Dynatrace says AI customers consume its platform at twice the rate

Dynatrace▼−0.40% — Dynatrace, down 0.40 percent today, whose observability software watches how applications and cloud infrastructure behave, has put the cleanest AI-monetization number in enterprise software on the record. Chief financial officer James Benson told BMO's technology conference on November 12, 2026↗, that customers running large language models or agentic workflows on the platform consume it at twice the rate of everyone else ¹; chief executive Rick McConnell said the same, more than twice, on the November 3 earnings call ². The claim deserves to be taken seriously because it names a mechanism. It has not yet reached the profit line.

Observability is one of the few software lines where AI raises the bill instead of compressing it: every agent step emits telemetry, and someone has to store and check it.

Agents are chatty, and telemetry is the bill

Agents are chatty. Every model call, tool use and autonomous step emits traces and logs, and Dynatrace, which charges by consumption, is the one that stores and checks them. Its own research puts the rise in log volume from AI workloads at 93% in a year↗ ².

The revenue line follows the telemetry: logs consumption is approaching $250 million a year ¹. Three quarters ago it was $100 million ¹. More than 1,100 customers already run AI workloads on the platform, and since October 1 the company has charged for its own agents as well ². For those customers, AI sprawl now has a line item.

Usage is outrunning bookings, and the proof is back-loaded

On that same earnings call, platform consumption grew more than 20% while net new annual recurring revenue was flat from a year earlier, as large deals slipped into the second half ². The stock fell almost 15% that day↗ ⁴.

Benson calls the slippage timing rather than demand, and the calendar is why he can: seven of ten annual renewals fall in the fiscal second half, most of them in the fourth quarter ¹. He expects AI-related consumption to rise significantly in calendar 2027, and says revenue growth lags consumption ¹.

One caution sits inside the headline number: the twice-rate cohort is self-selected, customers who chose Dynatrace for AI work in the first place ¹.

At 80 times earnings, shareholders own the lag

Dynatrace still trades near 80 times its trailing earnings↗; the S&P 500 sits at 22 ³. The net margin behind those earnings is 7.4% ³. Datadog▼−0.60% — Datadog, down 0.60 percent today, its larger observability rival, grew 35% in the same quarter ⁵, so the demand is not in question; Dynatrace's conversion of it is.

The tests are dated. The renewal wave lands in the fiscal quarter ending in March, and Trefis marks the February results, with net margin the number to watch ³. The usage half of the claim is already visible in consumption data and on customers' bills. The profit half is a conversion the company has scheduled, mostly, for one quarter, at a price that has already paid for it.