Scale AI is launching a “full-stack” generative AI platform that it says will allow large enterprise customers (big corporations, government agencies, etc.) to capture the benefits of large language models (LLMs) without having to send proprietary data out to a third party AI company.

Rather, the Scale AI solution, called Enterprise AI Platform, lets the customer host the LLM within its own walls, says Scale AI CEO Alexandr Wang.

“We integrate with all of the leading commercial and open source foundation models,” he says, while offering tools to ensure data security and safety. He points to Scale’s recent partnership with Anthropic as an example: “We basically pair that with a set of tools and user applications that enable customers to actually make use of these models in production.”

The enterprise customer is then free to use its own data to train the model, including reinforcement learning with human feedback (RLHF) and supervised fine-tuning. This customizes the LLM for use in specific enterprise applications.

Wang says some big enterprise customers, including Fox Corporation and Koch Industries, are already using the new Enterprise AI Platform. Koch, for example, is using the platform to give its traders and analysts market insights.

DEFENSE AND INTEL

Scale also announced a new decision-making hub called “Donovan” that was designed with government customers in mind. Donovan uses a large language model (running on secure government servers) to quickly assemble a wide variety of relevant operations or intelligence data in support of everything from battlefield decisions to intelligence gathering.

Donovan has already been deployed by the Army’s 18th Airborne Corps, Wang says. “They’re using it for AI–powered decision making,” Wang says, “using the LLM to ingest thousands of orders, situation reports, intelligence reports, all sorts of different data that they have sitting there, and operators can just use Donovan and ask questions.”

Scale may be scratching a major itch for intelligence and defense agencies. Both communities have collected massive amounts of data, but they’ve struggled to build the systems needed to make sense of it all. If Scale can help government agencies clean, segment, and organize their data, getting in shape to train AI models, it could play a large role in national security.