How to Invest in SpaceX: the Full Supply Chain, the Concept-Stock Hype, and the Real Ways to Own It

Research date: June 12, 2026 | OSINT supply-chain research on the public companies exposed to SpaceX and Starlink

Important disclaimer. This is OSINT-based research and educational analysis, not investment advice. I am not a financial advisor. SpaceX is a private company; nothing here is a recommendation to buy or sell any security, fund, or private interest. Market caps and fund weights are point-in-time (mid-June 2026), press-reported where noted, and move daily, especially around the pending IPO. “10X” scenarios are deliberately aggressive thought experiments. Do your own due diligence and consult a licensed advisor.

Companion tools - jump to the interactive dashboard (sortable scores, supply-chain map, rankings, risk controls) or download the Excel scoring model.

Executive summary

A viral list of “SpaceX supplier” tickers (Eson, Lite-On, Meiko, Compeq, Chicony, DISCO, E&R) is what prompted this piece. It is a fun starting point, but most of it is a proxy-trade narrative, not a supply chain. Three facts dominate everything below:

1. SpaceX is private. As of June 2026 it is in the final stretch of a reported IPO (press ticker SPCX, ~$1.5-1.8T target), which has set off a global “SpaceX concept stock” rally. Until it lists, you cannot buy SpaceX directly; you buy exposure.
2. SpaceX is ~80-85% vertically integrated. It builds its Raptor/Merlin engines, tank structures, avionics, Starlink satellites, the Dishy terminals, and even its inter-satellite laser links in-house. That structurally caps how much any external public company can earn from it, and lets SpaceX second-source or in-source almost any part.
3. So the honest map has five buckets, not one list: a short list of confirmed hardware suppliers, a tier of vendors and services that sell consumables to SpaceX (propellant, gases, cloud), a long list of concept stocks whose real driver is AI, a set of downstream customers that buy from SpaceX (telecom, resellers, launch customers), and the equity-proxy vehicles that are the cleanest way to actually own a slice of SpaceX.

The cleanest ways to own SpaceX today are the proxy vehicles - funds and listed companies that already hold a slice of SpaceX. The clearest examples are Baron Partners Fund (SpaceX is ~30% of the fund, and you buy in at that underlying value), Destiny Tech100 (DXYZ, ~16%, but it trades at a steep markup to what it actually holds), Alphabet (a ~6% stake you get as a free bonus inside the stock), and EchoStar (which literally holds ~2% of SpaceX) - plus the IPO itself once it lists. The highest-conviction confirmed supplier is Filtronic (E-band amplifiers, ~70% of revenue from SpaceX). On the new tiers: Linde is the only propellant vendor with a disclosed SpaceX contract, Google Cloud is a confirmed infrastructure vendor (it has hosted Starlink ground stations in Google data centers since 2021; note the headline ~$920M/month GPU deal in the S-1 actually runs the other way - Google paying SpaceX), and downstream the cleanest plays are KVH Industries (a Starlink reseller whose airtime resale is its growth engine) and Planet Labs (cheap Falcon rideshare drives its economics). Almost everything else is a feature, a rumor, or mislabeled.

Explore it yourself: the interactive dashboard

Here is the whole analysis in one place. Use the dashboard to sort or filter all 42 companies by any score, open the supply-chain map, and read through the rankings, watchlist, and risk controls, without leaving this page. If you would rather work in a spreadsheet, you can download the same data as an Excel model. The full write-up, company by company, follows below.

Open the dashboard in a full screen · Download the Excel model

The engine: how SpaceX actually makes and spends money

Before sorting tickers, it helps to understand the machine they are attached to, because the way SpaceX makes money is the reason its supply chain looks the way it does.

It starts with a rocket that comes back. Every time a Falcon 9 booster lands and flies again (one booster has now flown 35 times), SpaceX pays only for fuel, refurbishment, and a fresh upper stage instead of building a new rocket. That drops the internal cost of a launch far below the ~$62M sticker price it charges outside customers, to a figure analysts put somewhere around $15-28M. Cheap launch is the lever for everything else. SpaceX uses its own rockets to deploy Starlink satellites at a cost no rival can match, because a competitor has to buy launch at list price while SpaceX effectively pays itself near cost. Starlink then throws off the cash that funds Starship, a fully reusable vehicle designed to cut the cost per kilogram by another order of magnitude, which in turn makes the next, bigger generation of Starlink satellites economical. Launch feeds the network, the network funds the next rocket, and the next rocket makes the network cheaper still. By 2025 the loop was so dominant that SpaceX alone lifted more than 80% of all the mass humanity put into orbit, across a record 165 launches.

