A profile of data labeler Scale AI, which has a $249M contract with the US DoD and was the first AI company to have an LLM deployed on a classified network
the tech behind chatbots such as ChatGPT — deployed on a classified network after it signed a deal with the Army's XVIII Airborne Corps.” https://www.washingtonpost.com/ ... Charles Rollet / @charlesrollet1 : One of the largest US AI defense startups, @scale_AI, was partly inspired by its founder visiting a Chinese facial recognition firm & being ‘terrified’ https://www.washingtonpost.com/ ... via @washingtonpost [image] Elizabeth Dwoskin / @lizzadwoskin : Years ago Googlers protested the company's military contracts. Now a generation of AI companies faces the same questions, via @GerritD https://www.washingtonpost.com/ ... Hussein Kanji / @hkanji : “Defense is a customer like no other, and it requires deep institutional expertise to succeed. Selling to the Army isn't like selling to Uber.” https://www.washingtonpost.com/ ...
Context & Ripple Effects
This marks Scale AI’s move from data-labeling work into defense AI deployment, with classified-network access making the company a more consequential military technology vendor. The subsequent Pentagon contract to test and evaluate LLMs for planning and decision-making shows that evaluation, not only model access, became part of its defense role.
The arc continued toward broader data-preparation and model-testing work under a $100M DoD ceiling contract, before a later $500M DoD award for data analysis and decision support. That progression matters because it ties Scale’s commercial position to the operational adoption path for military AI.
First-order effects
- Scale AI gains a sizable defense customer and a classified deployment reference point, while the Army’s XVIII Airborne Corps gets access to an LLM in a restricted environment.
- The deployment turns the immediate work from supplying training data into operating and supporting AI capabilities under military security constraints.
Second-order effects
- Other defense-AI suppliers face pressure to demonstrate comparable security clearance, deployment reliability, and evaluation tooling—not merely model performance.
- Classified access raises the importance of data preparation, testing, and governance services around models, aligning with Scale’s later Pentagon evaluation and end-to-end services contracts.
Third-order effects
- If such deployments continue, defense procurement could favor a smaller group of AI vendors able to pair models with secure integration, validation, and operational support.
- Military use may become a durable source of commercial legitimacy for AI firms, while making government procurement decisions increasingly important to the sector’s structure.
The trend: This is an early instance of AI vendors evolving into state-compatible infrastructure providers, selling secure deployment and assurance alongside models and data services.