Today, Amazon officially killed the “Project Kuiper” name, rebranding its satellite initiative to Amazon Leo, and unveiled a hardware lineup designed to challenge SpaceX’s dominance in orbit. While the headline act is the “Ultra” antenna, Amazon has revealed a full family of terminals – the Ultra, Pro, and Nano – each targeting a specific slice of the connectivity market.
But make no mistake: this launch is a calculated strike against Starlink’s one remaining weakness – enterprise reliability.
Amazon Leo Hardware: Ultra, Pro, and Nano
At the top of the food chain is the Leo Ultra, a 20-by-30-inch “full-duplex phased array” terminal. Amazon claims it is the fastest commercial terminal in production, capable of 1 Gbps download speeds and, more critically, 400 Mbps upload speeds.
That upload figure is the “killer app.” For comparison, Starlink’s High-Performance Business kits generally top out at around 25 Mbps for uploads. If Amazon can deliver consistent 400 Mbps uplinks, it will become the default choice for data-heavy industries like geological surveying, live broadcasting, and remote telemetry, which need to send massive data to the cloud, not just consume it.
Below the Ultra, Amazon introduced two other tier options:
- Leo Pro: A standard 11-inch antenna designed for most businesses, supporting downloads up to 400 Mbps.
- Leo Nano: A highly portable 7-inch square terminal, offering speeds up to 100 Mbps, likely targeted at mobile workforce and logistics.
The AWS Moat: Private Networking
While Starlink has spent years aggressively capturing the consumer market – now boasting 8 million active subscribers globally as of this month – Amazon is pivoting hard to the enterprise.
The Leo network features “Direct to AWS,” allowing data to bypass the public internet entirely. Traffic moves from the dish to Amazon’s ground station and straight into the AWS cloud via Direct Connect. For a tea farm in Kericho or a government office in Mombasa, this means enterprise-grade security without the “wild west” risks of the public web.
The Reality Check: 153 vs. 8,800
However, the numbers expose the massive hill Amazon has to climb. The “1 Gbps” promise is currently supported by a constellation of just 153 satellites.
Contrast that with SpaceX. As of late 2025, Starlink has approximately 8,800 active satellites in orbit. Starlink has expanded aggressively, adding 42 new countries in the last year alone. Amazon is trying to fight a war on speed because it has lost the war on scale. It needs to launch thousands of satellites by mid-2026 to meet FCC license requirements, a deadline that is fast approaching.
The “Kenya Argument”: Why Slow Might Be Better
Amazon’s “slow and steady” enterprise focus might actually solve a specific pain point that has plagued Starlink in Kenya recently.
Starlink’s aggressive expansion in Kenya hit a wall earlier this year. In Nairobi, overwhelming demand caused network congestion, forcing Starlink to temporarily pause new subscriptions in high-density areas like Kiambu and Nairobi. Users reported speeds dropping to as low as 47 Mbps during peak times.
This vulnerability is Amazon’s opportunity. By limiting access to high-value enterprise clients rather than millions of households, Amazon can likely guarantee the throughput that Starlink’s congested cells struggled to maintain. They aren’t trying to connect every home in Westlands; they want to connect the telco’s backhaul and the bank server rooms.
The Verdict
Amazon Leo is technically superior on paper, especially with its 400 Mbps upload speeds. But until Amazon can launch rockets faster than Musk – currently a tall order – it remains a premium niche product for companies already locked into the AWS ecosystem.
