Jeff Bezos to Launch 2nd Satellite Broadband Service with 6Tbps Speeds, TeraWave

Talk about confusing. The founder of internet retail giant Amazon, Jeff Bezos, has announced that he intends to launch a second mega constellation of 6Tbps (Terabits per second) symmetric speed capable broadband satellites, albeit via his Blue Origin rocket company, called TeraWave, which seems to conflict a bit with the existing Amazon Leo project.

We’ll start with a recap. The existing Amazon Leo service (formerly Project Kuiper) is currently in the commercial beta phase and will start to launch properly during 2026. The service currently has approval to deploy and operate their own initial constellation of 3,236 LEO satellites (altitudes of between 590km to 630km). A total of over 180 Kuiper satellites have already been placed into orbit (they need at least 500 for basic global coverage) and many more are due to follow over the next few years.

The initial service that Amazon plans to launch will offer three different end-user terminals to cater for different markets and performance levels. Broadband speeds are expected to target 100Mbps (20Mbps upload) for roaming users, before rising to 400Mbps (100Mbps upload) for homes and business users, then topping out at 1Gbps (400Mbps) for enterprises and government. Latency times should be similar to Starlink’s, but details on pricing have yet to be disclosed.

Suffice to say that Amazon’s approach is already starting life at somewhat of a disadvantage, given that it will have to go up against Starlink’s massive and better-established network. The context here is important because Amazon’s founder, Jeff Bezos, has just announced – via his Blue Origin rocket company – that he now intends to launch a second mega constellation of broadband satellites, which seems to conflict a bit with Amazon’s project.

The new TeraWave on the block

The new TeraWave constellation is described by the CEO of Blue Origin, David Limp, as being “purpose-built for enterprise customers” (the network envisages serving around 100,000 such users) and able to offer “unmatched speeds of up to 6 Tbps through a multi-orbit constellation of 5,280 LEO and 128 MEO satellites with both RF and optical links“. At least that’s the goal, but it hasn’t launched any satellites yet.

Globally distributed customers will each be able to access up to 144Gbps (Gigabits per second) of capacity through Q/V-band links from LEO satellites (orbiting at between 520-540km), while up to 6Tbps point-to-point capacity can be accessed through optical links from satellites in a higher Medium Earth Orbit (MEO) – orbiting between 8,000 and 24,200km. MEO is good for coverage, but latency will suffer at those distances.

“This provides the reliability and resilience needed for real-time operations and massive data movement. It also provides backup connectivity during outages, keeping critical operations running. Plus, the ability to scale on demand and rapidly deploy globally while maintaining performance,” said Limp.

The official announcement adds that the service has been designed to address the “unmet needs of customers who are seeking higher throughput, symmetrical upload/download speeds, more redundancy, and rapid scalability,” before mentioning that it is intended to “complement fibre backhaul” rather than replace it.

The move is confusing given how Amazon Leo is also targeting the same enterprise market, albeit not to anything like the same levels of data performance. As such it could be seen as a response to SpaceX, which is working to launch Starshield using Starlink technology for customers in the government and military sectors.

However, Starshield does at least acknowledge and take advantage of Starlink’s existing network, while echoing some of the same branding. By comparison, Blue Origin seems to be keeping Amazon Leo and TeraWave totally separate, which seems like a recipe for unnecessary confusion and lost marketing potential. On the other hand, Jeff Bezos only owns about 8% of Amazon’s stock today (largest shareholder), so he’s not in total control there.

In any case, the deployment of the TeraWave constellation isn’t scheduled to begin until Q4 2027 and it will then take several years to fully deploy. The new constellation will also need to harness a sizeable amount of radio spectrum across those Q/V, E, Ka and S bands in order to deliver such capacity, which might present some challenges for regulators in other countries to consider.