Starlink distribution lowers customer-acquisition cost would support the premium.

Second-order effects favor NVDA only if orbital deployments translate into incremental accelerator demand rather than recycling hardware purchased by centralized cloud operators. GOOG faces limited near-term earnings risk: Starlink is more likely a connectivity complement in underserved markets than a substitute for terrestrial cloud, but a credible low-cost global backhaul offering could pressure long-run edge-network economics. The contrarian view is that the drawdown may still understate financing risk: rapid revenue growth can mask a capital cycle where each incremental dollar requires recurring launches, replacement satellites, and accelerator refreshes.

Key Decisions for Investors

• Do not initiate a directional SPCX long until the next filing provides segment gross margin, capex, backlog duration, and related-party/customer concentration. Treat verified AI recurring revenue and positive contribution margin as the entry trigger; absent those data, the valuation case is narrative-driven.
• For a 1-3 month hedge against a de-rating, consider a defined-risk SPCX put spread sized to a 30-40% downside scenario rather than an outright short. Cover if management discloses contracted multi-year AI backlog sufficient to support utilization through the next accelerator-refresh cycle.
• Maintain NVDA exposure separately from SPCX: add only on evidence of incremental accelerator orders and launch cadence, not on announced orbital-AI ambitions. The thesis is falsified if disclosed deployments use prior-generation inventory or if supplier revenue does not show a corresponding uplift within two quarters.
• Watch a relative-value long NVDA / short SPCX basket if SPCX trades materially above premium AI-infrastructure revenue multiples without demonstrated free-cash-flow conversion. Reassess after the first post-IPO earnings report; segment capex-to-revenue above expectations would be the primary confirmation.