Amazon Leo (formerly Project Kuiper) Satellite Internet

Amazon Leo is the rebranded Project Kuiper — a low Earth orbit broadband network that Amazon has been building toward for years and that finally has a name, an enterprise beta and a stated commercial target. On paper it is aimed at everyone Starlink currently serves by default: rural households with no wired option, airlines, telcos looking for backhaul, energy and logistics operators working well outside cell coverage, and government users. The three-tier terminal lineup makes that segmentation unusually explicit, with a small consumer-grade unit at one end and an enterprise box at the other.

What stands out at this stage is the seriousness of the engineering story rather than the customer experience. Custom baseband silicon, 100 Gbps optical links between satellites and a three-shell orbital design all suggest Amazon is not building a token competitor. The beta partner list is likewise a who's-who of carriers and aviation, which tells you Amazon is courting enterprise revenue first and consumers second. What it does not tell you is what any of this will cost, when you can actually buy it, or how it performs when thousands of subscribers share a beam.

The honest verdict is that this is not yet a reviewable product, and readers deserve to hear that plainly. There is no published price, no consumer availability, and the timeline itself is contradictory: a mid-2026 commercial launch was confirmed in April 2026, yet by mid-September 2026 service still had not started and no new satellites had gone up since early July. The FCC's decision to grant a limited, conditional waiver of the interim deployment milestone is the clearest signal available that deployment is running behind the licensed schedule.

The hardware claims are genuinely competitive if they hold up in the field. A seven-inch, 2.2-pound consumer terminal is small enough to be a real convenience advantage, and 100 Mbps is more than adequate for a household that currently has nothing. The Leo Ultra's up-to-1 Gbps down and 400 Mbps up figures are the kind of numbers that make satellite viable for branch offices and remote industrial sites. But these are headline maximums with no latency figures, no capacity-per-cell data and no independent testing behind them, and a constellation of a few hundred satellites cannot deliver that consistently at scale. Coverage and reliability will live or die on getting several thousand satellites up.

I'd rate this as promising infrastructure with a credible backer and an execution problem. If you're an enterprise buyer, it's worth tracking and worth a conversation. If you're a consumer hoping to cancel your existing connection, there is nothing here to buy and no price to plan around. The rating reflects real technical substance discounted heavily for slipping schedules, missing pricing and the absence of any verified real-world performance.