SpaceX Could Have Been Profitable in Q2 -- But Then AI Happened

SpaceX beat Q2 sales and loss expectations (revenue up 92% YoY to $7.8B vs. $6.8B; EPS loss of -$0.09 vs. -$0.29) but shares fell 13.6% before rebounding. Starlink grew subscribers to 12M and operating profit +79%, while the Space launch unit’s losses rose 47% to $542M. The AI division drove heavy cash burn—AI revenue +247% to $2.6B but losses at $1.3B—contributing to $16B cash burned in Q2 and -$25B used in the first half of 2026, with another 7% of insider shares unlocking on Aug. 21 after earlier unlocks.

Analysis

The market is still pricing SPCX like a scarcity asset, but the better lens is capital allocation risk: a profitable core is being diluted by a loss-making adjacent bet whose economics do not improve with scale. That matters because the equity story can only sustain a premium multiple if investors believe incremental dollars are earning above-cost returns; if not, the stock should trade more like a bundled sum of parts with a discount on the AI unit’s cash burn. The first unlock is less about raw share supply than about signaling: insiders taking any liquidity at the first window tends to cap upside when sentiment is already stretched.

Near term, the stock can remain technically supported if the post-earnings dip was absorbed and the market is willing to look through lockup expiry. Over 1-3 months, though, the next catalyst is not another sales beat; it is whether the company can prove the AI losses are contained. If AI spending stays heavy, the core satellite and launch businesses effectively become funding sources for a lower-quality growth narrative, which should pressure the multiple and favor competitors with cleaner unit economics.

The second-order winner is upstream AI infrastructure, not the integrator itself: continued AI losses likely mean more demand for compute, networking, and cloud capacity, which is a relative positive for NVDA versus SPCX on a 6-18 month view. The contrarian risk is that consensus may be overestimating insider selling pressure; if the float absorbs supply and management reframes AI as a strategic option rather than a cash drain, the stock could keep grinding higher. The bearish thesis is falsified if SPCX holds above the mid-$150s through the next unlock window and Q3 shows a material reduction in AI cash burn or a narrower net loss.

Key Decisions for Investors

• Short SPCX on strength into the next unlock window; prefer a 1-3 month put spread rather than outright short to define risk. Risk/reward improves if the stock remains near or above $150 and fails to make a new high after the supply overhang clears.
• Pair trade: long NVDA / short SPCX as a relative-value expression of AI capex flowing to picks-and-shovels rather than vertically integrated cash burners. Hold 3-6 months; exit if SPCX shows two consecutive quarters of materially lower AI losses.
• Set a tactical alert to cover part of the SPCX short if shares hold above the mid-$150s for 5+ trading sessions after the unlock, which would indicate the market is fully absorbing insider supply.
• If you want no-single-name risk, use a sector-neutral hedge: short a small basket of high-multiple unprofitable growth names against SPCX only if the next filing confirms continued heavy AI cash usage.