AstroForge plans a spacecraft that takes no commands from Earth

Why it matters

AstroForge's 2025 Odin mission exposed how much a deep-space startup depends on scarce ground communications. Autonomy-1 tests whether onboard decisions can reduce that dependence without putting the mission beyond recovery.

Matthew Gialich (@MattGialich) plans to put AstroForge's next spacecraft beyond the reach of ground commands. In a September 22nd announcement on X, AstroForge said Autonomy-1 will launch in 2027 and attempt every mission objective without a command from Earth after separating from its rocket. It is a plan for a flight, not a claim that the software has already done it.

Gialich, AstroForge's co-founder and CEO, previously led flight software and guidance, navigation and control teams at Virgin Orbit, then ran vehicle software at Bird. He founded AstroForge with Jose Acain, a former SpaceX and NASA engineer, to pursue asteroid mining. The new mission puts a problem from AstroForge's own flight history at the center of that ambition: keeping a spacecraft working when Earth cannot readily help.

AstroForge calls the onboard decision-making system Solo. According to AstroForge, it monitors the spacecraft's condition, identifies problems and coordinates responses across onboard systems. Solo sits above the deterministic, physics-based flight software that handles the underlying control tasks. That distinction matters. AstroForge is proposing to move mission coordination onto the vehicle, rather than replace every established flight algorithm with a model.

The sequence starts with DeepSpace-2, AstroForge's asteroid-bound spacecraft. TechCrunch reported that Solo will fly on that mission in "shadow mode," allowing engineers to assess its decisions before giving it control on Autonomy-1. AstroForge says Autonomy-1 will carry a NASA Goddard heliophysics payload and fly on Stoke Space's Nova Pathfinder. Stoke said on September 8th that Pathfinder, its first orbital configuration, was targeting a first flight in the first part of 2027. Autonomy-1's schedule therefore also depends on a rocket that has yet to make that flight.

The case for handing Solo more authority comes from Odin, AstroForge's 2025 deep-space mission. In its mission debrief, AstroForge described brief two-way contact with Odin and a series of ground-station failures that cost it crucial early opportunities to communicate. Odin did not complete its asteroid mission. AstroForge subsequently changed DeepSpace-2's communications and testing plans, according to Gialich's September account of the engineering work. Autonomy-1 extends that response: a vehicle that can make decisions locally has less need to wait for an available dish and an operator's instructions. Whether Solo can resolve the kinds of failures that affected Odin remains a flight-test question.

The economics are as important to Gialich as the flight demonstration. AstroForge says ground operations, including dish time, telemetry analysis and commanding, account for nearly a third of its overall mission cost. That is AstroForge's own cost estimate, not a saving Autonomy-1 has established. Reducing that burden would help its plan to fly spacecraft repeatedly in search of valuable asteroid material. AstroForge came through Y Combinator's Winter 2022 batch; Initialized Capital has also identified itself as an early investor.

Autonomy-1's proposed operating rules are unusually stark, though they may change before launch. Gialich told TechCrunch he did not plan to fly radios capable of receiving commands from Earth, while acknowledging that his engineers might persuade him otherwise. A spacecraft can send data home and still take no commands; the claim concerns the uplink and the absence of a human operator directing the mission. For AstroForge, the decisive test will be whether Solo can identify a problem, choose a safe response and finish the assigned work when ground control has no way to intervene.