The S-1 that SpaceX filed ahead of its June 2026 IPO showed how lopsided the economics have become. Of $18.7B in 2025 revenue, Starlink was $11.4B (61% of the company) and grew 48%, while the launch business grew just 8%. Starlink alone produced roughly $4.4B of operating profit. The connectivity network is the cash machine; the rockets are increasingly an internal utility that exists to feed it. Yet SpaceX still posted a ~$4.9B net loss on paper, because it is plowing everything back into Starship, satellites, and a large AI bet. The IPO priced at $135 a share for about a $1.75 trillion valuation, roughly 94 times trailing revenue. That one number is the whole debate: bulls see a consumer-subscription business wrapped around a launch monopoly, while skeptics see a company still losing money on paper at an extreme price.

What all that money builds is not a traditional aerospace shop but a set of high-volume factories. In Redmond, Washington, the satellite line stamps out roughly 70 Starlink satellites a week. In Bastrop, Texas, a 700,000-square-foot plant produces user terminals at around 15,000 dishes a day and is being doubled. At Starbase in Texas, SpaceX is building toward the capacity to produce 100 Starships and boosters a year. All of it is capital-hungry: capex jumped from $11.2B in 2024 to $20.7B in 2025. This is the central fact for anyone hunting a supply-chain investment - SpaceX spends like a heavy industrial manufacturer, but it spends most of that money inside its own walls.

The macro backdrop: a $600B economy bending toward $1.8T

Zoom out, and SpaceX sits at the center of one of the bigger structural growth stories in public markets. The global space economy reached a record ~$613B in 2024 and is widely forecast to roughly triple to ~$1.8 trillion by 2035 (the headline figure from a joint World Economic Forum and McKinsey study); the more cautious Morgan Stanley estimate of $1 trillion-plus by 2040 now reads as a floor rather than a target. The fastest-growing slice is connectivity. Satellite broadband is a ~$12-14B market today, heading toward ~$33B by the early 2030s, and the newer direct-to-cell (satellite-to-phone) segment is forecast at anywhere from ~$2.6B to ~$12B by 2030 depending on how you count it.

The reason that opportunity is real is human: about 2.6 billion people, a third of the planet, are still offline, and even in the United States roughly 500,000 square miles have no cellular coverage at all. Those dead zones are the entire pitch for direct-to-cell. Starlink is the runaway leader in attacking them, with more than 12 million subscribers across ~160 markets by mid-2026, around 75% of all active maneuverable satellites in orbit, and (through the T-Mobile partnership) the largest provider of basic phone coverage by land area on Earth.

That dominance is why the competition is a race against a clock. Orbital shells and radio spectrum are finite, and regulators reward whoever fills them first, so deployment cadence is itself the moat. Amazon’s Kuiper is the main Western challenger but is well behind its FCC deadline, with a few hundred satellites against a 3,236 target. The larger strategic story is China: the state-backed Guowang (~13,000 planned satellites) and Qianfan, or Thousand Sails (~14,000-plus), constellations together stake a claim to the same orbits, which US analysts increasingly frame as great-power competition. SpaceX’s ~75% orbital share and 165-launch-a-year cadence are a structural lead that rivals are scrambling to close before the window shuts.

Underneath the consumer story is a demand floor that de-risks the whole thesis: government. SpaceX has won on the order of $22B in cumulative federal awards. In 2026 alone it landed a $2.29B Space Force contract for a military data-relay backbone, a more-than-$4B role in the “Golden Dome” missile-defense initiative, and the largest share (~$5.9B) of the Pentagon’s NSSL Phase 3 launch program. Between Starshield, Golden Dome, and NSSL, there is a multi-tens-of-billions, multi-year stream of government money flowing into launch and space that does not depend on consumer-broadband economics at all.

