SpaceX's $4.16B Golden Dome deal: SB-AMTI sensor satellites, not interceptors

SpaceX won a $4.16 billion contract from the U.S. Space Force. The Verge reports it as missile-tracking satellites tied to President Donald Trump’s Golden Dome initiative, but reading the Space Systems Command announcement, the contract’s name is Space-Based Airborne Moving Target Indicator — SB-AMTI for short. It’s not a satellite that shoots missiles down; it’s an order for a sensor layer that detects and tracks airborne targets from space. The contract is an OTA (Other Transaction Authority) agreement, awarded May 29, 2026. An OTA is the acquisition vehicle the military uses for prototyping and short-term development — a faster lane than standard procurement rules.

Golden Dome news tends to draw loud headlines as “putting interceptors in space.” What was ordered here is the target-tracking layer that sits in front of that. It picks up sensor data from low-orbit satellites and passes it through comms links and ground processing to the Joint Force. Before any question of intercepting, it’s the part that continuously picks up what’s flying where.

Golden Dome’s total estimate is put at $185 billion, up $10 billion from the original. Adding the data network described below to this SB-AMTI award, SpaceX’s Golden Dome-related contracts come to roughly $6.45 billion. But SpaceX didn’t take the whole thing including interceptors. A separate group of companies is in for the space-based interceptor prototype; SpaceX’s role is strictly the find-and-carry side.

SB-AMTI is a satellite constellation for tracking airborne targets

Space Systems Command describes SB-AMTI as a sensor layer for “tracking and targeting airborne threats from space.” What it tracks isn’t just aircraft but also low, fast movers like drones and cruise missiles. Existing aircraft-based surveillance struggles to get close in airspace where the adversary’s A2/AD is strong. A2/AD stands for anti-access/area-denial — the mix of air defense, electronic warfare, and long-range weapons that keeps forces and sensors at bay. Instead of flying an aircraft into that airspace, the design tracks continuously from space.

The “Airborne Moving Target Indicator” in the name says what it is. AMTI is the technique for distinguishing airborne moving targets, a kind of radar MTI (moving target indication). MTI uses the Doppler effect to separate moving targets from stationary background. MTI has a counterpart: tracking ground moving targets is called GMTI (Ground Moving Target Indicator).

Type	What it tracks	Examples
AMTI	Airborne moving targets	Aircraft, drones, cruise missiles
GMTI	Ground/sea moving targets	Vehicles, ships

Both pick up movers by Doppler; what differs is the space they watch. AMTI has to output the target’s position, speed, heading, and altitude to friendly aircraft, a tougher bar than tracking ground targets with GMTI. The Space Force is separately planning to launch ground-target-tracking satellites in 2028; SB-AMTI is the airborne-target version.

In an April Space Force article, SB-AMTI is described not as a single satellite but as a system-of-systems combining sensors, AI-assisted ground processing, and secure comms links. The May 29 announcement likewise says SpaceX’s $4.16 billion OTA agreement is to build the initial SB-AMTI capability, with a satellite constellation expected to deploy by 2028. In positioning, it supplements from space the airborne early-warning aircraft (the U.S. Air Force’s E-7 Wedgetail) that can no longer get close to dense air defenses. Col. Ryan Frazier, who leads the effort, said putting this capability in the space domain gives the Joint Force persistent battlefield awareness over contested airspace.

Lump Golden Dome into one giant weapon and SB-AMTI’s place gets misread. This SB-AMTI is not the interceptor itself. Of the sensor to find airborne targets, the network to carry data, the software to process it on the ground, and the intercept capability that comes after — it’s closest to the first layer.

The sensor layer and the intercept layer play different roles. The sensor layer continuously picks up and carries “what is flying where, and how”; SB-AMTI and the data network below sit here. The intercept layer “actually shoots down the tracked target”; space-based interceptors and the like belong here. A different company was picked for that interceptor prototype — True Anomaly, for example, was selected for a space-based interceptor prototype in April and raised $650 million that same month. What SpaceX took is the sensing and transport — the find-and-carry side.

SDN Backbone also went to SpaceX days earlier

SpaceX’s large contracts don’t stop there. On May 27 (announced Tuesday evening), Space Systems Command also issued SpaceX a $2.29 billion OTA contract for the Space Data Network Backbone. This is a comms foundation of many satellites in low orbit (proliferated LEO), a mesh network for carrying data from sensors to weapon systems at low latency and high volume. The satellites are meant to connect via laser links (optical comms), running global tactical communications and broadband without relying on ground relays. The full-operational prototype deadline is set for the end of 2027.

Rather than replacing the existing SDA Transport Layer, this SDN Backbone combines with it. The SDA Transport Layer is the low-orbit comms and data-relay network the Space Development Agency has been building, with tranche 1 and 2 already operating. The Space Force says it will zero out the budget for future Transport Layer tranches in fiscal 2027, while running the existing tranche 1 and 2 alongside the SDN Backbone and folding them into a single open architecture called the Space Data Network. This data network supports not just Golden Dome but also the communications of the all-domain command-and-control concept CJADC2.

