Product1 publisher3 min readPublished
AT&T routes Amazon Leo through its own managed contract rather than a standalone satellite deal
AT&T Business will sell Amazon's Leo capacity through contracts its customers already hold, which settles who owns the ticket queue well before anyone can quote a price for the satellite layer itself.
The Product Desk · Product desk

What happened
- AT&T Business said Amazon's Leo satellite service will run as a managed connectivity layer on top of its own fiber and wireless network services, sold to AT&T Business customers.
- AT&T contrasted its "unified architecture" with a competitor's "standalone service", pitching one provider and one experience across terrestrial and satellite paths.
- Amazon revealed the final design of its Leo Ultra enterprise terminal late last year, a full-duplex phased array rated for downloads up to 1 Gbps that mounts on a building or sits on the ground.
- The FCC granted Amazon a conditional waiver dropping the requirement to have half the Leo constellation in orbit by July 30, after refusing a February request for a two-year extension.
- The offer is framed around redundancy and extended reach, with AT&T managing the satellite layer alongside the terrestrial services it already supplies.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- decision Without a price or an availability date, the live choice for network teams is not whether to buy Leo but whether to renew or extend the backup path they already pay for at single-homed sites.
- constraint Buying the diverse path from the incumbent that supplies the primary circuit puts both paths inside one contract and one support organisation, which narrows the failure modes the redundancy actually covers.
- contradiction The business-unit pitch treats Leo as enterprise-grade architecture while the CEO describes satellite to investors as corner cases and rural cost relief, and buyers should plan against the second version.
- exposure Sites in areas where AT&T wants assets and costs removed are the ones most likely to end up on satellite by carrier decision rather than by a customer's procurement choice.
The buyer this is aimed at has a few sites with exactly one physical path in and a renewal date on the calendar. What turns that into a purchase order is pricing, a service availability date, and a supported configuration, and AT&T has published none of the three [4].
So the only firm commitment on the table is structural. Leo runs as a managed layer above AT&T's own fiber and wireless services [1], which means the carrier owns provisioning, monitoring and the escalation path for the satellite hop the same way it owns them for the circuit underneath. AT&T says its confidence in Leo is "grounded in extensive technical validation" and that the two companies are moving through a "structured, phased deployment" intended to establish business-grade reliability [9]. Phased deployment is roadmap language, and the missing price list confirms the same thing more plainly: the service is not yet available to buy.
There is real plumbing behind it. AT&T struck a deal earlier this year to move more of its internal operating systems onto AWS while AWS taps AT&T's fiber footprint for cloud connectivity [10], and AT&T is migrating workloads onto a managed on-premises version of AWS Outposts [11]. AWS has already added AT&T's network footprint to its managed Interconnect service, where a private link is provisioned with an AWS-generated activation key [12]. That key is the useful benchmark. If ordering a Leo failover path at a site eventually looks like that, the managed claim is earned; if it looks like a satellite project with a separate account team, it is a reseller arrangement with better branding.
The more informative signal is what AT&T tells investors rather than buyers. CEO John Stankey described satellite as fitting "corner cases" in the carrier's operations [13], and specifically as a way to shrink the terrestrial footprint at what he called "poverty sites", low-traffic locations where satellite could keep continuity while improving cost structure [14]. He was blunt about the direction: he is trying to shrink the fixed footprint, shutting down wire centers, and wants assets and costs out of those areas, staying prevalent where he can build fiber and hang wireless off it [15]. Enterprise redundancy and rural cost relief are two different products with two different budgets. The rural cost relief version has a named executive sponsor; the enterprise redundancy version does not.
On timing, the FCC's conditional waiver removed the mandate to have half the constellation up by July 30 [7], which leaves "commercial service later this year" as the only dated public statement, against roughly 400 satellites currently in orbit [16].
For sites, the sort that matters is two questions: whether the location already has a second terrestrial path at a cost you would actually approve, and whether AT&T is already the primary carrier there. Only the no-second-path, AT&T-primary quadrant is a Leo candidate in the next planning cycle; everywhere else, evaluating it now is a procurement exercise without a price. And the clause worth arguing over at renewal is attribution: whether minutes lost on the satellite layer count against the site's service level or against a separate satellite agreement that nobody in your organisation has read.
What to watch
- Whether AT&T publishes Leo pricing and availability, and whether satellite outage minutes count against the site service level or a separate satellite agreement.
- Whether Amazon Leo enters commercial service this year from roughly 400 satellites in orbit, now that the July 30 milestone has been waived.
- Whether AT&T names specific wire centers or rural sites moving onto satellite as it pulls assets out of low-traffic areas.