SpaceX Stock Is Down 50% From Its Peak. Here's 1 Reason to Buy the Dip, and 1 Reason to Run for the Hills.

SpaceX (Space Exploration Technologies) stock has given investors a rollercoaster ride—rallying to $225 before falling ~50% to about $110 as of Aug. 6—highlighting valuation risk despite strong growth. In Q2 2026 it generated $7.8B revenue (+92% YoY), including Connectivity $4.29B (+66% YoY) and AI $2.56B (+247% YoY). However, with trailing-12-month revenue of ~$23B versus a ~$1.5T market cap, the article cites a ~65x P/S (vs. Nasdaq-100 ~6.4x), arguing forward P/S could fall to ~10 only if the AI business reaches a $100B annual run rate by end-2026.

Analysis

SPCX is being priced less like a diversified growth asset and more like a long-duration call option on flawless execution across three capital-intensive businesses. The near-term issue is not TAM; it is whether the new AI revenue is recurring, high utilization, and high-margin enough to justify a premium multiple after depreciation, launch spend, and working-capital drag. If the next 1-3 months do not show backlog conversion, the stock can keep de-rating even with strong top-line growth.

The clearest second-order beneficiary is NVDA, because every credible version of the orbital-AI story still needs accelerators, networking, and storage. GOOGL is a tactical counterparty beneficiary if it is leasing or consuming compute, but it is not the cleanest way to express this theme. On the loser side, satellite broadband and enterprise connectivity peers such as VSAT and IRDM face margin pressure first, then share loss, because a bandwidth step-up can force price competition before volumes fully inflect.

The contrarian mistake in the market is treating headline revenue as proof of economic durability. The more important variable over 6-18 months is free cash flow conversion: if capex and launch cadence scale faster than contract monetization, the AI narrative can destroy value even while revenue accelerates. The thesis is falsified if backlog does not translate into utilization by the next print, or if launch/satellite deployment slips enough to push the 2027 revenue ramp materially right.

Key Decisions for Investors

• Avoid initiating a new long in SPCX until the next reported quarter confirms AI backlog conversion and utilization; if that evidence is absent, expect continued multiple compression over the next 1-3 months.
• Long NVDA / short SPCX as a 3-6 month pair: NVDA captures the AI capex spend regardless of which infrastructure model wins, while SPCX still has to prove economics; attractive if SPCX stays above 50x sales.
• Use GOOGL only as a secondary AI infra beneficiary, not as the main SpaceX expression; buy on weakness for core cloud exposure, but do not overpay for the SpaceX adjacency.
• Watch VSAT and IRDM as short candidates into any verified V3 launch cadence; if Starlink bandwidth improves faster than enterprise pricing power, these names can see 10-15% downside before fundamentals fully reflect it.
• Set a hard thesis check on SPCX around the next earnings/backlog update: if reported AI bookings do not visibly shorten the path to a lower forward sales multiple, reduce exposure rather than averaging down.