So in a few days at the end of May, SpaceX took back-to-back large contracts for the sensor layer and the data-transport layer around Golden Dome. As The Verge notes, a separate company is in for the interceptor prototype, so it isn’t a setup where SpaceX alone holds intercept, sensing, and communications. Still, as the company with low-orbit satellite mass production and laser links, it sits close to the center of the orders.

graph TD
  T[Airborne threats<br/>aircraft, drones, cruise missiles] --> S[SB-AMTI<br/>space sensor layer<br/>4.16B USD]
  S --> N[SDN Backbone<br/>data-transport layer<br/>2.29B USD]
  SDA[SDA Transport Layer<br/>tranche 1 and 2] --> N
  N --> P[AI ground processing]
  P --> J[Joint Force<br/>decision and intercept]
  J -.handled by other firms.-> I[space-based interceptors, etc.<br/>intercept layer]

The Space Force side also writes that it won’t concentrate on a single vendor. SB-AMTI had a pool of nine vendors assembled as of April (names withheld for security reasons), and it plans to issue several more contracts over the next year to widen the field. Col. Ryan Frazier also said they’ll work with a diverse pool of traditional and non-traditional vendors rather than relying on a single provider. On the budget side, the fiscal 2026 reconciliation allocation requested $9.2 billion for target tracking and fiscal 2027 requested $7 billion for SB-AMTI, with this initial contract as the entry point. Still, if the large initial-capability contract went to SpaceX and the SDN Backbone also went to SpaceX, the actual interface design and schedule get decided on the premise of SpaceX’s implementation. A multi-vendor policy is stated, but the initial core contracts are concentrating at SpaceX.

Japan is standing up the same two layers for defense

The same structure of “stacking a sensor layer and a comms layer separately” started moving in Japan in early 2026. The dollar figures aren’t on the order of SB-AMTI or the SDN Backbone, but in February 2026 Japan’s Ministry of Defense signed contracts in quick succession for the find side and the carry side.

The sensor layer is the satellite constellation development-and-operation program. On February 19, 2026, seven companies — Mitsubishi Electric, SKY Perfect JSAT, and Mitsui & Co., plus SAR-satellite makers Synspective and QPS Research Institute, optical-satellite maker Axelspace, and Mitsui Bussan Aerospace — set up a special-purpose company, TriSat Constellation, and signed with the Ministry of Defense. SAR stands for synthetic aperture radar; because it images the ground with radio waves, it can catch targets at night or under clouds. The program’s purpose is to supply the imagery needed for stand-off defense — striking distant targets with long-range missiles — at high frequency and stably, via a constellation of imaging satellites and dedicated ground facilities. The idea of holding the “eyes” for striking in a low-orbit satellite group overlaps with SB-AMTI.

The comms layer is the next defense communications satellite. Also in February 2026, Mitsubishi Electric won the contract for the successor to the current X-band defense communications satellite “Kirameki-2.” With strengthened jam resistance and a digital communications payload that lets beam coverage and capacity be changed after the fact, it’s the side that carries data from sensors to units. It’s the layer that corresponds to the U.S. SDN Backbone.

That said, the contents aren’t identical. What SB-AMTI tracks are airborne moving targets like aircraft and drones, while what TriSat acquires is imagery. Tracking targets like North Korean missiles or Chinese and Russian hypersonic glide vehicles (which fly low at more than five times the speed of sound and change course to evade interception) moves separately — as a U.S.-Japan low-orbit detection satellite network and as an infrared-sensor demonstration aboard the HTV-X cargo spacecraft. The procurement shape contrasts too: where the U.S. concentrated the core at SpaceX, Japan splits the roles across a seven-company SPC and Mitsubishi Electric.

The story is the LEO sensors and ground processing, not the rockets

This blog earlier covered Artemis II and the cislunar Lagrange points. That was about how to move people and things near the Moon. This SB-AMTI is the same space infrastructure, but on the side that lines up sensors in low orbit, connects them with a comms network, and turns them into target data through ground processing. More than the rockets or satellites themselves, the latency and connectivity from sensor to decision come to the front.

If you follow the contracts, the next thing to watch isn’t the satellite count itself but the connection terms among SB-AMTI, the SDN Backbone, and the SDA Transport Layer. Space Systems Command writes that for the SDN Backbone it will build the LEO portion’s hybrid mesh data network together with the SDA’s Transport Layer. By what route the data SB-AMTI picks up gets carried, through what processing, and to which system. The more that comes out as public information, the easier it is to track whether Golden Dome is moving from a concept stage into orders for an operational architecture.

A lot also remains undecided. The specific number of satellites, orbits, sensor methods, data-processing latency, sharing scope with allies, and the connection to the live-intercept side aren’t clear from the announcements alone. Whether the end-2027 SDN Backbone prototype and the 2028 SB-AMTI initial deployment proceed at the same tempo isn’t clear yet either. If those two deadlines slip apart, you get an asymmetric state — sensors but no way to carry the data, or a network but not enough target data coming in.