AstroForge Bets on AI to Pilot Its Next Spacecraft

Why autonomy matters for a startup

Traditional NASA missions rely on large teams of flight controllers on Earth, as illustrated by the OSIRIS-REx asteroid rendezvous in 2018, which required 100 operators working eight‑hour shifts. Such manpower is out of reach for a startup focused on asteroid mining.

From prototype failures to a new AI stack

Founded in 2022 with $56 million in venture capital, AstroForge has already launched two test spacecraft. The 2025 Odin mission reached deep space, but communication difficulties—stemming from the scarcity of large Earth antennas and narrow contact windows—prevented the company from regaining control. The loss highlighted the need for onboard intelligence that could troubleshoot without ground assistance.

“Would that have been recoverable with all the data on the spacecraft? I don’t know, but I can tell you nothing onboard tried it, and I would love something onboard to try if the spacecraft is unrecoverable at launch,” said Matthew Gialich, AstroForge’s co‑founder and CEO.

Gialich framed the choice as a cost calculus: building a global ground network of five dishes would cost roughly $200 million, whereas an autonomous system could eliminate that expense.

Building Solo: a transformer‑based control stack

Armand Awad, head of flight software, explained that AstroForge combined traditional control algorithms with machine‑learning models. Specific subsystems—such as power generation and navigation—use models trained on test data, while an overarching intelligence layer draws from about 2,500 onboard sensors.

“I’m not saying I’m going to make general spacecraft autonomy or general autonomy for the world,” Gialich added. “I’m making a constrained autonomy at a very low sensor input, following the basic training of a transformer model.”

The system, named Solo, is designed to detect anomalies, diagnose root causes—like a misaligned star tracker causing a power anomaly—and execute corrective actions, potentially as simple as a reboot.

First autonomous mission slated for 2027

AstroForge plans to fly Solo on its first autonomous spacecraft in 2027, launching aboard the inaugural rocket from Stoke Space. NASA will support the flight, which aims to collect scientific data about the Sun. This would mark a milestone: most spacecraft still rely on conventional algorithms, and the first neural‑network‑based orbital positioning control was only demonstrated last year.

Proving Solo before full independence

The company’s third vehicle, DeepSpace‑2, is scheduled to launch alongside Intuitive Machines’ third lunar mission by the end of 2026. Solo will ride in “shadow mode,” allowing engineers to test its capabilities while the spacecraft remains under conventional control.

Gialich has stated that Autonomy‑1 will initially operate without any Earth‑based radio links.

“I don’t plan on flying radios that can receive from Earth on Autonomy‑1,” he said. “We have to go all in right now. The team’s probably going to talk me into it by the time we fly it. But right now, I’m telling them no radios.”

If Solo proves itself, AstroForge could demonstrate that a constrained, sensor‑driven AI can safely manage a spacecraft without the massive ground infrastructure that traditional operators depend on.