Datadog (DDOG) Q4 FY25 earnings call: guidance and Q&A

How will the rise of agentic frameworks and LLMs impact the observability category, its defensibility, and Datadog's evolution, especially regarding UI and workflows for human vs. agentic SREs?

Olivier stated that AI accelerates complexity and productivity, leading to more applications and breakage, which increases the need for observability to validate and ensure safety. He believes observability will expand to new domains. Datadog is evolving by building products for both human UIs and direct agent interaction, exemplified by the MCP server's explosive usage growth.

“The most important part of the whole AI development life cycle today.”

asked by Sanjit Singh · answered by Olivier Pomel

What drove an 8-figure deal with an AI model company, given the perception that open-source tooling is cheaper, and how does this validate Datadog's value proposition against such alternatives?

Olivier explained that the situation is similar to other customer lands: companies initially use homegrown or open-source solutions, but Datadog quickly demonstrates value by improving developer velocity and solving complex problems, making it more cost-effective than internal efforts, especially for fast-growing AI companies.

“It's cheaper to do it yourself is usually not the case.”

asked by Raimo Lenschow · answered by Olivier Pomel

How does Datadog maintain its moat against LLMs becoming better anomaly detection tools, and what is the limiting factor for LLMs taking share from observability?

Olivier highlighted Datadog's moat in assembling context for LLMs and its ability to perform proactive, in-stream analysis on massive, real-time data volumes. He argued that while LLMs are good for post-hoc analysis, they cannot handle the scale and real-time requirements for preemptive resolution, which is where Datadog's embedded data plane and specialized models excel.

“You need to be embedded into the data plane, which is what we run.”

asked by Gabriela Borges · answered by Olivier Pomel

How does Datadog ensure customers perceive value when their bill increases due to more AI usage, and what steps are taken to manage this perception?

Olivier stated that customers buy products to either make or save money, and Datadog consistently proves this value. He noted that consolidating onto Datadog's platform for new products is generally more cost-effective than integrating additional vendors or products.

“There's only 2 reasons people buy your product is to make more money or to save money.”

asked by Gabriela Borges · answered by Olivier Pomel

Could you elaborate on the level of conservatism in the FY26 guidance, particularly regarding the largest customer's growth assumption and potential concentrations within the AI cohort?

David explained that the guidance applies conservatism to growth trends. The core business (excluding the largest customer) is modeled to grow at least 20%, implying a more conservative growth assumption for the largest customer. He added that the AI cohort, with 650 customers, is highly diversified and does not show unusual concentration.

“We took a very conservative assumption there.”

asked by Ittai Kidron · answered by David Obstler

How is the rise of LLMs impacting competition and share shifts in the market, and how will Datadog be affected?

Olivier stated that there has been no significant change in the competitive landscape, with Datadog continuing to gain market share. He acknowledged 'noise' from recent M&A but noted these companies had little market impact. He emphasized that competing in observability is a demanding, innovative market, and Datadog is confident in its strategy.

“Competing in observability is a very, very full-time job.”

asked by Todd Coupland · answered by Olivier Pomel

With hyperscaler CapEx exceeding $500 billion, how much of this is training-related, and when will it convert to inferencing, creating a predictable ramp for Datadog's LLM observability revenue?

Olivier noted that the massive CapEx points to significant increases in system complexity and reach, which will greatly benefit Datadog's business. He cautioned that directly mapping CapEx to future revenue is difficult but expects it to drive more applications and intelligence, ultimately helping Datadog.

“It definitely points to very, very, very large increases in the complexity of the systems, the number of systems and the reach of the systems in the economy.”

asked by Mark Murphy · answered by Olivier Pomel

A year ago, you discussed in-sourcing observability as a cultural choice, not rational for most. Has this view changed given recent industry advancements?

Olivier reiterated that his view has not changed. While some customers may in-source for cultural reasons, it rarely makes economic or strategic sense for the vast majority of companies. He noted that even teams at hyperscalers, with extensive resources, choose Datadog for a more direct path to problem-solving.

“Economically or from a focus perspective, it doesn't make sense for the very vast majority of companies.”

asked by Koji Ikeda · answered by Olivier Pomel

How does Datadog capture fair value for the productivity Bits AI brings, and how will it compete against startups and general-purpose AI platforms in this space?

Olivier explained that Bits AI's value is in significant cost savings from reducing incident resolution time and internal involvement. He emphasized that Datadog's long-term advantage lies in proactive, in-stream analysis of massive, real-time data, which is much harder for general-purpose AI platforms to achieve compared to post-hoc analysis.

“I think that's what is going to be our advantage and where it's going to be much harder for others to compete, especially general purpose AI platforms.”

asked by Peter Weed · answered by Olivier Pomel