Briefing: SpaceX IPO Valuation: Am I Missing Something?

Published: June 7, 2026 | Language: English | Author: E. J. Original article: https://posts.ejsays.com/spacex-ipo-valuation-am-i-missing-something/ Chinese version: https://www.ateasthillside.com/p/59-spacex-ipo

Disclaimer (author's note): All analysis is speculation only. Written for fun and intellectual curiosity on a Sunday afternoon with a calculator. Not financial advice. Not an accusation of anyone, any organization, or anything. Just a middle-aged woman trying to figure out what she is missing.

Core claim: Morgan Stanley's SpaceX roadshow model projects $3.4 trillion revenue by 2040 — 181x growth from 2025 in 15 years, requiring 41.3% compound annual growth rate. No company in modern history has sustained 41% CAGR at this scale for 15 consecutive years. The author's question: what is she missing?

The benchmark problem:

None of them came close to 41% over 15 years. Not one.

The three businesses — each examined:

AI (Colossus data center): SpaceX acquired xAI in February 2026, inheriting 220,000 Nvidia GPUs and 300MW of power in Memphis. Current tenants: Anthropic ($1.25B/month through 2029), Google ($920M/month through June 2029). Combined: ~$25B/year. Both contracts allow 90-day exit after December 2026. The GPUs are being rented to competitors actively building models that compete with Grok. xAI's founding team has largely departed. Grok has not topped any benchmark in months. Its most culturally notable moment: generating undressing images on X. Nearly 100% of compute capacity is now generating rental income. This is a data center landlord business with short-term leases and depreciating assets. GPU rental is not a 41% CAGR business.

Connectivity (Starlink): Real, impressive, $11.4B revenue in 2025. But: Amazon Kuiper just won Delta Air Lines — aviation is Starlink's highest-ARPU segment (~$300K/aircraft/year). Delta is one of the world's largest airlines. OneWeb is expanding. China's constellation is coming. Japan is testing HAPS (stratospheric airships at 20km altitude, 0.3ms latency, direct smartphone connection, no dish required) — a different cost structure that competes at the edges where Starlink struggles. Every serious government with sovereignty interests has reason to avoid permanent dependence on infrastructure controlled by a single American citizen who publicly feuds with heads of state and retains contractual rights to cut access if he deems it harmful to humanity. No country's connectivity spending has ever grown at 41% annually. With these factors, the range narrows further.

Rockets and space: 620+ Falcon 9 flights, 99% success rate, 80%+ of global mass to orbit since 2023. Genuinely one of the most impressive engineering achievements in human history. But: defense contracts are political, cyclical, and contested. Blue Origin is operational. Governments dislike single-source dependency. Mars: first uncrewed mission projected late 2026, human landing earliest 2029 (more likely 2031). Revenue from Mars does not appear in any plausible model for at least a decade. The 41% CAGR must start now, from today's businesses.

The underwriter structure:

Morgan Stanley is: lead underwriter, stabilization agent, manager of SpaceX employee direct share plan, owner of Shareworks (where SpaceX employees hold stock and options), and owner of E*TRADE (where retail investors would buy). Every door leads to Morgan Stanley.

The author's honest position: If she is wrong — which is entirely possible, even likely, she tells herself — $2,000 is not much to bet on the greatest growth story since the internet. But $2,000 is half a month's rent in San Francisco. She would rather help a beached whale in Monterey Bay. Bad news: she just checked. It's gone.

SpaceX Revenue Projection Math

Starlink Competitive Pressure