The single most important macro fact for a supply-chain investor, though, is simpler: launch capacity is the bottleneck for the entire industry. The number of active satellites in low orbit is doubling about every two years, a pace no plausible expansion of rocket production can match. Whoever controls cheap, high-cadence launch controls the supply chain, and that is precisely the position SpaceX has built. It is also the reason the cleanest public ways to play the theme are so indirect: the company that owns the bottleneck is private.

Why SpaceX is unusually hard to invest in

Most “supply chain” theses assume a prime contractor that buys heavily from a tiered base of public suppliers (the way Apple or Nvidia does). SpaceX breaks that model. It is the most vertically integrated hardware company in the space, by design: in-housing is how it cut launch and satellite costs by an order of magnitude. Practical consequences:

• Confidentiality. SpaceX publishes no supplier list, and its May 2026 S-1 names competitors but not component vendors. So almost every “link” is established through teardowns, supplier-side disclosures, or trade press, not SpaceX’s own filings.
• Displacement risk is the master risk. Any supplier that becomes meaningful is a candidate for in-sourcing or dual-sourcing. SpaceX already cut the beamformer-chip count in a Dishy terminal from ~80 (Gen1) to ~6 (Gen3).
• The demand is real and enormous. Starlink revenue was ~$11.4B in 2025 (~61% of SpaceX revenue), tracking toward ~$20B+ in 2026 on more than 12M subscribers and ~15,000 terminals/day; a full Starship stack burns ~4,900 tons of propellant per launch.

So the goal is to separate confirmed from rumored from mislabeled, and to be honest that the best risk-adjusted “SpaceX trade” for most readers is a proxy vehicle, not a Taiwanese small-cap.

Supply-chain dynamics: geography, materials, and chokepoints

If you trace the physical inputs rather than the stock tickers, three things stand out: where the suppliers are, what the hardware is made of, and where the genuine single points of dependence actually sit.

The geography: a Taiwan cluster migrating to Vietnam. The reason so many confirmed Starlink suppliers are Taiwanese is structural. The Dishy terminal is, underneath its exotic radio core, a high-volume consumer-electronics product: an advanced multi-layer circuit board, RF front-end modules, an assembled router, and a die-cast housing. That is exactly what Taiwan’s electronics cluster was built to deliver cheaply and fast, which is why the confirmed names (Compeq for boards, Wistron NeWeb for the terminal, Universal Microwave for RF, Chicony for power) all sit on the island. That concentration is now the risk. Since late 2024, Reuters has reported SpaceX pressing those suppliers to move production off Taiwan, explicitly citing the danger that a China-Taiwan conflict could sever the supply of its single most volume-critical product. The destination is Vietnam, and the shift is partly underway (Wistron NeWeb is already building Starlink gear in Ha Nam province), but final assembly and networking are moving first while the harder advanced-PCB and RF work stays Taiwan-anchored for now.

The materials: stainless steel, kerosene, methane, and a helium scare. Falcon 9 is a conventional aerospace machine: aluminum-lithium tanks, carbon-fiber fairings, burning refined kerosene and liquid oxygen. Starship is the radical departure. Musk abandoned carbon fiber for 301/304L stainless steel because steel costs a few dollars a kilogram instead of hundreds, tolerates far more heat on reentry, and actually grows stronger at cryogenic temperatures. Starship burns liquid methane and oxygen, and a full stack swallows on the order of 4,900 tons of propellant per launch. This is where the industrial-gas vendors enter: Linde and Air Products supply the liquid oxygen, nitrogen, and argon, and Linde built an air-separation plant next to Starbase precisely to feed it. Helium is the wildcard input, since rockets use it to pressurize tanks, and a 2026 supply shock (a large share of global helium briefly went offline) is a real cost-and-schedule risk for kerosene rockets, though notably Starship engineers most of that dependency out by pressurizing its own tanks with vaporized propellant.

