Next upHack for Humanity: San Francisco (powered by Google Gemini)
News

Dynatrace acquires AI-observability startup Arize for $915 million

Dynatrace agreed to buy AI-observability startup Arize for $915 million to add hallucination detection and AI-agent evaluation to its monitoring platform.

D
Aug 15, 2026 · 1 min read

Dynatrace said on Aug. 13 that it agreed to acquire Arize, an AI-observability startup, for $915 million, adding tools that detect AI hallucinations and evaluate the output of AI models and agents running in production.

The purchase is a bet that monitoring AI systems — catching hallucinations and grading agent output — becomes standard infrastructure as companies push generative AI into production, the same way application-performance monitoring became routine for conventional software.

The deal comprises roughly $815 million in cash plus replacement equity awards for Arize employees joining Dynatrace, the software-intelligence company said. Dynatrace plans to fund it with cash on hand or its existing credit facility and expects the transaction to close later this quarter or early next, subject to regulatory review, according to the announcement.

Arize’s platform detects hallucinations, evaluates model and agent outputs, and monitors AI applications in production. It is used by Fortune 500 companies and AI-native developers, according to Dynatrace, and its capabilities would slot into Dynatrace’s broader observability suite for enterprise IT.

Dynatrace Chief Executive Rick McConnell said the AI-observability market opportunity is “enormous” as AI moves into production. Arize co-founder and CEO Jason Lopatecki said joining Dynatrace would let the company “take that mission much further.”

The $915 million figure and the roster of Arize customers come from Dynatrace and have not been independently verified, and the deal still requires regulatory clearance before it closes. Terms of the replacement equity awards were not disclosed.

Whether the acquisition pays off depends on how many enterprises decide that watching what their AI models and agents actually do in production is worth paying an observability vendor to handle.

More news