This is a sample of CapRelay's research. CapRelay provides similar research and more for every US-listed company.

Start free trial
SpaceX -

Bear Case

SPCX
Industry:
Telecom Aerospace & Defense Software

Thesis Summary

SpaceX has built a genuinely dominant space and connectivity business, but the xAI acquisition has transformed it into something far more complicated and risky. The core bearish thesis: SpaceX is now a massively capital-intensive conglomerate running three simultaneous high-stakes bets, funded by a single cash engine — Starlink — whose unit economics are visibly deteriorating. The AI segment is absorbing capital at a scale that dwarfs anything SpaceX has attempted before, with returns uncertain at best. Meanwhile, the single asset tying everything together — Starship — has not yet delivered a single commercial payload after 12 flight tests.

Key thesis points:

  • Starlink ARPU is collapsing and the subscriber growth needed to compensate is slowing. The cash engine is deteriorating faster than expected.
  • The AI bet is uniquely dangerous. SpaceX acquired a deeply loss-making AI company and is now one of the largest AI compute spenders in the world, competing against entrenched hyperscalers with no meaningful enterprise AI relationships.
  • X advertising revenue is in secular decline. The foundational data moat for Grok is eroding.
  • Starship is the single point of failure for everything. Every long-term growth driver — V3 satellites, orbital AI compute, Starlink Mobile Gen2 — depends on a vehicle still in flight testing.
  • Debt and capital allocation complexity have increased significantly. The $20B SpaceX Bridge Loan matures in September 2027, and the company is simultaneously funding Starship, Starlink, AI data centers, and a $19.6B spectrum acquisition.

Starlink's Cash Engine Is Deteriorating

Starlink is the company's only segment generating material cash today, but its unit economics are moving in an unfavorable direction that will be difficult to reverse without Starship:

  • Monthly Starlink Subscriber ARPU fell from $99 in FY2023 to $91 in FY2024 to $81 in FY2025, and further to $66 in Q1 2026 — a 33% decline in roughly two years.
  • SpaceX has explicitly guided that ARPU will continue declining as international expansion into lower-income markets accelerates and lower-priced service tiers are added.
  • The offset mechanism — deploying higher-capacity V3 satellites at lower cost-per-bit — requires Starship, which has not yet delivered a commercial payload.
  • While subscriber growth remains strong (105% YoY to 10.3M as of March 2026), growth rates will inevitably slow as the highest-willingness-to-pay markets (rural North America, Western Europe, Australia) approach saturation. Incremental subscribers will come from progressively lower-ARPU geographies.
  • Connectivity Segment Adjusted EBITDA grew to $7.2B in FY2025 and $2.1B in Q1 2026, but with ARPU now at $66/month and continuing to fall, the volume required to sustain EBITDA growth compounds with every passing quarter.

The model works while subscriber growth offsets ARPU compression. The moment growth slows — and it will, as penetration in addressable markets rises — the entire capital allocation flywheel stalls.

The AI Segment is a Massive, Unproven Bet Against Entrenched Incumbents

The xAI acquisition has placed SpaceX in direct competition with the best-capitalized technology companies on Earth in the most competitive software market in history, at a time when SpaceX is simultaneously funding Starship and Starlink:

  • AI segment capex was $12.7B in FY2025 and $7.7B in Q1 2026 alone, annualizing above $30B. This compares to $7.2B in Connectivity Segment Adjusted EBITDA generated in all of FY2025.
  • The AI segment generated an operating loss of $(6.4)B in FY2025 and $(2.5)B in Q1 2026. SpaceX acknowledges a "multi-year investment horizon" before AI Segment Adjusted EBITDA turns positive.
  • Grok competes against OpenAI (backed by Microsoft), Google (Gemini), Meta (Llama, open-source), and Anthropic (backed by Google and AWS) — all with larger developer ecosystems, deeper enterprise relationships, and more proven monetization models.
  • As of March 31, 2026, only 117M of 550M MAUs used Grok's AI features, and paid subscribers were approximately 6.3M — a thin monetization base relative to the billions in capital being deployed.
  • The Anthropic cloud services agreement ($1.25B/month through May 2029) provides near-term compute monetization, but is terminable by either party on 90 days' notice and does not represent committed long-term revenue. It also cements SpaceX's role as a commodity compute provider rather than a differentiated AI platform.
  • AI model performance benchmarks change rapidly. Grok's claimed advantages — real-time X data integration, compute speed — can be replicated or made irrelevant by competitors. OpenAI, Google, and Meta are also investing heavily in compute infrastructure.

