WASHINGTON — Slingshot Aerospace today announced a $27 million contract to help modernize training of Space Force Guardians, including the use of the company’s TALOS AI to simulate an adversary’s actions during orbital warfare scenarios.
Trained on Slingshot’s extensive library of real-world orbital observations, the AI is meant to respond realistically and dynamically to the trainees’ moves in the wargame, the company said, rather than following a rigid, pre-programmed script.
“This ensures that TALOS can easily adapt to new scenarios as the space environment evolves and can dynamically exploit new algorithms and tools as soon as they become available,” Slingshot CEO Tim Solms told Breaking Defense in an email.
Slingshot and its subcontractors will also provide other software tools to simulate friendly (“Blue”) forces and to act as virtual referees (“White”), Solms said.
The 18-month contract was awarded through a Space Force Commercial Solutions Opening (CSO), a streamlined procurement process made popular by the Defense Innovation Unit. It builds on a 39-month 2022 Strategic Funding Increase award worth $25 million that allowed the Space Training and Readiness Command to test TALOS capabilities, according to the company’s announcement.
“This latest CSO award is a system-of-systems integration program supporting the Space Force Operational Test and Training Infrastructure (OTTI), where Slingshot is priming a team of subcontractors to deliver a consolidated, classified training and assessment environment,” Solms said. “The objective is to integrate ‘industry-proven’ Red/White/Blue cell tools into a single solution so Guardians can train in realistic, high-side scenarios.”
“Slingshot’s TALOS is one of the components being integrated into that solution, specifically acting as the thinking adversary / Red Cell capability. Multiple blue cell tools are also being integrated into the combined solution all deployed into a classified environment to support scenario development, management, and visualization,” he added.
Slingshot launched TALOS in July 2025, according to the announcement, to “imitate” the behavior of satellites on orbit for training and simulations, as well as to “learn and replicate real-world operations and change as the orbital environment changes.”
The secret sauce, the company says, is that it was able to train TALOS’s AI algorithms on a massive amount of real world-data: “Slingshot tracks roughly 95% of all payload-sized objects across all orbital regimes, from LEO through xGEO, and our systems can track in both day and night, 24/7/365. The resulting data set represents the largest corpus of commercially available astrometric and photometric data today,” Solms said.”
That is further multiplied by the number of derived data products we generate from those raw observations, which includes event detections, pattern of life information, photometric fingerprints, etc.,” he added.