Datadog acquires Adaptive ML to enhance AI operations

Datadog just bought itself an AI research lab. The cloud observability company announced the acquisition of Adaptive ML, a startup specializing in reinforcement learning operations that will fold directly into Datadog’s AI research division.

The deal, announced June 30, comes with no disclosed price tag. What it does come with is a team of engineers from some of AI’s most recognizable names and a platform that has already processed trillions of tokens in enterprise deployments.

What Adaptive ML actually does

The startup’s core product, called the Adaptive Engine, uses reinforcement learning to continuously optimize specialized large language models and AI agents after they’re deployed. The platform handles synthetic data generation, specialized AI judges that evaluate model outputs, A/B testing for different model configurations, and feedback loops designed for real-time applications. Enterprise customers have used it for search optimization and customer support, among other use cases.

Adaptive ML was co-founded by CEO Julien Launay, who previously worked at LightOn and Hugging Face. The startup raised $20M in a seed funding round led by Index Ventures.

Datadog’s AI acquisition spree

This isn’t Datadog shopping on impulse. The Adaptive ML pickup is the third AI-adjacent acquisition the company has made in roughly 14 months.

Datadog acquired Metaplane in April 2025, adding data observability capabilities. Then came Eppo in May 2025, which brought experimentation and feature flagging tools into the fold. Adaptive ML rounds out a trilogy of deals that, taken together, paint a clear picture: Datadog is building a full-stack platform for companies that need to not just monitor their AI systems, but actively improve them.

Why this matters for investors

Datadog’s decision to position Adaptive ML as its AI research lab rather than just absorbing the team into an existing product group suggests awareness of the integration challenge that comes with acquiring startups.

Investors should also note what the undisclosed deal terms might mean. For a startup that raised $20M in seed funding and had already demonstrated enterprise-scale deployments processing trillions of tokens, the acquisition price was likely meaningful enough that Datadog chose not to publicize it. Whether that number shows up in future earnings filings will be worth tracking.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.