Elodie Collins 
Astris AI, a Lockheed Martin subsidiary, has introduced Astris AI for Government, a new offering that promotes trustworthy and secure artificial intelligence adoption to support high-assurance missions.
Astris AI for Government provides a unified platform that combines Oracle Cloud Infrastructure’s U.S. network for unclassified and classified regions, Meta’s open source AI models; NVIDIA’s AI Enterprise software; and Astris AI’s AI Factory machine learning operations, or MLOps, and generative AI, Lockheed Martin said Thursday.
“By integrating the best of American technology, the Astris AI for Government full-stack solution eliminates fragmentation and procurement barriers that slow adoption for secure, mission-ready AI—without vendor lock-in or complex integration challenges,” stated Sarah Hiza, senior vice president of technology and strategic innovation at Lockheed Martin.
Astris AI was established in December 2024 to provide MLOps to the U.S. defense industrial sector.
Learn more about how agencies are using or plan to use AI, ML and automation at the Potomac Officers Club’s 2026 Artificial Intelligence Summit on March 19. This critical GovCon networking event will bring together government and industry leaders to provide insights into how AI is transforming the world today. Get your tickets here.
What Is Astris AI for Government?
According to Lockheed Martin, Astris AI for Government will work to enable agencies to continuously ingest and process data on the field and deliver insights to operators and analysts.
One potential application for the platform is to support the Genesis Mission, a government initiative to bolster the use of AI for scientific discovery.
The company also said the Department of the Interior can deploy the platform for faster and more informed decision making to minimize property destruction and loss of life during wildfires.
What Comes Next for Astris AI?
Astris AI is expanding its partner network and advancing development of Persistent Environment for AI, Readiness and Learning, or PEARL, a large-scale AI and simulation environment intended to drive operations analysis, domain-specific models, AI test and evaluation, and simulation-to-real integration.