The chokepoint: silicon. The terminal’s whole advantage is custom chips that collapsed the part count; SpaceX cut the beamformer chips in a dish from roughly 80 to about 6. But it does not fabricate that silicon itself. The terminal’s RF and main processor chips are made by STMicroelectronics under a decade-long co-design partnership, with ST citing on the order of 5 million chips a day to keep up. That is the closest thing to a genuine single-point supplier dependence in the whole chain, which is exactly why SpaceX has begun hiring to bring even that in-house for its direct-to-cell chips. The satellites, separately, use AMD’s radiation-tolerant Versal chips, which are fabricated by TSMC. For the optical laser links between satellites there is no external supplier at all: SpaceX builds its own, and the one notable merchant player, Mynaric, collapsed into bankruptcy and was absorbed by Rocket Lab.

The thread tying all of this together is the same one that makes the stock-picking so hard: vertical integration plus mass manufacturing is the moat. SpaceX designs the difficult parts in-house and buys only commoditized, second-sourceable inputs, which deliberately keeps every external supplier small, interchangeable, and low-margin. That is excellent for SpaceX’s costs and frustrating for anyone hunting a supplier pure-play.

SpaceX supply-chain map

flowchart TD
    SX["SpaceX (private) -- ~85% vertically integrated -- IPO reported mid-2026"]

    SX --> LAUNCH["Launch / Starship (mostly in-house)"]
    SX --> SAT["Starlink satellites (Redmond WA)"]
    SX --> TERM["User terminals 'Dishy' (Bastrop TX)"]
    SX --> CONS["Vendors / consumables<br/>LIN, APD, AIQUY (gases) - GOOGL (cloud) - AON (insurance)"]

    LAUNCH --> LSTR["Structures / materials: KRMN, CRS, ATI, HWM"]
    SAT --> SRF["Satellite RF / PCB: FTC.L, 2313.TW, 3491.TWO"]
    SAT --> SSIL["Satellite silicon: AMD Versal, STM"]
    TERM --> TODM["Terminal ODM + power: 6285.TW, 6412.TW"]
    TERM --> TSIL["Terminal RF silicon: STM (confirmed), QRVO (merchant)"]

    SX -.->|"customers buy from SpaceX"| DOWN["Downstream<br/>Telecom D2C: TMUS, VEON, KDDI - Resellers: KVHI - Launch: PL, GSAT, ASTS, LUNR"]
    SX -.->|"cleanest exposure"| VEH["Equity-proxy vehicles<br/>BPTRX, DXYZ, GOOGL, SATS (~2% stake), ARKVX"]
    SX -.->|"if it lists"| IPO["SpaceX IPO (SPCX)"]

    CONCEPT["Concept / rumored (AI stocks + Starlink rumor)<br/>2301.TW, 6088.HK, 3665.TW, 6787.T, 300433.SZ, 5243.TW, 8027.TWO, 6146.T"] -.->|"weak link"| SX

The dashed lines are the honest part of the map: for most investors the real exposure runs through the proxy vehicles, the downstream customers, and (soon) the IPO, not the component tickers.

Reading the map, layer by layer. Follow the money from the top. SpaceX’s own capex, $20.7B in 2025, is the source of every flow on the page. Most of it never leaves the company: the engines, structures, satellites, terminals, and laser links it builds itself are the solid lines running down the middle. What does leak out to public suppliers does so in three thin streams. The first is hardware, a short list of confirmed component makers (RF amplifiers from Filtronic, satellite boards from Compeq, the terminal itself from Wistron NeWeb), mostly small and foreign-listed because SpaceX hands out only the commoditized work. The second is consumables and services, the industrial gases that fuel the launches (Linde, Air Products) plus cloud and insurance around the edges, where the suppliers are giants and SpaceX is a rounding error to them. The third stream, on the dashed side, runs the other direction entirely: it is money flowing from customers into SpaceX, as telecoms pay to resell Starlink to phones, maritime and aviation operators resell its bandwidth, and satellite companies buy cheap launches. Off to one side sit the equity-proxy vehicles, the funds and holding companies that simply own SpaceX shares, which for most investors is the only clean way to stand on the same side of the table as the company itself. The visual punchline is the one the dashed lines make: the thick flows are internal, and the public market mostly gets to stand at the edges.

