LEO satellite.
Whoever owns the rocket.

Amazon has the money, the satellites, and a deal to power Apple's iPhone satellite features. What it doesn't have is enough rides. In September it was buying more launches on a European rocket held back by a turbopump problem, and the Blue Origin rocket it was also counting on blew up on the pad in May.

That's the odd thing about low Earth orbit: the satellites are the easy part. LEO is the band a few hundred to about two thousand kilometers up, close enough for satellites to deliver fast internet and talk to ordinary phones. The trade-off is that each one only covers a small patch of the planet at a time, so a useful network needs thousands of them, replaced every few years. Building satellites has gotten cheap. Getting thousands of them into orbit is still the hard part.

Nobody doubts the business anymore. Starlink had 12 million subscriber lines at the end of June, SpaceX is now public, and the rest of the industry is consolidating to keep up, with Rocket Lab buying Iridium and Amazon buying Globalstar. The market for LEO satellite stocks has stopped asking whether LEO works.

The 2027 question for LEO satellite stocks is who can actually get a network into orbit on schedule, and the list is sorted by the answer:

• Own the rocket: they launch their own networks.
• Wait in line: network operators that depend on someone else's rocket.
• Satellite makers: paid to build, whoever launches.
• Earth data: imaging companies selling answers, not bandwidth.

The companies that can put their own satellites up when they want, at their own cost.

SpaceX NASDAQ: SPCX

SpaceX is the network that launches itself. Starlink had 12 million subscriber lines at the end of June, connectivity revenue grew 66% in Q2, and in September Starship carried its first batch of next-generation V3 satellites to orbit. Each V3 is designed for about a terabit per second of downlink, and Starship is meant to carry up to 60 at a time.

Nobody else can add capacity that fast or that cheaply, because nobody else owns the rocket and the network together. The trade-off is that SpaceX isn't a clean Starlink stock: it also comes with the launch business and an AI business.

Rocket Lab NASDAQ: RKLB

Rocket Lab is building the SpaceX model at a smaller scale: rockets, satellites and now a network. Iridium shareholders approved its takeover, worth roughly $8 billion including debt, in September, which brings a global satellite phone network, valuable spectrum and steady service revenue. The deal still needs FCC and other approvals, with a close targeted for mid-2027.

Q2 revenue was a record $234 million, up 62%, with a $2.36 billion backlog. Its bigger Neutron rocket hasn't flown yet. Until it does, Rocket Lab owns a small rocket and is about to own a big network.

These operators have the satellites, spectrum and customers, but they have to book someone else's rocket to get into orbit. In 2027, the launch schedule is their schedule, and it's what will move these LEO satellite stocks.

Amazon NASDAQ: AMZN

Amazon's satellite network, Amazon Leo, has the deepest pockets in LEO, and it's lining up customers. It's buying Globalstar, Apple's current satellite partner, in a deal expected to close in 2027, and has signed on to power the satellite features on iPhones and Apple Watches. In September it bought six more Ariane 6 launches, bringing that contract to 24.

But Ariane 6 is slowed by a turbopump problem, and Blue Origin's New Glenn is still recovering from a pad explosion in May. Amazon has everything a satellite network needs except the one thing SpaceX makes itself. You don't buy Amazon for Leo, but Leo is where its money meets the launch bottleneck.

AST SpaceMobile NASDAQ: ASTS

AST SpaceMobile wants your normal phone to connect straight to a satellite, with no dish. It has agreements with more than 60 mobile carriers covering over 3 billion subscribers, about $1.3 billion of contracted revenue, and more than $3.7 billion of pro forma cash after selling convertible notes in July. Three more BlueBird satellites launched in August.

Each BlueBird is huge, and the service only works at scale once dozens are up. The cash is there. What the stock needs now is a launch cadence that matches the factory.

Eutelsat Euronext Paris/LSE: ETL

Eutelsat owns OneWeb, the only other global LEO broadband network flying today, and Europe wants it as its own alternative to Starlink. On September 10 it gave Airbus the go-ahead on 229 more OneWeb satellites, an estimated €1 billion investment, and its CEO expects about 1,300 in orbit by 2030. LEO revenue grew nearly 70% last fiscal year to a quarter of the company. Its growth now depends on getting those satellites launched, on SpaceX and Europe's still-thin Ariane 6 schedule, which is the bottleneck this list is about.

Telesat NASDAQ/TSX: TSAT

Telesat is an old satellite operator rebuilding itself around a new LEO network called Lightspeed. In August it signed a C$2.3 billion Canadian military contract, expanded Lightspeed to 225 satellites and pushed its LEO backlog to about C$5.6 billion pro forma, with global service targeted for 2028. The legacy business is shrinking fast, so Lightspeed is now the whole story. The stock has more than tripled in three years on it.

MDA Space NYSE/TSX: MDA

MDA Space builds satellites for other people's networks, so it gets paid whoever wins. It's building Telesat's Lightspeed constellation, picked up another C$474 million order when Lightspeed expanded, and is finishing Globalstar's 17 replacement satellites, the first eight of which launched in August. Q2 revenue was up 34% to about C$499 million, on a C$4 billion backlog. Big programs make the cash flow lumpy, but every new constellation needs a factory.

Broadband gets the headlines. Imaging companies sell answers to governments that want to know what changed on Earth today.

Planet Labs NYSE: PL

Planet runs about 200 imaging satellites and photographs the planet's landmass every day. It's moving upmarket into sharper satellites and defense and sovereign contracts, and it's working: fiscal Q2 revenue hit a record $116.1 million, up 58%, with about $865 million in cash and investments. The satellites matter less than the archive and the software that turns pictures into answers.

BlackSky NYSE: BKSY

BlackSky is smaller and more defense-focused than Planet, selling very sharp images on fast revisit times, with analytics built in. Q2 revenue was $33.3 million, up 50%, including a record $24.5 million from space-based intelligence and AI services. Defense buying is lumpy, but governments increasingly want commercial imagery now instead of waiting years for their own satellites.

SpaceX is assembling everything it needs to become a phone company.

Today satellite-to-phone service is a partnership. Carriers rent their spectrum to SpaceX or AST and sell the service under their own brand. But in 2025 SpaceX agreed to buy a block of mobile spectrum from EchoStar for about $17 billion, later expanded to nearly $20 billion. Regulators signed off in May, and the deal is set to fully close by late 2027. Then SpaceX will own the rocket, the satellites, the bandwidth and the airwaves.

Once V3 satellites are flying by the hundred, SpaceX won't need a carrier to reach your phone. It could sell coverage directly, starting where cell towers are thin and moving in from there. The carriers signing with AST are partly buying insurance against that.

In 2027, watch whether SpaceX starts selling phone service under its own name. If it does, LEO satellite stocks stop trading like a satellite industry and start trading like a telecom fight, against an opponent that launches its own towers.