China is copying SpaceX's playbook — and finding out the playbook isn't the point

Space 4 min 80 sources

SpaceX's record $75 billion IPO has set off a wave of Chinese space listings chasing the same goals. But the numbers, and the structure underneath them, show how far the copy still is from the original.

Key takeaways

SpaceX's record $75 billion IPO has set off a wave of Chinese space startups racing to copy its model — reusable rockets and satellite mega-constellations.
The numbers show the distance: SpaceX made nearly $19 billion in 2025 with ~10,400 Starlink satellites; LandSpace made $5.2 million, and no Chinese firm has yet landed a reusable booster.
The hardest thing to copy isn't the rocket — it's SpaceX's structure, where its own internet business creates demand for its own launches. China's fragmented, state-led sector has no equivalent.

SpaceX went public this week at a record valuation, and the shockwaves reached Beijing fast. Its $75 billion IPO — the largest in history — made Elon Musk the world’s first trillionaire on paper [13]. In China, at least seven rocket and satellite startups are now racing toward their own public listings, betting that what worked for SpaceX will work for them [13]. The frenzy is real. So is the gap between the imitation and the original.

The race to copy

China’s space firms are chasing the two things that made SpaceX’s economics work: reusable rockets that land and fly again, and giant satellite constellations that beam internet down to phones and ships [13]. One early investor in LandSpace — China’s closest private answer to SpaceX — says his 2016 stake is now worth roughly 100 times what he paid [13]. Soochow Securities forecasts China’s commercial space market could pass $1 trillion by 2030 [13].

But the firms are going public ahead of meaningful revenue and without the proven technology that underpins SpaceX’s numbers, a mismatch analysts say will cap how much they’re worth [13]. The scale of the gap is stark. LandSpace reported 36.4 million yuan — about $5.2 million — in revenue for the first half of 2025 [13]. SpaceX’s 2025 revenue rose by a third to nearly $19 billion, around three-fifths of it from its Starlink internet service [13].

The rocket that hasn’t landed

The harder gap is technical. No Chinese firm has successfully launched and recovered a reusable rocket [13]. LandSpace flew the first test of its Zhuque-3 booster in December, but it failed to complete a controlled landing and was not recovered [13]. The ability to return, refurbish and re-fly a booster — the single thing that slashes the cost of reaching orbit, because the rocket is most of the price — remains unproven across the whole Chinese sector [13]. One Chinese space executive said the most optimistic case has China matching Starlink’s current scale around 2033 [13]. Starlink already has roughly 10,400 satellites in orbit; China’s two main constellation projects, Guowang and Qianfan, have a combined few hundred [13].

The part that doesn’t copy

The deepest difference isn’t a rocket or a balance sheet. It’s the shape of the company. SpaceX runs a vertical model: Starlink needs launches, and SpaceX provides them, so its own broadband business generates demand for its own rockets [13]. “The big move of SpaceX was to move revenue generation away from launch and to broadband constellations,” said Gabriel Deville of consultancy Novaspace [13].

China has no equivalent. Its sector is fragmented — startups depend on orders from state-backed constellation operators whose schedules sit outside their control [13]. You can see the fragmentation in this week’s other China news: the government shortlisted four commercial firms — Galactic Energy, CAS Space, OrienSpace and LandSpace — to launch a single new cargo spacecraft to its Tiangong space station [64]. Four companies competing for one job is the opposite of one company feeding itself. As one consultant put it, a private Starlink may never emerge in China at all: “If you want to be a telco in China, there are no private telcos in China” [13].

Elsewhere this week

Japan’s H3 rocket returned to flight after a December failure that lost a navigation satellite, launching a new three-engine variant carrying six small satellites [48]. NASA named the crew for its Artemis III mission — though that flight will test landers in Earth orbit, not land on the Moon [5]. And Europe’s space sector heard a warning of its own: a panel noted that strong stated demand for European launchers has yet to translate into actual orders, while a separate piece argued Europe is still asking only how to get satellites up, not how to keep the ones already up there working in an increasingly crowded orbit [73][38].