Bucket 1: the confirmed hardware suppliers (short, real, concentrated)

Names with primary-source or strong trade-press evidence of an actual SpaceX/Starlink relationship. Note how small and foreign-listed most are.

Filtronic (FTC.L) is the single highest-conviction confirmed supplier (E-band GaN amplifiers, escalating contracts, a SpaceX warrant, ~70% of its order book) - which is also its biggest risk: near-total customer concentration. Universal Microwave (3491.TWO) is the purest Taiwan name (~80% LEO), Compeq (2313.TW) is the most investable liquid one (confirmed satellite PCBs, >20% LEO, profitable), and STMicroelectronics (STM) is the only confirmed silicon partner with easy US access (but Starlink is immaterial to it). Wistron NeWeb and Chicony are confirmed on the terminal side but are thin-margin, easily multi-sourced roles, and Karman (KRMN) is the best US-listed launch-structures proxy.

Bucket 2: vendors and services - the consumables that flow TO SpaceX

Beyond hardware, SpaceX buys propellant, industrial gases, cloud compute, and insurance. The evidence here is sometimes excellent, but the exposure is immaterial because these are giants.

Linde (LIN) is the best-evidenced vendor of the entire report - a publicly disclosed multi-year liquid-oxygen/nitrogen/argon supply agreement and a $100M air-separation plant it built near Starbase in Brownsville (started up in early 2026); press and analyst estimates (not an official Linde disclosure) put its support at ~70% of SpaceX launches. The catch: that is roughly $100M of a ~$230B company; SpaceX is well under 1% of revenue. Air Products (APD) is similar (a planned 2028 Cocoa, FL launch-gas plant that names SpaceX, not yet operating) and is also the cleanest large-cap helium-shortage play; Air Liquide is a logical regional supplier but its disclosures point to NASA, not SpaceX.

The standout services link is Google Cloud (GOOGL): since 2021 SpaceX has hosted Starlink ground stations inside Google data centers and run Starlink traffic on Google Cloud, a genuine vendor relationship. One important correction: the June 2026 S-1’s headline ~$920M/month (~110,000 NVIDIA GPUs) compute figure is widely misread - it is Google paying SpaceX for xAI/Colossus compute capacity, i.e. SpaceX-as-vendor revenue, not SpaceX buying cloud. Either way it is immaterial to Alphabet, and Alphabet’s real significance here is as a proxy vehicle via its ~6% equity stake. Microsoft (Azure Space) and the insurance brokers (Aon, plus Marsh and Gallagher) have real but tiny links; note Starlink is entirely self-insured.

Several categories simply have no investable public play: liquid methane is trucked to Starbase by private GenOx (the public LNG names like Cheniere have no disclosed contract); helium pure-plays are tiny speculative juniors (Avanti, Pulsar) that do not supply SpaceX; Starbase construction is self-performed with private subcontractors; and marine recovery and transport are done in-house (Falcon Landing LLC). Treat this whole bucket as evidence-rich but exposure-poor.

Bucket 3: the concept stocks (mostly AI names wearing a Starlink rumor)

Real, often excellent companies whose 2026 growth is driven by AI infrastructure, with a SpaceX link that ranges from rumored to mislabeled. Buying them “for Starlink” is usually the wrong thesis.

Three corrections the viral list gets wrong: Eson Precision is not a SpaceX launch supplier (its tie is supplying Foxconn’s satellites, which flew on Falcon 9); E&R / “$LPK” are AI advanced-packaging equipment with no verified SpaceX link; and the screenshot mislabels some lines (the “$7.48B connectors” line is FIT Hon Teng, not BizLink). And there is no public pure-play for Starlink’s optical inter-satellite links - Mynaric was wiped to zero and absorbed by Rocket Lab; SpaceX builds its own laser terminals.

Bucket 4: the downstream - the public companies that buy FROM SpaceX

The other side of the chain: telecoms that resell Starlink Direct to Cell, verticals that adopt Starlink, and satellite operators that pay SpaceX to launch.

Direct-to-cell telecom partners. These resell Starlink’s satellite-to-phone service to fill cellular dead zones. Be clear-eyed: for almost all of them it is a coverage/marketing feature, not a revenue driver, and Starlink is simultaneously a long-term threat that can disintermediate their own networks.

