Quote · All-In with Chamath, Jason, Sacks & Friedberg
Google's AI Brain Drain, SpaceX's Huge Quarter, Airtable's 90% Collapse, US Data Fuels China AI
Where this was said
SpaceX's big quarter: Terafab, AI Capex, $1T revenue projection?
At 28:40 · chapter starts 20:39
SpaceX's first earnings report as a public company is the kind of quarter that rewrites what's possible: $7.8 billion in revenue up 92% year over year, AI compute revenue more than tripling to $2.6 billion, and CapEx of $18.4 billion — a 6x year-over-year surge that has the market asking hard questions about financing [1] — Jason Calacanis "SpaceX's first public earnings were staggering: $7.8B in revenue, up 92% year over year, with AI compute revenue tripling quarter over quar…" 20:56 . Elon Musk simultaneously pulled forward his $1 trillion ARR target from 2031 to 2030 while Morgan Stanley's bullish estimate sits at $325 billion. Brad Gerstner sets the context: SpaceX is down about 40–50% from its IPO peak, following the same historical pattern where tech stocks almost always correct 50% within six months of going public. The market's concern isn't the business — it's the math of financing 6 additional gigawatts of compute next year, which could require $300 billion in CapEx and forces the question of dilutive equity, debt, or Nvidia backstops. David Friedberg makes the Starlink bull case in granular detail [2] — David Friedberg "Starlink generated $4.3B in Q2 revenue with $2.6B in adjusted EBITDA, 12 million subscribers doubling year over year, and an average revenu…" 27:00 : $4.3 billion in Q2 revenue, $2.6 billion in adjusted EBITDA, 12 million subscribers that have doubled year over year at $66 ARPU, growing 20% quarter over quarter. Extrapolate to $40 billion in top-line revenue with perhaps $30 billion in free cash flow, apply a 30x multiple to a high-renewal subscription business, and Starlink alone supports a trillion-dollar market cap — with TerraFab, Groq, Cursor, and Starship as pure upside. David Sacks explains why Starship matters so much to this thesis: each Starship launch deploys 60 V3 satellites adding 60 terabits per second of capacity, versus Falcon 9's 27 V2 satellites adding 2.6 terabits per second — over 20x more capacity per flight. Brad Gerstner closes with the meta-argument: Elon's unique competitive advantage in this AI arms race isn't software, it's his unmatched ability to build physical infrastructure faster than any competitor — a skill forged through Gigafactories and rocket manufacturing that now applies directly to data centers.
SpaceX's first public earnings were staggering: $7.8B in revenue, up 92% year over year, with AI compute revenue tripling quarter over quarter to $2.6B. Elon also pulled forward his $1T ARR target from 2031 to 2030 — a timeline that makes even Morgan Stanley's bullish $325B estimate look conservative.
SpaceX reported $7.8 billion in Q2 revenue, up 92% year over year and 67% quarter over quarter, marking its first earnings report as a public company.
Elon Web Services (SpaceX's AI compute rental division) more than tripled quarter over quarter to $2.6 billion in Q2.
SpaceX's Q2 CapEx hit $18.4 billion, representing a 6x year-over-year increase, implying a roughly $75 billion annualized CapEx run rate.
Elon Musk stated SpaceX's spot compute pricing is in the $30–$50 per watt range, implying $30–50 billion in annual revenue per gigawatt of compute deployed.
Starlink generated $4.3B in Q2 revenue with $2.6B in adjusted EBITDA, 12 million subscribers doubling year over year, and an average revenue per user of $66 per month. At a 30x multiple on projected free cash flow, the Starlink business alone could support a trillion-dollar valuation — and that's before Starship or direct-to-cell.
Starlink generated $2.6 billion in adjusted EBITDA in Q2 from $4.3 billion in revenue, with 12 million subscribers doubling year over year.
Starlink reached 12 million subscribers, having doubled year over year, and is growing at roughly 2 million new subscribers per quarter on the consumer side.