Far from the launch pads

Two quieter findings landed this week. The James Webb Space Telescope weighed a “sleeping giant” — a black hole 6 billion times the mass of our Sun, seen as it was 10 billion years ago — by measuring how its gravity bends the light of stars around it [63]. And astronomers reported a flood of newly found small, dark “irregular moons” around the outer planets, hinting the outer solar system saw violent upheaval more recently than thought [31]. Neither is a launch or a listing. Both are reminders that the universe keeps its own books, indifferent to anyone’s IPO.

02 · Lesson · why it matters

You can copy what a winner does without copying why it works

The visible part of any success — the moves, the milestones, the structure on paper — is the easy part to imitate. The part that actually made it work is usually invisible, and it doesn't travel.

The move everyone can see

When SpaceX went public this week at $75 billion, China’s space industry didn’t just take notes. It started copying. Seven Chinese startups are now racing toward their own listings, chasing the two SpaceX trademarks: reusable rockets and giant satellite constellations.

This is a natural thing to do. When something works spectacularly, you look at what it did and you do that. SpaceX has a reusable rocket, so build a reusable rocket. SpaceX has Starlink, so build a constellation. SpaceX went public, so go public. The playbook is right there, visible to anyone watching.

The trouble is that the visible playbook is almost never the reason the thing worked.

What’s easy to copy, and what isn’t

Look at what China has actually reproduced so far. The IPOs: real. The ambition: real. An early LandSpace investor’s stake is up roughly 100 times. By those measures the copy is going well.

Now look at what hasn’t reproduced. No Chinese firm has landed and re-flown a rocket — LandSpace’s booster failed its recovery test in December. LandSpace’s revenue last year was $5.2 million; SpaceX’s was nearly $19 billion. Starlink has about 10,400 satellites in orbit; China’s two big projects have a combined few hundred.

Notice which things copied cleanly and which didn’t. The announcements copied — listings, targets, press releases. The results didn’t. That’s the tell. When a strategy is easy to imitate but the outcome isn’t, you’ve copied the surface and missed the engine.

The engine was never the rocket

Here’s the part that’s genuinely hard to see. The thing that made SpaceX’s economics work isn’t any single rocket or satellite. It’s the shape of the company.

SpaceX needs launches for its own Starlink network, and SpaceX provides those launches. Its internet business creates demand for its rocket business, which it owns. The money loops inside one company. “The big move of SpaceX was to move revenue generation away from launch and to broadband constellations,” one analyst put it. The rocket is famous. The loop is what pays for it.

That loop is invisible from the outside. You can photograph a landing booster; you can’t photograph a demand structure. So when China copied SpaceX, it copied the booster and the IPO — the things you can see — and the loop quietly didn’t come along.

Why the engine didn’t travel

And it couldn’t have, because the engine depends on conditions China doesn’t share. SpaceX’s loop works because one private company owns both the demand and the supply. China’s space sector is fragmented: dozens of startups depend on orders from state-backed operators whose schedules they don’t control.

You can see it in this week’s smaller news. China shortlisted four commercial firms to launch a single new cargo spacecraft. Four companies fighting over one job is the exact opposite of one company feeding itself. “If you want to be a telco in China, there are no private telcos in China,” a consultant noted — so the one piece that made the original work, a private firm generating its own demand, may simply never appear there.

The copy isn’t failing because the engineers are worse. It’s failing because the thing worth copying was a structure, and structures grow out of conditions — ownership rules, who’s allowed to sell internet, how the state procures. Move the move to new soil and the soil rejects it.

What this is really about

This isn’t a story about China, or rockets. It’s about a mistake everyone makes, including the reader. We watch someone succeed, we extract the visible playbook, and we run it — the founder’s morning routine, the company’s reorg, the friend’s career move, the diet that worked for someone built differently. We copy what we can see because that’s all we can see. And then we’re puzzled when the same moves give us different results.

The honest position is that you usually cannot tell, from the outside, which part of a success was load-bearing. The visible part — the part that got written up, the part you can imitate — is rarely it. The part that mattered was some quiet structure, some condition, some loop that never made the photo. You are copying too, right now, somewhere in your life. The useful thing isn’t to stop. It’s to hold the copy loosely — to assume the reason it worked for them is something you can’t see, and may not have, and to watch for that gap before you bet everything on the playbook.

03 · Lab · your turn

Build Your Copy

Copy SpaceX's visible playbook and watch the result stall, then learn the load-bearing piece is an invisible structure your conditions won't let you copy.