T-Mobile (TMUS) is the flagship and best-marketed (with a unique cross-carrier angle that poaches AT&T/Verizon subscribers), but satellite is immaterial to a $200B incumbent. VEON/Kyivstar (VEON) has the most material adoption (3.0M direct-to-cell users, Europe-first) because the service is genuine network resilience in wartime Ukraine - but geopolitical risk dwarfs the upside. KDDI is the technical leader. Telstra illustrates the disintermediation risk perfectly: it resells the very Starlink rural broadband that commoditizes its own regional network. (Others - Singtel/Optus, Telefonica’s Virgin Media O2, One NZ via Infratil, Salt, Entel - are delayed, diluted in holding companies, or not accessible to US investors.)

Starlink verticals and resellers. Adopters that put Starlink on planes, ships, and enterprise sites.

KVH Industries (KVHI) is the cleanest downstream play in the whole report - an authorized Starlink maritime reseller that re-architected its entire business around it (2026 guidance: revenue +24%, EBITDA +66%, on a $45M Starlink purchase). Starlink airtime resale is literally its growth engine. The catch is structural: a thin-margin middleman exposed to SpaceX disintermediation, and a micro-cap. For airlines (United) Starlink is a free-wifi cost; for cruise lines (Royal Caribbean, and Carnival/Norwegian) it is marginally better because they monetize wifi as a high-margin package - but in every case it is tiny relative to revenue.

Launch customers. Satellite operators that pay SpaceX to launch, and benefit from cheap, reliable, frequent access.

Planet Labs (PL) is the cleanest launch-enabler story: cheap SpaceX Transporter rideshare structurally improves its smallsat economics, it is posting +42% growth, and (unlike the others) it is not a Starlink competitor. Globalstar (GSAT) has active 2026 Falcon 9 launches serving its Apple Emergency-SOS contract (though Amazon agreed to acquire it for ~$11.6B in April 2026, so the standalone thesis is being overtaken by a buyout); AST SpaceMobile (ASTS) relies on Falcon 9 to deploy - but is also a direct-to-cell competitor; Intuitive Machines (LUNR) is a high-risk lunar play; and Iridium is a former anchor customer that now competes. The special case is EchoStar (SATS), which swapped spectrum for a ~2% direct stake in SpaceX (worth ~$11B) plus ~$8.5B of cash, in a deal valued at ~$19.6B total - making it, despite a declining legacy business, one of the most direct equity-linked SpaceX wrappers on the market.

Bucket 5: the cleanest way to actually own SpaceX (equity-proxy vehicles)

Because the operating exposure is thin, the most direct way to own SpaceX is funds and holding companies that own its equity.

Baron Partners Fund (BPTRX) is the best risk-adjusted proxy (SpaceX ~30% of NAV, priced at NAV, daily liquid - bundled with Tesla). Destiny Tech100 (DXYZ) is the most concentrated listed proxy (~16%) but trades at a brutal premium to NAV. Alphabet (GOOGL) owns ~6% as a free embedded option, and EchoStar (SATS) holds ~2% directly. When the SPCX IPO lists, it becomes the only true pure play - and that event would likely collapse the proxy premiums (a key risk for DXYZ/VCX holders).

Not a clean SpaceX play: competitors and disrupted incumbents (context)

Scoring model: 42 names, 1-10 across 11 dimensions

Each name is scored 1-10 where 10 is most attractive (for CONC and GEO, 10 = lowest risk). Scores are judgment applied to the OSINT evidence, not a formula. The Composite rewards quality and confirmed exposure, so it skews toward proxy vehicles, quality vendors, and confirmed suppliers; read the X10 column on its own for genuine multibagger optionality.

Legend: SXP = SpaceX exp - EVID = Evidence - GRW = Growth - MOAT = Moat - CONC = Divers - VAL = Value - BS = Balance - ACC = US access - GEO = Geo risk - CAT = Catalyst - X10 = 10X.

