Amazon is getting ready to flip the switch on its Leo satellite internet service, and the biggest clue is its rocket schedule: the company now plans to more than double its annual launch rate to more than 20 missions a year as it moves from experimental deployments into the early stages of commercial service. For a project that has spent years living under the codename Project Kuiper, the new cadence is Amazon’s way of signaling that Leo is no longer a slide deck—it is a network that needs fresh hardware in orbit, fast.

Right now, Amazon has launched more than 200 Leo satellites and already has another 200-plus stacked, tested, and sitting in payload processing facilities in Florida and French Guiana, waiting for their ride to space. That puts Amazon in the somewhat surreal position of running the third‑largest active satellite system on the planet—behind SpaceX’s Starlink and China’s growing constellations—before a single consumer in a rural village, fishing vessel, or remote oil field has paid for a monthly Leo plan. Every new launch adds coverage and capacity, and Amazon is now openly saying it is “on pace” to complete 11 missions in its first year of deployment, then push beyond 20 missions in year two as heavy‑lift rockets come online.

On the ground, Leo already looks less like a scrappy startup and more like a scaled Amazon operation. The satellites themselves are built at a dedicated factory in Kirkland, Washington, which can produce up to 30 spacecraft per week when running at full tilt. Amazon has dialed production up and down based on how ready launch vehicles are, but the key detail is that Leo can manufacture “multiple satellites per day” and has “hundreds of flight-ready satellites” in storage, ready to load into rocket fairings as soon as slots open up. To support that pace, Amazon has built a 100,000‑square‑foot payload processing facility at Cape Canaveral—described as the largest of its kind in the world—and is duplicating some of that capability on the U.S. West Coast at Vandenberg.

The rocket lineup is where Leo’s ramp‑up becomes very visible. Amazon is using a mix of legacy and next‑gen heavy‑lift vehicles: Arianespace’s Ariane 64, Blue Origin’s New Glenn, United Launch Alliance’s Vulcan Centaur, and ULA’s older but ultra‑reliable Atlas V. The first major heavy‑lift mission of 2026, dubbed Leo Europe 1 (LE‑01), flew in February on Ariane 64 and carried 32 satellites in a single shot, with future Ariane 6 configurations expected to cram even more hardware into the same 7‑meter fairing. New Glenn and Vulcan Centaur will push that even further; initial Leo missions will pack up to 48 satellites on New Glenn and 40 on Vulcan, with headroom to increase payloads as those rockets mature.

Even Atlas V—the workhorse rocket quietly heading toward retirement—is getting a last‑minute upgrade for Leo. Working with ULA, Amazon’s engineers took advantage of a higher‑performance RL10C upper‑stage engine and redesigned their dispenser system to add a fourth satellite “layer,” bumping the Atlas payload from 27 to 29 Leo satellites and setting a new record as the heaviest payload ever flown on that rocket. It is the kind of incremental optimization that matters when you are trying to put thousands of satellites into orbit on a tight schedule and every extra kilogram of payload turns into more coverage on Earth.

Behind the scenes, Amazon has been writing very large checks to make sure those rockets can launch often enough. The company says it has invested more than 200 million dollars into infrastructure upgrades at ULA’s facilities at Cape Canaveral, including a dedicated vertical integration building just for Leo missions, a new rail system and a second mobile launch platform to shuttle stacked rockets to the pad, and even a second ship to ferry boosters from the factory in Alabama to Florida. In practical terms, these upgrades are designed to let ULA double its launch capacity from Space Launch Complex 41, cutting down the time between flights not just for Amazon, but also for other government and commercial customers sharing the pad.

All of this rocket and factory talk is really in service of one thing: turning Leo into a real broadband option for people and businesses that have never had decent connectivity. Amazon’s satellite network is built around a planned constellation of just over 3,200 low Earth orbit satellites, connected with high‑speed optical links and a global mesh of ground gateways and fiber interconnects. Customers on the ground will connect through three main antennas: Leo Nano, a compact square dish aimed at lighter‑duty residential and consumer use; Leo Pro, a slightly larger unit that can deliver up to about 400Mbps; and Leo Ultra, an enterprise‑grade phased‑array antenna advertised at up to 1Gbps down and roughly 400Mbps up.

In November 2025, Amazon quietly began an enterprise preview of Leo, seeding early hardware into the hands of selected businesses and governments to test performance and management tools in the real world. Those trials emphasized things that matter less to a rural home user and more to a CIO: private network interconnects, end‑to‑end encryption, and the ability to link remote sites directly into Amazon Web Services without ever traversing the public internet. For Amazon, this is a strategic advantage; Leo is not just another satellite ISP; it is an on‑ramp into the broader AWS ecosystem for mines, ships, offshore rigs, and remote offices that currently rely on patchy VSAT links or expensive private circuits.

Timelines have been a bit messy, which is one reason the new 20‑plus‑launch‑per‑year target matters so much. Under its original license, Amazon was supposed to have half of the Leo constellation—about 1,618 satellites—deployed and operating by July 2026. Launch delays, limited heavy‑lift availability, and a year‑ish of schedule slippage meant that by early 2026, Amazon had orbited only a few hundred satellites and was warning regulators it would likely be short of that milestone, even with an impressive ramp. The company has asked the U.S. Federal Communications Commission for an extra 24 months, to July 2028, to hit the halfway mark and argues that the combination of multiple launch providers and over 100 missions booked gives it enough margin to catch up.

For end users, the exact regulatory date matters less than the practical question: when does Leo start feeling like a real option? The answer is: in pockets, soon, and more broadly as the 20‑mission‑per‑year cadence kicks in. Amazon and its partners say early commercial availability is aimed at 2026, with particular focus on markets like the U.K. and other priority countries where spectrum and regulatory approvals are already in place. As more satellites go up, coverage should expand toward roughly 56 degrees north and south latitude, which includes most of Europe, North America, and big chunks of the Southern Hemisphere.

Of course, Leo will not exist in a vacuum. Starlink already has several thousand satellites in orbit, millions of subscribers, and widely available hardware, especially in rural North America and parts of Europe. But Amazon does not need to beat Starlink on raw satellite count to matter; instead, Leo is positioning itself as a tightly integrated part of Amazon’s broader devices and cloud stack. Think eero‑style home networking setups that can fail over to Leo during outages, ships and remote industrial sites that treat Leo as a kind of wireless AWS Direct Connect, and governments that want a second major supplier in the LEO broadband market.

The economic story is just as interesting. Satellite broadband historically has been a niche, expensive, and often frustrating service of last resort. Low Earth orbit networks flipped that assumption by dramatically lowering latency and unlocking modern‑feeling connectivity in places where fiber is impossible and 5G is a marketing phrase on a coverage map. If Amazon executes on this accelerated launch plan, Leo will add another large-scale player to that mix, likely pushing prices down over time and giving rural users, enterprises, and maritime operators genuine choice instead of “take it or leave it” contracts.

In that sense, the decision to double the annual launch rate is not just about hitting a regulatory checkbox or bragging about rocket statistics. It is a signal that Amazon is crossing the line from experiment to infrastructure, with Leo poised to become one more invisible layer of connectivity in the background of people’s lives and businesses—just like AWS did for servers a decade ago, and Prime did for shipping. The next 18–24 months, and whether those 20‑plus launches a year actually fly on schedule, will determine how quickly that vision turns into something as mundane—and as essential—as a Wi-Fi password on the fridge in a place that never had decent internet before.