Competitive Bear Case: X Advertising is in Secular Decline, Eroding Grok's Core Differentiator

Grok's primary claimed competitive differentiator is real-time integration with X's approximately 350 million daily posts. But the X platform itself is structurally challenged:

  • X advertising revenue fell 11.5% from FY2023 to FY2024 as advertisers departed following Elon Musk's acquisition of Twitter. Revenue continued declining: Q1 2026 advertising revenue was $343M, down 23% YoY from $443M in Q1 2025.
  • Major advertisers have not returned to the platform at scale, and brand safety concerns remain unresolved.
  • X competes for advertising dollars against Meta, Google, TikTok, and YouTube, all of which have larger, more engaged audiences and more sophisticated performance advertising tools.
  • A declining and disengaging user base on X would reduce both the quantity and quality of real-time data available to train and differentiate Grok — the one moat the AI segment can claim.
  • The company is rebuilding X's advertising platform from scratch (the new X Ads Manager rolled out in April 2026), introducing material execution risk and near-term revenue disruption; Q1 2026 advertising revenue fell partly due to an advertising platform overhaul.

Starship is the Single Point of Failure for Every Long-Term Growth Driver

Every meaningful long-term revenue opportunity — V3 Starlink satellites, Starlink Mobile Gen2, orbital AI compute, the lunar economy — depends entirely on Starship achieving commercial operations at scale:

  • SpaceX has executed 12 Starship flight tests as of May 2026. Payload delivery to orbit is expected to begin in the second half of 2026, but SpaceX acknowledges that multiple flight tests have resulted in FAA mishap determinations, each requiring return-to-flight conditions.
  • V3 Starlink satellites (1 Tbps downlink per satellite vs. ~50 Gbps for V2 Mini) cannot be deployed on Falcon 9. Without V3, the cost-per-bit improvement needed to sustain Connectivity margins at lower ARPU does not materialize.
  • V2 Mobile satellites for Starlink Mobile Gen2 also require Starship. The $19.6B EchoStar spectrum acquisition is largely stranded value without V2 Mobile deployment.
  • Orbital AI compute — SpaceX's most ambitious long-term argument for AI differentiation — requires Starship at scale with full reusability. This technology is untested and targeted for first deployment as early as 2028, a timeline with no margin for error.
  • SpaceX has spent over $15B on Starship development and is spending $3.0B per year in R&D on it. This spending caused the Space Segment Adjusted EBITDA to decline from $1.2B in FY2024 to $653M in FY2025, and the segment posted a $(351)M Adjusted EBITDA loss in Q1 2026.

A meaningful delay in Starship's commercial operations cascades immediately across all three segments: Connectivity margins worsen, the spectrum acquisition sits idle, and orbital AI compute becomes even more speculative.

Leverage and Capital Allocation Complexity Create Structural Fragility

SpaceX is now running three simultaneous capital-intensive investment cycles while carrying $29.1B in debt:

  • The $20B SpaceX Bridge Loan matures September 2, 2027 (with extension options to March 2028). SpaceX is contractually required to apply IPO proceeds toward repayment within six months of receipt — alongside funding AI compute expansion, Starship, and Starlink constellation growth simultaneously. The allocation among these uses is not committed.
  • Interest expense was $1.9B in FY2025 and $664M in Q1 2026, representing a persistent cash drain on a business generating $6.8B in operating cash flow in FY2025.
  • Total capex was $20.7B in FY2025 and $10.1B in Q1 2026. At Q1 2026 annualized rates, capex is approaching $40B per year — a level that dwarfs operating cash flow.
  • The company has committed to $19.6B in equity and cash consideration for EchoStar spectrum licenses, with the equity portion alone representing ~261.8M new Class A shares to be issued at closing in November 2027.
  • Additional obligations include the Cursor option (if exercised, approximately $60B in Class A stock) and three equipment sale-leaseback transactions with Valor Equity Partners (a director-affiliated entity) totaling over $20B in aggregate lease payments.

If AI segment returns do not materialize on the expected timeline, or if Starlink ARPU erosion accelerates faster than subscriber growth compensates, SpaceX's ability to service debt, repay the Bridge Loan, fund Starship, and expand Starlink simultaneously will be strained in a way that its current financial profile does not suggest investors are pricing.

Using data as of 2026-06-03