The pattern is the honest punchline. The highest composites mix proxy vehicles (Alphabet, Baron) with great companies where SpaceX is immaterial (Linde, Microsoft, T-Mobile) - safe ways to touch the theme that cannot 10x. The highest X10 scores sit with the small, levered names: Filtronic and KVH (confirmed and concentrated), AST SpaceMobile and Intuitive Machines (launch enablers with binary risk), and the small confirmed suppliers. The concept stocks (Eson, E&R, Lens) score lowest because the SpaceX evidence is weakest.

Best ways to play SpaceX, by goal

1. Most direct exposure, lowest hassle: Baron Partners (BPTRX) - ~30% SpaceX at NAV. Then Destiny (DXYZ) only on premium compression.

2. Own a literal slice of SpaceX equity: Alphabet (GOOGL, ~6%, a free option) and EchoStar (SATS, ~2% stake plus a multi-billion cash check).

3. Highest-conviction confirmed supplier: Filtronic (FTC.L) - contract-backed, ~70% SpaceX, with a warrant. Size small.

4. Best downstream P&L linkage: KVH Industries (KVHI) - the rare name where Starlink adoption directly drives revenue (+24% guided).

5. Best launch-enabler: Planet Labs (PL) - cheap rideshare improves its economics, with real accelerating growth; Globalstar (GSAT) as the runner-up.

6. Best-evidenced vendor (immaterial, but real): Linde (LIN) for propellant; Google Cloud (GOOGL) for compute.

7. Easiest US-listed supplier: STMicroelectronics (STM) terminals; Karman (KRMN) launch structures.

8. Highest-risk / highest-upside (size tiny): Filtronic, Universal Microwave (3491.TWO), KVH (KVHI), AST SpaceMobile (ASTS), Intuitive Machines (LUNR) - each a concentration, reseller, or binary bet.

9. The pure play that does not exist yet: the SPCX IPO. Watch it - and watch the proxy premiums (DXYZ/VCX) collapse toward NAV when it lists.

Watchlist (bull / base / bear)

Filtronic (FTC.L) - E-band amplifiers (confirmed supplier)

• Bull: Starlink gateway/payload buildout keeps escalating; SpaceX warrant aligns incentives; content scales with V3 satellites.
• Base: Order book grows with Starlink but slower; volatile around each contract headline.
• Bear: SpaceX in-sources or dual-sources SSPAs; with ~70% concentration that is existential. -50%+.
• Monitor: new SpaceX contracts, concentration disclosure, GaN capacity, SpaceX in-house RF hiring.

KVH Industries (KVHI) - Starlink maritime reseller (downstream)

• Bull: Maritime Starlink adoption compounds; KVH’s airtime resale + pooled-data model drives +24% revenue / +66% EBITDA; re-rates off a micro-cap base.
• Base: Growth continues but reseller margins stay thin.
• Bear: SpaceX goes direct to maritime and disintermediates the channel; margin compression. -40%+.
• Monitor: Starlink airtime gross margin, subscriber adds, any SpaceX direct-maritime move, the next purchase-commitment size.

EchoStar (SATS) - equity-linked (~2% SpaceX stake)

• Bull: SpaceX IPO marks the ~2% stake (~$11B) up sharply; the ~$8.5B cash slug (part of the ~$19.6B deal) transforms the balance sheet.
• Base: Legacy DISH/Hughes decline offsets the spectrum windfall; value hinges on capital allocation.
• Bear: Core businesses erode faster than the stake appreciates; deal/timing risk.
• Monitor: SpaceX IPO/valuation, deal close + cash receipt, what management does with the proceeds.

Baron Partners Fund (BPTRX) - proxy vehicle

• Bull: SpaceX IPO marks the ~30% position up; captured at NAV, daily liquid.
• Base: SpaceX compounds privately; tracks it plus Tesla.
• Bear: Tesla drawdown or a soft IPO drags NAV; not a pure play.
• Monitor: quarterly SpaceX weight/mark, IPO timing, Tesla correlation.

