For the first time, its SwarmOS platform extends into orbit, integrating satellites with drones, ships, and ground assets under one embodied AI orchestration layer.
The contract signals a shift in how the Pentagon approaches multi-domain autonomy — from platform-specific control to heterogeneous swarm coordination with no single point of failure.
Drones fly. Ships sail. Satellites orbit. And until now, each operated inside its own informational silo — seeing what its own sensors saw, responding to its own mission thread, blind to what the others knew.
The paradox of modern autonomous warfare is that the more platforms you field, the less you see.
More data. Less understanding.
The Air Force Research Laboratory just awarded the company a $4.2 million contract to solve exactly that contradiction. The project is called HANGTIME — Hierarchical Adaptive Networked Game-Theoretic Integration of Multiple Echelons — and it extends its SwarmOS platform into orbit for the first time.
Satellites, drones, ships, and ground systems will share a single AI framework. They will share intelligence, adapt to changing conditions, and act in sync across four domains simultaneously.
A Single Brain for Distributed Bodies
SwarmOS is the defence variant of the company's Pilot embodied AI software. It was designed for one hard problem: making heterogeneous platforms from different manufacturers collaborate without depending on centralized infrastructure that adversaries can target.
The standard approach today is platform-specific. A drone runs its own autonomy stack. A ship runs another. A satellite — if it has any onboard intelligence at all — runs a third. They do not talk to each other unless a human operator bridges them.
HANGTIME inverts that. One AI framework sits above all four domains. Platforms retain their individual control systems, but SwarmOS becomes the orchestration layer — assigning tasks, fusing sensor data, and redistributing intelligence across the entire network in real time.
Palladyne CTO Dr. Denis Garagic described it as a team that thinks and acts together, sharing what they see and learning as conditions change. AFRL program manager Caleb Williams called it a critical step in multi-domain autonomy for coordinated execution in challenging environments.
Satellites Enter the Swarm
The satellite integration is what makes HANGTIME different from earlier multi-domain experiments. Previous cross-domain demonstrations — Joint All-Domain Command and Control (JADC2) efforts included — connected platforms through data links and human decision-makers. HANGTIME connects them through an autonomous AI that decides, not just transmits.
The patented SwarmOS operates on edge hardware in the field, in degraded and denied communications environments. It does not need a cloud backhaul or a satellite link to function — though when satellites are available, it incorporates their sensor data as another input stream.
This matters for a simple reason: space-based sensors see the widest picture, but that picture is useless if it cannot reach the tactical edge faster than the adversary can move. SwarmOS collapses the decision loop by processing satellite data alongside drone and ship data inside the same AI model, at the edge.
Four Domains, One Decision Loop
The contract covers integration across space, air, maritime, and land domains. The company has already demonstrated cross-platform autonomous flight with Red Cat (three heterogeneous UAS operating collaboratively) and holds a U.S. patent for multi-domain swarm coordination. HANGTIME scales that from a handful of air vehicles to a full multi-domain operational network.
AFRL's choice of the company is itself revealing. The lab awarded the contract under its Disruptive Applications program — the same vehicle that selected it for Army SwarmOS and Gremlin-X contracts in June. It is not a prime contractor in the traditional sense (no 100,000-employee workforce, no Capitol Hill lobbying apparatus). It is a 90-person engineering company in Salt Lake City.
That the Pentagon is entrusting a core multi-domain coordination problem to a company of this size tells you something about where the defence innovation centre of gravity is shifting.
Not to Beltway primes. To engineers who ship software.
What happens when autonomous swarms become the default operating model for multi-domain operations?
Probability: 70% — AFRL's HANGTIME is an SBIR/STTR-scale validation. The natural next step is a tiered prototyping contract (e.g. OTA or Middle Tier of Acquisition). The incumbent primes (Lockheed, Northrop, Boeing) will bid their own solutions, but the architectural advantage belongs to companies that build software-first, hardware-agnostic platforms — exactly the position this company holds.
✅ Arguments for
AFRL's Disruptive Applications program has a proven pipeline from SBIR to major procurement (see how Anduril's Lattice scaled from prototype to SOCOM integration).
The Pentagon's own CJADC2 roadmap explicitly calls for heterogeneous, vendor-neutral orchestration — SwarmOS matches that requirement better than any prime's proprietary stack.
Confirmation criteria: The company wins a follow-on contract valued above $50M within 18 months, or is selected as the autonomy integrator for a major platform program (e.g.CCA, RCCTO).
❌ Arguments against
Primes have their own multi-domain orchestration programs (Lockheed's JADO, Northrop's IBCS) and will lobby aggressively to keep the architecture inside the traditional industrial base.
SBIR-to-POR transition is rare: fewer than 10% of AFRL SBIR contracts reach a program of record. The technical win does not guarantee the procurement win.
Disconfirmation criteria: AFRL does not award a follow-on within 24 months, or the follow-on goes to a prime (Lockheed, Northrop) instead.
The company's Q3/Q4 earnings calls — any mention of HANGTIME milestones, follow-on funding, or prime partnership
AFRL BAA awards in the autonomy and multi-domain C2 category — competing small-business awards would validate the category even if it does not win them all
Anduril and Shield AI contract announcements in multi-domain autonomy — if either wins a comparable AFRL award, the category is on a POR trajectory
Pentagon budget submissions for FY2028 — look for a new Program Element under PE 0604710F (Advanced Aerospace Systems) for heterogeneous swarm C2
Development scenarios
🟢 Optimistic scenario (25%)
Implications: Its revenue trajectory shifts from sub-scale R&D contracts to production-adjacent delivery. The stock re-rates toward defence-software multiples (15-20x revenue).
🟡 Base-case scenario (55%)
Implications: It remains a viable niche supplier but does not achieve prime-level scale. Revenue grows gradually from existing contract vehicles. The technology direction is validated; the business model remains unproven at scale.
🔴 Pessimistic scenario (20%)
Implications: The company's defence revenue plateaus below $20M annually. The SwarmOS architecture remains technically respected but commercially marginal. Shareholder dilution funds ongoing R&D without a clear path to break-even on the defence side.
The HANGTIME contract is small — $4.2 million is a rounding error in a $900 billion defence budget. But it belongs to a category of awards that matter more than their dollar value suggests: first-of-kind architecture contracts that define how a technology domain will be structured.
AFRL could have given this work to Lockheed. It gave it to a 90-person company in Salt Lake City.
The signal is not about the company. It is about what the Pentagon believes it needs — and who it trusts to build it.