PiLogic, the company building explainable artificial intelligence for complex space systems, has won a prime Phase I award under SpaceWERX’s Adaptive and Intelligent Space (AIS) Challenge, a U.S. Space Force STTR program aimed at making satellite operations more autonomous and resilient. Starcloud is subcontracting to PiLogic on the effort, hosting the demonstration on its ground-based FlatSat and, later, on-orbit aboard Starcloud-1.
PiLogic will lead development of Manifest, an onboard diagnostic capability designed to operate alongside a satellite’s existing fault-detection software, using available onboard computing resources to help operators understand what’s happening on a satellite and why, rather than working from a raw alarm feed. PiLogic is developing Manifest with Kennesaw State University as a research partner.
“As satellite constellations expand, operators may benefit from additional onboard tools to help assess satellite anomalies,” said Johannes Waldstein, CEO of PiLogic. “The operators we talk to benefit from diagnostic tools that prioritize relevant information and an onboard system intended to help assess anomalies and identify potential contributing factors. That’s the direction this whole industry has to move, and it’s the problem we built PiLogic to solve.”
The goal is a meaningful cut to both the time it takes to spot problems and the amount of manual review operators have to do. That’s where Starcloud comes in. PiLogic plans to test Manifest first on Starcloud’s ground-based FlatSat, followed by a planned on-orbit demonstration aboard Starcloud-1, supporting an end-to-end Phase I demonstration path from ground testing to a planned on-orbit demonstration.
“Starcloud-1 has demonstrated the ability to run AI workloads reliably in orbit,” said Philip Johnston, CEO of Starcloud. “Hosting Manifest on our FlatSat and, if integrated, in orbit would give PiLogic an environment to evaluate the capability and would provide additional information on the platform’s ability to support the demonstration.”
PiLogic is also subcontracting on two more Phase I awards under the same AIS Challenge: one with Starcloud, focused on validating Starcloud’s on-orbit compute and cybersecurity architecture, and one with Sophia Space, supporting mission-level analysis and a software demonstration on Sophia’s ground testbed as part of its work toward flight validation.
“Sophia Space’s architecture is built around self-sufficient, distributed compute TILEs operating together as a single orbital data center. Each TILE generates its own power, manages its own thermal environment, and coordinates compute with its neighbors,” said Rob DeMillo, CEO of Sophia Space. “At scale, models like PiLogic’s could route workloads away from a node showing signs of degradation, before it affects overall system performance. PiLogic’s work on the Sophia Orbital Operating System testbed will provide data to inform future development and flight-validation activities.”
Three different teams, three different compute environments, one underlying question: Can onboard diagnostic tools provide useful, traceable information about satellite anomalies? That’s the case for Exact AI. A model that can only tell you it’s 90% confident isn’t enough when the other 10% is the difference between a routine sensor anomaly and an acute system failure. PiLogic’s approach produces calibrated, traceable answers instead, intended to help operators assess satellite anomalies, rather than outputs that are difficult for users to interpret.
As satellite constellations scale past what any ground crew can watch by hand, traceability and calibrated outputs are important factors to consider in assessing the applicability of autonomous capabilities.
This STTR effort adds to PiLogic’s growing footprint in space applications, following a NASA SBIR Phase I award to apply its technology to spacecraft fault detection and a separate research agreement placing its technology on Air Force Research Laboratory satellite testing platforms.
PiLogic builds Exact AI — exact probabilistic reasoning for mission-critical aerospace and defense. Its models, Manifest (satellite diagnostics) and Resolve (radar tracking), compute correct, explainable answers in milliseconds, onboard, on existing hardware, with no training data and no hallucination. PiLogic works with the Air Force Research Laboratory under a two-year Cooperative Research and Development Agreement and is backed by Scout Ventures, Seraphim Space, Sovereign's Capital, Flex Capital, and FN Fund. For more information, visit pilogic.ai.