Bernstein comments on Datadog (DDOG.US) Q3 earnings report: Strong performance but “not surprising enough”, downgraded AI business revenue expectations

The Zhitong Finance App learned that Bernstein released a report indicating that Datadog (DDOG.US)'s revenue performance in the third quarter of fiscal year 2026 was strong, but investors were disappointed by the increase in the company's full-year guidance. At the same time, changes in the usage of the largest AI native customer (the research report estimates OpenAI) have added subtle variables to growth prospects. Based on this, Bernstein lowered the 2027-2028 ARR forecast for the AI business by about 25% and lowered Datadog's target price from $350 to $273, but maintained an “outperforming market” rating.

The following is a summary of the core contents of the research report:

Strong Q3 results, but the increase in guidance was “not surprising enough”

Datadog's third-quarter revenue was US$1,197 million, exceeding expectations by US$57.5 million. This was the largest margin in the company's history (previously Q1 and Q2 were 50 million and 47 million, respectively, both historical highs at the time), and the first time since 2022 that it exceeded expectations by more than 5% for two consecutive quarters.

However, Bernstein believes that the problem lies in the guidelines: the company only “flipped” the portion that exceeded expectations into the full-year guidance, and did not amplify the increase in the guidelines to 2-3 times beyond expectations as before. If the company can once again exceed expectations of 50 million to 60 million US dollars in Q4, its Q4 year-on-year growth rate will only be about 30% to 31%, lower than the 35% that was just achieved in Q3.

AI customers present a “subtle differentiation”: the largest customer usage is declining, and other AI-native customers are still expanding

The research report points out that Datadog's AI-related business has split. According to the company's disclosure, non-AI revenue increased by more than 20% in Q3, and was about 1 percentage point faster than in Q2; however, revenue growth for native AI customers slowed in Q3, mainly due to a decrease in usage by its largest AI native customer (Bernstein presumably OpenAI). Bernstein's model shows that the largest AI native customer's usage on Datadog began to decline in Q3, and it is expected that Q4 will remain sluggish and reset to a lower operating rate.

However, Datadog is still signing million-dollar contracts with other large AI-native customers, whose usage expansion is still strong. Bernstein's AI-native customer revenue model previously overestimated Q3 revenue by about $34 million, of which about $20 million was due to the time difference — that is, the intensity of demand in Q3 may be reflected in Q4 revenue; another $14 million was due to the “ratio of token consumption to revenue” set too high in the model. The updated model has taken into account a quarter's lag effect.

Management communication: largest customer guidelines tend to be conservative, and early contract renewals bring discount headwinds

At the post-earnings conference call, Datadog management's Q4 guidelines for the largest customer were based entirely on the minimum contract usage and did not reflect the customer's recent usage trend. Management stressed that this treatment should not be interpreted as a sign of a further slowdown in consumption. The company's guidance for 2027 will also continue to eliminate the upward assumption of this largest customer.

Additionally, some rapidly growing customers (including AI native customers) renewed their contracts early after spending more than promised to lock in higher discounts. This led to a decline in Datadog's short-term unit revenue, creating a discount headwind of around $15 million against the Q4 guidance. Management said it usually takes about a quarter for consumption to return to pre-discount revenue levels, but this mainly reflects strong consumption and greater commitment rather than weakening demand.

Based on the AI revenue revisions mentioned above, Bernstein lowered the AI business's ARR forecast for 2027-2028 by about 25%, but kept the rest of the model unchanged. At the same time, the research report pointed out that despite the short-term headwinds of fluctuating AI native customer usage and renewal discounts, Datadog's core growth trend is still healthy, non-AI business growth is steady, and early Q4 signals indicate that demand may be improving, laying the groundwork for potential growth in Q1 2027.