Prediction: SpaceX Shares Can Reach $220 by End of 2026

SpaceX shares are near $149 (just above the $135 IPO price) after pulling back from a ~$225.6 post-IPO peak, as investors refocus on Starlink expansion, Starship progress, and AI compute deals. The article cites 2025 revenue up 33% YoY to $18.7B but a $4.9B net loss, alongside Alphabet paying $920M/month (Oct 2026–Jun 2029) for ~110,000 Nvidia GPUs and an Anthropic compute-access deal, which Reuters estimates could total ~$26B annually if fully realized. Valuation remains expensive (~51.4x base-case 2026 sales) and the stock’s end-2026 price is framed around a likely 20–25% multiple compression offset by revenue growth, implying ~$213–$227 per share (base case ~$220).

Analysis

The cleanest read-through is that NVDA is the real economic toll collector here. Multi-year compute commitments, even if they land inside a private-space narrative, still validate GPU scarcity and extend pricing power for the silicon layer; the risk is less demand and more supply allocation, which supports NVDA’s multiple better than any single “AI platform” story supports SPCX’s.

SPCX is the harder asset to underwrite because the market is already paying for several years of execution that are not yet observable in cash flow. That creates a classic post-IPO setup: near-term enthusiasm can persist for weeks, but over 1-3 months the stock likely trades on evidence of backlog conversion, launch cadence, and disclosed margin structure rather than headline TAM; absent that, multiple compression is the more probable path. The biggest second-order risk is capital intensity: any delay in Starship or monetization cadence forces investors to capitalize growth farther out, which is where high-multiple names de-rate fastest.

GOOGL is a mixed beneficiary: strategic insurance against compute scarcity, but also a signal that hyperscalers may keep renting capacity instead of internalizing all of it, which can cap margin expansion if external procurement scales. The contrarian miss is that “big contract” does not equal near-term earnings power; the value is back-loaded, and the market may be overestimating how much of this turns into visible 2026-2027 free cash flow versus headline bookings. If the next 1-2 public updates fail to show sharper revenue conversion or Starship progress, the post-IPO premium should compress materially.

Key Decisions for Investors

• Long NVDA vs short SPCX for a 3-6 month relative-value trade: own the more certain beneficiary of AI compute demand while fading the richer, execution-dependent multiple; thesis breaks if SPCX starts showing faster-than-expected revenue conversion and launch cadence.
• Do not chase SPCX strength above the current post-IPO range; wait for the first hard disclosure on contract conversion and margin economics. If that evidence is absent by the next update cycle, use rallies to fade or structure put spreads.
• Treat GOOGL as a watchlist name, not a conviction long: external compute procurement can be a capacity hedge, but the earnings impact is likely muted unless it clearly lowers time-to-inference/training spend. Add only on weakness if subsequent disclosures show better capital efficiency.
• Set a catalyst alert on Starship/test cadence and any filed details on the compute contracts over the next 1-3 months; if either slips, reduce any long SPCX exposure immediately because the valuation depends on the market believing the back-half ramp.
• If you need a cleaner expression of the AI infrastructure beneficiary set, prefer NVDA over any space-platform proxy until the revenue mix becomes visible in audited filings.