Datadog Inc (NASDAQ:DDOG) has acquired Adaptive ML, a frontier AI startup, and will add it to Datadog AI Research to support its work on world models and agentic LLM post-training for observability.
Adaptive ML develops a Reinforcement Learning Operations, or RLOps, platform designed to help enterprises build, own and deploy specialised agents and models.
Datadog said the acquisition will accelerate its investment and research efforts. Datadog AI Research focuses on fundamental technical problems and works with the company’s product and engineering teams to turn research into products.
Julien Launay, co-founder and chief executive of Adaptive ML, said the aim was to give enterprises the ability to continually improve their AI. He said production scale was the hardest part, and that Datadog’s access to real-world infrastructure would help accelerate continuous intelligence.
Ameet Talwalkar, Datadog’s chief scientist, said the lab was focused on using the company’s data and domain expertise to build specialised agents and models and turn data into first-party intelligence. He said bringing Adaptive ML on board was a natural fit as Datadog continued to strengthen its R&D efforts.
Datadog said it has invested over $1bn in research and development annually. It said this has supported its observability and security products, including Toto 2.0, Bits Investigation, Bits Code and Bits Security Analyst, which have already carried out hundreds of thousands of investigations for customers.