In short: Amazon’s satellite internet service, rebranded from Project Kuiper to Amazon Leo in November 2025, entered enterprise beta on April 8, 2026, with commercial availability targeted for mid-2026 per Andy Jassy’s annual shareholder letter. The service offers three terminal tiers delivering up to 1 Gbps for enterprise users, with Verizon, AT&T, Vodafone, JetBlue, and NASA among the beta partners. Amazon has approximately 210 to 241 satellites in orbit against a Federal Communications Commission requirement of 1,618 by July 30, 2026, has applied for a two-year deadline extension, and has contracted 22 additional launches to close the gap.

From Project Kuiper to Amazon Leo, the rebrand and the beta

Amazon received Federal Communications Commission approval for a 3,236-satellite low-earth-orbit constellation in 2020, then spent five years building the hardware, regulatory infrastructure, and carrier partnerships needed to turn that approval into a commercially viable service. The first production satellites launched in April 2025 aboard an Atlas V rocket operated by United Launch Alliance, and by November 2025 Amazon had enough operational hardware in orbit to retire the Project Kuiper name in favour of Amazon Leo, a rebrand that signals a deliberate shift from development programme to commercial product. A business preview programme opened to select enterprise partners shortly after the rebrand. The full enterprise beta launched on April 8, 2026. The following day, Jassy’s annual letter to shareholders confirmed mid-2026 as the commercial launch window, placing Leo alongside Amazon’s $50 billion Trainium chip investment as one of the defining bets in the company’s current capital allocation cycle. Beta customers span Verizon and AT&T in North America, Vodafone and Vodacom across Europe and Africa, JetBlue for in-flight connectivity, NBN Co in Australia, Vrio in Latin America, and NASA, along with enterprise logistics clients Hunt Energy and Crane Worldwide Logistics.

Three terminals, three speed tiers

Amazon has engineered three terminal models to address distinct market segments without forcing a single hardware compromise across all of them. The Leo Nano is the consumer and light-enterprise unit: seven inches square, 2.2 pounds, and rated to 100 Mbps download. The Leo Pro is aimed at small businesses, rural operators, and mobile backhaul deployments: eleven inches square, 5.3 pounds, priced at under $400, and rated to 400 Mbps. The Leo Ultra is the enterprise flagship, a 20-by-30-inch installation weighing 43 pounds and capable of 1 Gbps download with 400 Mbps upload, designed for maritime vessels, commercial aircraft, and large-campus enterprise deployments. Jassy claimed in his shareholder letter that Leo terminals deliver six to eight times better uplink performance and twice the downlink performance compared with the satellite internet alternatives currently available to enterprise customers, a claim that will be scrutinised closely once commercial service begins and independent benchmarks are possible.

The FCC deadline and the launch shortfall

Amazon’s FCC licence for its Generation 1 constellation requires exactly half the planned 3,236 satellites, or 1,618, to be in orbit and operational by July 30, 2026. As of early April 2026, Amazon has between 210 and 241 satellites in orbit, a figure that makes the original deadline effectively unreachable. The company filed a formal request with the FCC in January 2026 for a two-year extension, citing a shortage of available launch vehicles. Alongside the extension filing, Amazon disclosed ten additional Falcon 9 launch contracts with SpaceX and twelve additional New Glenn contracts with Blue Origin. Bezos is betting heavily on orbital infrastructure beyond Leo itself: Blue Origin separately filed with the FCC for a 51,600-satellite Project Sunrise constellation and a 5,408-satellite TeraWave optical backhaul network, making the New Glenn launch pipeline central to multiple overlapping ambitions simultaneously. The FCC separately approved Amazon’s Generation 2 constellation in February 2026, clearing the path to a potential 7,727-satellite network once the current launch bottleneck is resolved. The contracted vehicle fleet now spans Atlas V and Vulcan Centaur (United Launch Alliance), Falcon 9 (SpaceX), Ariane 6 (Arianespace), and New Glenn (Blue Origin).

Taking on Starlink, and the Globalstar play

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Starlink is not a vulnerable incumbent. SpaceX’s satellite internet service generated $10.6 billion in revenue in 2025 at a 54 per cent EBITDA margin and serves more than 10 million paying subscribers across more than 100 countries, operating a constellation of 7,600 to 8,000-plus satellites. SpaceX has filed for the largest IPO in history, seeking to raise $75 billion at a valuation of up to $1.75 trillion, potentially as early as June 2026, which would cement Starlink’s position as a capital-markets-validated infrastructure business before Amazon Leo has completed its initial rollout. Amazon’s response involves two distinct moves. The first is distribution: Leo is being sold primarily through carrier partners and enterprise integrators in its launch phase, using Verizon’s, AT&T’s, and Delta’s existing customer relationships rather than competing for consumers directly. Delta has contracted Leo for in-flight Wi-Fi across 500 aircraft starting in 2028, with free access available to SkyMiles members. The second move is spectrum acquisition. Amazon is reportedly in talks to acquire Globalstar for approximately $9 billion, a deal that would give Leo access to L-band spectrum currently anchoring Globalstar’s existing satellite network and Apple’s emergency satellite connectivity service. Apple holds a 20 per cent stake in Globalstar through a $1.5 billion investment, adding complexity to any acquisition. If the deal closes, Amazon would arrive at commercial launch with not just a new constellation but a second frequency band and an established spectrum position. The year 2025 established satellite internet as a serious enterprise infrastructure market rather than a connectivity experiment, and Amazon Leo’s mid-2026 commercial launch arrives precisely as that market enters its most contested phase.