Subscribe
Share
Share
Embed
Chapters
00:00 Today's headlines
00:45 Musk on SpaceX and Tesla merger
02:23 Telesat offered $3.2 billion rescue package
03:45 Eutelsat capacity deal with Amprion
04:38 SpaceX quarter: space, Starlink and AI
07:29 SpaceXSI AI compute, who pays for buildout
The money, missions, and power shaping space. Space Stakes is a daily briefing on the business of the new space race — launches, contracts, constellations, and the capital behind them. Every episode cuts through the industry's hype: signed contracts vs. paper constellations, funded orders vs. "up to" ceilings, launch dates vs. track records. Who pays, who profits, and who actually delivers — told so both aerospace insiders and curious newcomers finish feeling it was made for them. Space Stakes is an AI-voiced podcast, created and built by a real human, Brian Lampert. Nothing on this show is financial advice.
Also from Brian Lampert: Quickly Quantum and Concrete Compute.
SpaceX just posted about $14.6B in quarterly revenue, Payload reports, with net income of $4.4B — and for the first time its AI business out-earned Starlink. The complication is who actually pays for that AI buildout, and whether subscriber fees or launch money are funding a compute bet that jumps to $35B to $40B in spending next quarter. So who eats that cost if the payback takes longer than promised? That's what today's episode will work out. Welcome back to Space Stakes, your daily brief on the business of space. It's Wednesday, October 21, 2026. Let's get into it.
Elon Musk says merging SpaceX and Tesla would be a good idea — but even he says it is nowhere near done. The news broke on SpaceX's third-quarter earnings call Tuesday, according to Ars Technica, where Musk was asked about formally combining his two largest companies. His answer, quote, “Do I think it would be a good idea? Yes,” with the caveat that shareholders have to be supportive. At shareholders' request, he said both companies have formed special committees of independent directors to discuss the idea, and he said, “I’m a conflicted party; I can’t actually be in the discussions of the special committees.” Those committees have not reached an agreement on terms, and Musk added, “Obviously there’s no guarantee that they will, but I do think there’s a lot of merit to combining the companies as we are pretty much in the superintelligence singularity,” Musk said. The logic he offered was cooperating on superintelligence software, chip design, the Terafab, robotics, mass to orbit with the Starmind constellation, and connectivity. For you, even if you never buy the car or the rocket ride, this matters because it would rank among the largest corporate tie-ups ever, tying your phone's future connectivity and AI to one balance sheet. As @business posted on X, quote, "Elon Musk said SpaceX and Tesla have formed special committees to explore a potential merger, a preliminary step toward what would rank as the largest corporate tie-up of all time" — that ranking is Bloomberg's read on X, not the show's fact. I think committees are cheap and merger terms are expensive, so I am filing this as exploration until a price and a vote appear.
Staying with that merger chatter — Reuters posted on X that PODCAST: SpaceX-Tesla merger, Witkoff in Russia and Kansas voters.
Facing a December wall, Telesat just got a proposal that would keep it out of default — if it takes the deal. SpaceQ Media Inc. reports lenders have offered the Ottawa satellite operator a $3.2 billion rescue package, as reported by The Wall Street Journal Monday, to repay debt that comes due in December. The run of figures here is all from that reporting: the lenders would help Telesat raise $1.7 billion in new equity and borrow $1.5 billion. Existing shareholders could buy the new shares at $31.50, about half Monday's close, and Telesat's stock fell almost 10 per cent on the Toronto Stock Exchange on Tuesday. For newcomers, Telesat runs older geostationary satellites — parked over one spot — while building Lightspeed, its low Earth orbit broadband constellation that circles close to Earth. Telesat GEO, the subsidiary that runs the older fleet, owes about $2.4 billion on a term loan and notes that mature in December. Telesat's second-quarter filings say its cash flows and cash on hand are not expected to cover that debt, raising “substantial doubt” about the subsidiary's ability to pay. The first payment is due Dec. 6, and lenders including Silver Point Capital, GoldenTree Asset Management and Sound Point Capital Management sued over a 2025 transfer of Lightspeed assets — claims that have not been proven in court. Who blinks before December 6 tells you who funds Lightspeed next.
From lenders to lights on the grid — Europe's sovereignty bet just added a new site. Eutelsat has signed an expanded capacity deal with EuroSkyPark, a German specialist in secure communications, to support Amprion, one of Germany's largest electricity transmission system operators, according to Via Satellite. The expansion extends satellite connectivity from Amprion's onshore substations across Germany to its North Sea offshore infrastructure, announced Oct. 20. The setup mixes Eutelsat satellite capacity — both GEO, those parked satellites, and OneWeb LEO — with EuroSkyPark's dedicated networks for primary and backup communications, using Eutelsat 10B's flexible GEO capacity for redundancy independent of terrestrial lines. No contract value or term was disclosed, so I am not guessing at size. For you, this is the unglamorous business that actually pays — keeping power-grid control links alive when ground cables fail.
