Datadog AI Research Launches New Open-Weights AI Foundation Model and Observability Benchmark

Rhea-AI Summary

Datadog (NASDAQ: DDOG) has launched two significant AI research initiatives through its Datadog AI Research division. The first is Toto, an open-source time series foundation model (TSFM) trained on Datadog's internal telemetry metrics, designed specifically for observability. Toto offers zero-shot forecasting capabilities for anomaly detection and capacity planning, outperforming existing TSFMs.

The second initiative is BOOM, which becomes the largest public benchmark of observability metrics, featuring 350 million observations across 2,807 real-world multivariate series. Both projects are immediately available under a permissive license, aiming to advance observability forecasting capabilities in the broader research community.

Chief Scientist Ameet Talwalkar emphasized that these launches mark the beginning of continuous AI project releases, with plans to collaborate with applied AI teams to develop practical customer solutions.

Positive

• Launch of Toto, an innovative open-source foundation model achieving state-of-the-art performance in time series forecasting
• Introduction of BOOM, the largest public benchmark for observability metrics with 350 million observations
• Free availability and permissive licensing of both tools for the research community
• Strategic collaboration between research and product teams to develop practical customer solutions

Negative

• None.

Insights

AI and Machine Learning Research Scientist

Datadog's new open-source Toto model and BOOM benchmark represent significant AI advancements in time series forecasting for observability.

Datadog has made a strategic move by launching Datadog AI Research with two significant contributions to the AI community. The first release, Toto, is particularly notable as it represents the first open-weights foundation model specifically designed for observability data. This positions Datadog at the forefront of specialized AI for time series analysis.

What makes Toto technically impressive is that it achieves state-of-the-art performance compared to existing Time Series Foundation Models (TSFMs), despite being trained exclusively on Datadog's internal telemetry metrics. The zero-shot forecasting capability is a technical breakthrough that solves a critical scaling challenge in cloud monitoring—the ability to detect anomalies and plan capacity instantly across billions of ephemeral time series without per-series tuning.

The second release, BOOM, establishes what is now the largest public benchmark of observability metrics with 350 million observations across 2,807 real-world multivariate series. This benchmark will significantly accelerate research in this domain by providing a standardized testing ground that captures the unique characteristics of production telemetry data—scale, sparsity, spikes, and cold-start issues.

From a research perspective, Datadog's decision to open-source these models under a permissive license demonstrates a commitment to collaborative innovation. This approach not only benefits the broader technical community but also positions Datadog to harness collective improvements to their technology foundation. The involvement of Chief Scientist Ameet Talwalkar suggests Datadog is building serious research capabilities while maintaining focus on translating theoretical advances into practical customer solutions.

Toto-an open-weights, zero-shot, time series foundation model-and BOOM, the largest public benchmark of observability metrics, are the first launches from Datadog AI Research

New York, New York--(Newsfile Corp. - May 21, 2025) -  Datadog, Inc. (NASDAQ: DDOG), the monitoring and security platform for cloud applications, today unveiled the first two launches from Datadog AI Research, which is tackling cutting-edge research challenges that are firmly rooted in real-world problems within cloud observability and security. Datadog AI Research is actively contributing to the broader research community by publishing findings and open-sourcing model artifacts.

The initial releases from Datadog AI Research are Toto and BOOM. Toto is the first open-source foundation model focused on observability. Time series foundation models (TSFMs) are to time series what LLMs are to language. A type of AI model trained on massive datasets that can be adapted to a wide range of downstream tasks, foundation models learn general patterns and can be fine-tuned for various applications.

Toto is an open-weights model that is trained with observability data sourced exclusively from Datadog's own internal telemetry metrics, which achieves state-of-the-art performance by a wide margin compared to all other existing TSFMs. Its zero-shot forecasting will enable instant anomaly detection and capacity planning with no per-series tuning, which is critical when monitoring billions of ephemeral time series. While existing TSFMs struggle with telemetry data, Toto heightens performance-not only for observability data but for time series forecasting more broadly-and is freely available.

BOOM introduces a time series benchmark that focuses specifically on observability metrics, which contain their own challenging and unique characteristics compared to other typical time series. It instantly becomes the largest public benchmark of observability metrics, providing 350 million observations across 2,807 real-world multivariate series to capture the scale, sparsity, spikes and cold-start issues unique to production telemetry. BOOM is an actively maintained resource for the community and will allow researchers to advance their forecasting models.

"Today marks the launch of our first open-source foundation model and we expect to continuously release AI projects through Datadog AI Research," said Ameet Talwalkar, Chief Scientist at Datadog. "The lab offers an exciting opportunity to develop research ideas and build prototypes that will contribute to the community. We will also collaborate with applied AI teams to build tools that will solve customer problems and transform how engineers work."

Collaboration between Datadog AI Research and Datadog's product and engineering teams will help translate research advances, like Toto and BOOM, into tangible benefits for Datadog customers.

Toto and BOOM are immediately downloadable under a permissive license and Datadog invites the research and the OSS communities to push observability forecasting forward with these open-source projects.

To learn more about Datadog AI Research, please visit: https://www.datadoghq.com/blog/ai/toto-boom-unleashed/.

About Datadog

Datadog is the observability and security platform for cloud applications. Our SaaS platform integrates and automates infrastructure monitoring, application performance monitoring, log management, user experience monitoring, cloud security and many other capabilities to provide unified, real-time observability and security for our customers' entire technology stack. Datadog is used by organizations of all sizes and across a wide range of industries to enable digital transformation and cloud migration, drive collaboration among development, operations, security and business teams, accelerate time to market for applications, reduce time to problem resolution, secure applications and infrastructure, understand user behavior and track key business metrics.

Forward-Looking Statements

This press release may include certain "forward-looking statements" within the meaning of Section 27A of the Securities Act of 1933, as amended, or the Securities Act, and Section 21E of the Securities Exchange Act of 1934, as amended including statements on the benefits of new products and features. These forward-looking statements reflect our current views about our plans, intentions, expectations, strategies and prospects, which are based on the information currently available to us and on assumptions we have made. Actual results may differ materially from those described in the forward-looking statements and are subject to a variety of assumptions, uncertainties, risks and factors that are beyond our control, including those risks detailed under the caption "Risk Factors" and elsewhere in our Securities and Exchange Commission filings and reports, including the Quarterly Report on Form 10-K filed with the Securities and Exchange Commission on May 6, 2025, as well as future filings and reports by us. Except as required by law, we undertake no duty or obligation to update any forward-looking statements contained in this release as a result of new information, future events, changes in expectations or otherwise.

Contact
Dan Haggerty
press@datadoghq.com

To view the source version of this press release, please visit https://www.newsfilecorp.com/release/252742

FAQ

What is Datadog's Toto AI model and how does it work?

Toto is an open-weights time series foundation model (TSFM) developed by Datadog that enables zero-shot forecasting for anomaly detection and capacity planning. It's trained on Datadog's internal telemetry metrics and outperforms existing TSFMs in observability data analysis.

What is BOOM in Datadog's new AI research launch?

BOOM is the largest public benchmark of observability metrics, containing 350 million observations across 2,807 real-world multivariate series. It helps researchers advance their forecasting models by providing real-world telemetry data.

How can developers access Datadog's new AI tools?

Both Toto and BOOM are immediately downloadable under a permissive license, allowing developers and researchers to freely access and use these tools for observability forecasting.

What are the key benefits of DDOG's new AI research initiatives?

The initiatives provide state-of-the-art performance in time series forecasting, enable instant anomaly detection without per-series tuning, and offer the largest public benchmark for observability metrics, advancing both research and practical applications.