The Foundation of SpaceX’s Trillion-Dollar Valuation: Who Is Dividing Up Musk’s Annual $10 Billion Capital Expenditures?

If you missed Apple’s supply chain in 2010, Tesla’s supply chain in 2020—and even NVIDIA’s supply chain over the past two years, which has left you kicking yourself—SpaceX’s supply chain is only just getting started. Of course, I don’t think chasing SpaceX itself outright looks particularly attractive right now. On its IPO debut, its stock surged 19%—jumping from a $135 offering price to $160—with a price-to-sales ratio approaching 100x, while the company remains deeply unprofitable. For retail investors rushing in on Day One, the pressure is far from trivial. So what I’m really highlighting here are the companies supplying SpaceX.

History has repeatedly validated the same logic: super-terminals deliver explosive, downstream benefits to their entire supply chains. In 2010, when Apple launched the iPhone 4, Luxshare’s annual revenue stood at ¥1 billion; ten years later, it soared to ¥92.5 billion, and its share price rose 30-fold. In 2019, as Tesla’s Shanghai Gigafactory went live, CATL’s market cap hovered just above ¥100 billion—five years later, it surpassed ¥1 trillion. And with NVIDIA’s explosive growth over the past two years, Zhongji Xukong’s market cap rocketed from several billion to over ¥100 billion. Apple, Tesla, NVIDIA—each time, a super-terminal stands center stage, but the real wealth creation for a broad cohort of investors happens behind the scenes, within the supporting supply-chain companies.

SpaceX spends hundreds of billions of dollars annually—on chips, materials, components, and industrial gases. These procurement orders are gradually turning into real, tangible revenue on certain companies’ books. Only after its IPO prospectus was made public did this supply chain become quantifiable for the first time. So let’s start by tracing where SpaceX’s money comes from—and where it goes. Its business falls into three main segments:

1. Starlink: Generated $11.3 billion in revenue last year—60% of group revenue—with over 10 million global subscribers. This is SpaceX’s only consistently profitable segment—and arguably, it subsidizes all its other capital-intensive ventures.

2. Rockets: Falcon and Starship R&D consumed $3 billion last year, enabling SpaceX to achieve the lowest commercial launch cost globally. It plans 100 launches in 2026, requiring 1,500 Raptor engines.

3. AI: Lost over $6 billion last year. On the ground, it’s building the Colossus supercomputer—stacked with 220,000 GPUs. In orbit, it’s planning a space-based data center.

So the money flow is straightforward:
Starlink profits → fund rockets to drive down launch costs → low-cost launches deploy AI hardware into orbit → rent out AI compute capacity and collect revenue.
That’s the core loop—and every year, it generates hundreds of billions in procurement orders. Who pockets that cash? We can classify suppliers into three tiers based on replaceability.

Tier 1: Irreplaceable (at least in the short term)
– NVIDIA (NVDA): All 220,000 GPUs in the Colossus supercomputer are NVIDIA chips. But NVIDIA’s true moat isn’t hardware—it’s CUDA, the software ecosystem underpinning virtually all AI training worldwide. As long as SpaceX keeps building supercomputers, NVIDIA keeps getting paid.
– Eutelsat (SATS): Holds critical radio-frequency spectrum for satellite communications—Elon Musk’s direct-to-cellphone feature must route through Eutelsat.
– Filtronic (FTC): Supplies millimeter-wave signal amplifiers for Starlink satellites—and SpaceX accounts for 83% of Filtronic’s revenue.
– Materion (MTRN): The world’s only fully integrated beryllium producer—from mine to finished product—controlling ~56% of global supply.
– STMicroelectronics (STM): Designs and supplies phased-array antenna chips for SpaceX, having delivered over 5 billion units to date.

Tier 2: Technically replaceable—but prohibitively costly to switch
– Honeywell (HON): Provides flight control and inertial navigation systems for rockets.
– Carpenter Technology (CRS): Produces specialty steel alloys used in Raptor engines.
– Hexcel (HXL): Supplies aerospace-grade carbon fiber.
– Broadcom (AVGO): Manages 10-gigabit-per-second data exchange between ground and space.
– Linde Group: Invested $100 million to build an air separation plant near Starbase, Texas—dedicated to producing liquid oxygen and nitrogen.

Tier 3: High-volume, ultra-low-cost manufacturing
Starlink terminals need to be deployed globally at 30 million units—where technical sophistication is secondary to reliability and cost discipline.
– Wistron (6285.TW): World’s largest ODM for Starlink terminals and routers.
– Upstream A-share players include:
– Sunway Communication (300136.SZ): Sole global supplier of high-frequency connectors.
– Parker New Materials (605123.SH): The only Chinese supplier of forgings for Starship’s airframe and engines.
– Western Superconducting (002149.SZ): Supplies niobium alloy for Raptor engines.
– Yingliu Shares (603308.SH): Supplies critical castings for Raptor turbopumps.
– Smaller-tier suppliers:
– Tianyin Electric (300177.SZ): Builds star trackers (stellar inertial sensors).
– Tongyu Communications (002311.SZ): Supplies Starlink ground-station antenna modules.
– U.S.-listed names include Trimble (TRMB), Astronics (ATRO), and CTS (CTSH).

You might ask: “These companies have existed for years—why now?” Three reasons:
1. Procurement volumes are only just beginning to ramp up.
2. Transparency has just opened: With the IPO prospectus public, supply-chain order growth is now trackable and verifiable.
3. Historical analogy fits: Today’s SpaceX supply chain resembles Tesla’s in 2018—mass production just starting, suppliers newly locked in, and order growth entering its steep, early-phase acceleration.

Finally, buying SpaceX on its IPO debut means paying a premium to back Elon Musk’s vision—a very expensive dream of space. But we can shift perspective: Follow the supply chain instead—and bet on something else entirely. Because regardless of SpaceX’s stock price, its $hundreds-of-billions-in-annual procurement orders must be fulfilled. Those orders are decoupled from equity valuations—they’re just monthly, reliable revenue hitting the bank account.

This article does not constitute investment advice. Caveats remain: Beryllium markets are cyclical; Taiwanese manufacturers face geopolitical discounts; smaller firms suffer from illiquidity; and certifications may be reset by rapid tech iteration—requiring case-by-case analysis. But if you miss the IPO allocation, consider a different strategy: skip the top-line hype, and look instead to the quiet, dependable suppliers doing the heavy lifting. The giant has ignited its engines—and this time, the shovels are well within reach.

[nini]