Risk controls

1. Vertical-integration / in-sourcing (the master risk). Any supplier or reseller that matters can be displaced. Favor specialty roles SpaceX chooses to outsource (RF like Filtronic/UMT, gases like Linde) over commodity roles it can pull in-house. Control: treat single-customer names as high-beta options.
2. Concept-stock / proxy-trade hype. The IPO has inflated a basket of loosely-related tickers - many are AI stocks with a Starlink rumor stapled on. Control: demand evidence (teardown, contract, supplier disclosure) before paying a “SpaceX multiple.”
3. Downstream disintermediation. Resellers (KVH) and telecom partners (Telstra) resell the very Starlink product that can later bypass them. Control: prefer enablers SpaceX cannot easily replace (launch customers like Planet) over pure resellers.
4. IPO premium collapse. DXYZ and VCX trade at large premiums to NAV that an actual SPCX listing would compress. Control: prefer at-NAV vehicles (Baron) or buy listed CEFs only on premium compression.
5. Customer / single-thesis concentration. Filtronic (~70% SpaceX), UMT (~80% LEO), KVH (Starlink resale), Globalstar (Apple). Control: size small; diversify across the basket.
6. Foreign-listing / access risk. Most confirmed suppliers are TWSE, TPEx, HKEX, AIM, or Tokyo. Control: confirm the tradable line and liquidity (ADR/OTC) before sizing.
7. Geopolitics / supply-chain security. Starlink is US/defense-adjacent; China names (Lens) face export pressure; VEON carries war risk; SpaceX is migrating suppliers to Vietnam. Control: discount China-listed links and stress-test the geopolitically exposed names.
8. Valuation and immateriality. Much of the basket re-rated on the IPO, and for the giants (Linde, Microsoft, T-Mobile, Alphabet) SpaceX is a rounding error. Control: separate “real SpaceX exposure” from “story premium,” and own the giants on their own merits, not as SpaceX bets.

Methodology, sourcing, and data-quality flags

This report synthesizes live web research (June 2026) across nine parallel streams (user terminals, satellites, launch/Starship, semiconductors/RF/optics, US beneficiaries and equity vehicles, propellant/gases/energy, construction/logistics/insurance/IT, downstream telecom, and downstream verticals/launch customers), prioritizing supplier-side disclosures, product teardowns, SEC/regulatory filings, the SpaceX S-1, fund holdings (DXYZ 424B5, Baron letters, ARK CSV), and credible trade press (DigiTimes, Reuters, company press releases). SpaceX publishes no supplier list, so links are graded confirmed / rumored / mislabeled rather than asserted.

Data-quality flags:

• Caps are point-in-time and FX-converted (NT$/USD ~32.5, HK$/USD ~7.8, EUR/JPY/AUD as noted); foreign small-caps move fast around the IPO.
• Screenshot corrections: the “$7.48B connectors” line is FIT Hon Teng, not BizLink; Eson is a Foxconn-satellite supplier, not a SpaceX launch supplier; E&R / “$LPK” are AI-packaging equipment with no verified SpaceX link. (The screenshot’s ~$10.4B Compeq cap is roughly correct - an earlier draft wrongly cut it to ~$3.6B.)
• Mynaric is delisted (equity wiped to zero; absorbed by Rocket Lab, April 2026) - there is no public pure-play optical-ISL Starlink supplier.
• Vendor evidence vs materiality: Linde and Google Cloud have disclosed SpaceX relationships but SpaceX is well under 1-3% of either company - own them as quality businesses, not SpaceX bets.
• Downstream framing: direct-to-cell is a coverage feature for telecoms (immaterial revenue, plus disintermediation risk); KVH is the rare reseller where Starlink directly drives the P&L; launch customers (Planet, Globalstar, ASTS) are enablers carrying their own standalone risk, and several are also Starlink competitors.
• SpaceX valuation/IPO figures are press-reported (~$1.5-1.8T target) and change weekly. Verify against live filings before acting.
• Scores are directional judgment, not a mechanical formula; the Composite rewards confirmed exposure and quality, so it favors the proxy vehicles and quality vendors over the highest-torque small caps.

Prepared June 12, 2026. Figures are point-in-time and will change, especially around the pending SpaceX IPO. This is research and analysis for educational purposes - not investment advice, not a recommendation, and not a solicitation. SpaceX is private and not directly investable until/unless it lists. Markets involve risk of loss; “10X” scenarios are low-probability thought experiments. Verify all figures independently and consult a licensed financial advisor before making any decision.