That's the rest of the rundown. Now, the story that deserves the long look. SpaceX just delivered its first full quarter as a public company, and Payload reports it blew past expectations Tuesday with ~$14.6B in revenue and net income of $4.4B. For anyone new here, SpaceX went public this year, so this was its first full quarter reporting like a public company — rockets and Starlink internet used to fund everything, and now there is a third leg. The revenue number is up 87% from Q2 and 161% YoY—plus substantially higher than the ~$12.8B that analysts had expected. Once again, Payload notes, SpaceX's earnings are as much an AI story as a space one. So is SpaceX now primarily an AI-compute company that launches rockets? My answer is not yet, but this quarter strengthened the AI side fast. Start with space, the smallest piece: space revenue in Q3 jumped 84% compared to the previous quarter to $1.8B despite fewer launches — 35, compared with 38 in Q2 — as more of them carried paying customers. The segment lost just $7M, nearly breaking even after a $542M loss in Q2, despite spending $1.2B on R&D, mostly for Starship. SpaceX flew 12 customer launches, up from 10 in Q2. The company dubbed Flight 14 — SpaceX's first orbital Starship flight — its first revenue-generating Starship mission, delivering 26 Starlink V3 satellites last month. SpaceX will try to catch both the booster and the ship next month on Flight 15. During the earnings call, Elon Musk called success “more likely than not” and said he expects weekly Starship flights in the first half of 2027. Starlink, the commercial internet service beamed from low Earth orbit, is still the profit engine: earning $2.4B in operating income — basically profit from operations — on $5.2B in revenue. That revenue now makes up 36% of the company’s total, down from 55% in Q2, as AI pulled ahead. SpaceX added 1.6M consumer subscribers, bringing the total to 13.6M. That count no longer includes managed enterprise and government accounts, such as airlines and ships. The average subscriber paid $69 a month, up from $66, as more growth came from developed markets where plans cost more. Enterprise and government revenue nearly doubled from a year ago, to $2.1B. And Starlink Mobile now has 22M monthly users, up 37% since Q2 — which to me strengthens our earlier read that SpaceX wants a direct role in mobile service, not just as a supplier to carriers.
Now the turn, because the AI line is where the quarter flips. SpaceXSI brought in $7.6B in Q3, nearly triple its Q2 total and more than half of SpaceX's revenue, Payload reports. Cloud deals accounted for most of the jump, with Cursor and Grok adding the rest. SpaceXSI turned a profit, earning $431M in operating income after losing $1.3B in Q2. It also makes up the bulk of spending, accounting for $16.4B of SpaceX's $19.6B in Q3 capex — capex meaning cash spent on hardware like data centers and ships. Cursor — acquired in August — has already brought in $608M. Executives talked payback fast: Johnsen said new compute pays for itself in less than a year, and Musk said SpaceX could rent out its compute “several times over.” For now, SpaceX is keeping contracts short so it can charge more at renewal. So who actually pays for that buildout? What portion is funded third-party contracts versus internal demand is still open, and these are company figures and projections, not audited conclusions. One more thing from the call: SpaceX still spends more cash than it brings in, and Q4 will cost even more. Management anticipates that the spending will pay off fast, forecasting $33B to $37B in revenue for Q1 2027. Musk took estimates a step further, putting annualized revenue at $500B to $1T by the end of 2027. Those are aspirations wearing a calendar until funded contracts show up — and as for the Tesla merger, Musk said calling it a done deal “would be overstating it.” I believe the revenue mix shift is real, but I am demanding receipts on payback and funded backlog before I call this an AI company that flies rockets. Time for the Hype Check. I land at 7 because $14.6B with $4.4B net income and AI over half of sales is too concrete to dismiss. If SpaceX hits that $33B to $37B Q1 revenue guide while showing funded AI contracts by March, that tells us the compute bet is customer-paid.
If today's AI-over-rockets shift matters to you, follow today's episode and share it with the colleague tracking Starlink or Tesla. This has been Space Stakes, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!
Watching on YouTube? Hit subscribe, and tomorrow's briefing will be waiting for you. I also host Quickly Quantum: a daily quantum computing briefing you don't need a physics degree to follow. The breakthroughs, the funding rounds, and how much substance is really under each claim. Find it wherever you get your podcasts.