Deep|DDOG: BitsAI, A Hidden AI Agent Play?

New Growth Driver could Drive Revenue Upside and Re-rating

Datadog’s share price has corrected sharply since late 2024 on fear of its largest customer OpenAI cutting off from them, and has fully recovered since bottoming in April. The strong rebound was partially driven by the stabilized relationship with OpenAI, tailwind from AI agent adoption, as well as the unexpected S&P500 index inclusion. We sense that market has been desperate to find another AI winner in infra software space besides SNOW, and DDOG has become a ideal candidate as AI agents bring create new demand for observability. However, today we would like to discuss another overlooked AI tailwind to DDOG: its new Bits AI agent offering. Our research shows that, despite early, Bits AI could expand a whole new TAM for DDOG, potentially boosting its revenue growth as well as multiple re-rating.

Datadog and OpenAI

Before diving into our formal analysis, let's discuss Datadog and OpenAI.

We previously noted in our Snowflake report that OpenAI has been implementing a Grafana/Clickhouse + Clickhouse/SNOW/Databricks architecture to replace Datadog. By late last year, they were pushing for a complete transition, but this triggered significant outage. Consequently, the replacement pace has slowed, though the migration efforts continue. This makes Guggenheim's research report from yesterday particularly valuable. From an investment perspective, once the OpenAI risk factor is fully priced in, Datadog presents a more compelling opportunity.

What’s Bits AI?

Bits AI was initially previewed during DASH 2023 in Aug as an AI assistant sitting on DDOG’s observability + security platform, which went GA in mid 2024. The product then was more of a copilot chatbot that learns from customers’ telemetry to answer questions and recommend fixes. During DASH 2025 in June, the product was upgraded to a family of agents, including Bits AI SRE, Bits AI Dev Agent and Bits AI Security Analyst and went Limited Availability (LA).

Bits AI SRE is currently the most advanced among all, given observability is DDOG’s bread and butter. To put it in context, there are about 1m SRE/On-call Ops globally, 20% of which are based in North America. According to Glassdoor, Junior SREs in North America are paid on average US$135k per year, with senior ones at US$166k. Therefore, the global SRE market is at least over US$100bn per annum.

Bits AI SRE already seeing meaningful efficiency gain

We have talked to multiple selected DDOG customer that had early access to Bits AI product, and are surprised with their universally positive feedback. Note that these customers started POC in 2024 or early 2025, it’s fair to assume the products are less sophisticated then vs now. When a system alert triggers, SREs need to analyze all the logs to identify root causes, then start to fix bugs accordingly. Bits AI SRE is now very helpful in auto-hypothesizing root cause before the pager even fires. Customers during POC are seeing at least 20-30% time saved in identifying root causes, with certain cases could be pinpoint as much as 95% faster. In fixing bugs, Bits AI capability varies depending on customer’s infra complexity and data readiness. In the long run, it’s likely Bits AI could replace some SRE (at least the juniors), similar to how coding agents are pressuring the programmer job market now. In fact, one most aggressive expert admits they already moving half of SRE team to other positions due to Bits AI.

Top UK bank expert

• Started POC in early mid 2024 and purchased in late 2024

• Most helpful in pre-emptive alerting and root cause analysis

• A Bits AI Agent could handle the workload of 3-6 human SRE at 30% of cost

• Previously running an SRE team of 26-27 persons at about US$300k per year, which is more than halved to 10-11 currently. Expect to reduce to 2-3 at the end

• Not planning to lay off the SREs given lack of talent that could understand AI internally

• Also trying out DevOps and SecOps agents

US Fortune 500 hardware company expert