Dynatrace Completes Arize Acquisition, Keeping Its AI Tools Available Separately

AI builders and the teams keeping their software running now have a common owner for two different views of reliability. Dynatrace said on October 1, 2026, that it had completed its acquisition of Arize. The deal brings AI tracing and evaluation into its business, while Arize’s existing products remain available rather than immediately giving way to a combined platform.

Dynatrace first announced its intent to acquire Arize in August. Closing moves the companies from a proposed transaction into joint product planning. Dynatrace says the immediate priorities are continuity for customers and developers, alongside identifying where Arize’s capabilities can add value to its platform. It has not set an integration timetable.

A healthy service can still deliver a wrong answer

Arize focuses on the behavior of AI applications and agents. Its tracing tools record the path through a task: model calls, retrieved information, tool use and context. Engineers can inspect that path, assess performance and cost, evaluate results and compare experiments. Those workflows help teams investigate why an application behaved a particular way.

Dynatrace serves the teams responsible for the surrounding digital environment. Its monitoring covers applications, services, infrastructure, user experiences and business processes. The proposed connection is between these two perspectives: what an agent did internally, and how that activity relates to the software it touches and the outcomes users receive.

Dynatrace’s case for combining them rests on a distinction: technical health is not the same as a correct result. It says an AI system can produce a plausible but wrong response without triggering an error. An agent might finish a workflow while using stale context or updating the wrong record.

Evaluation does not end at deployment

The company also argues that checks before launch are insufficient. Changes to models, prompts, retrieval methods, tools, data or agent workflows can change behavior after deployment. Dynatrace wants evaluations and experiments to continue through operation, with lessons from failures becoming new datasets, regression tests or evaluators that check whether later changes introduce problems.

That ambition reaches beyond response quality. Dynatrace says inefficient agent paths can cause unexpected spending, while inappropriate access or actions can create security and compliance risks. Its argument is that infrastructure measurements alone cannot explain which information an agent retrieved, which tools it called or why it took a particular action.

Separate products, with a shared roadmap ahead

• Phoenix remains available as an open-source project. Its workflows include tracing, evaluation, failure investigation, dataset curation and experiment comparison.
• Arize AX remains a standalone offering, providing those AI-building workflows in a managed platform across development and production.
• The Dynatrace platform also remains available separately; the acquisition does not replace these offerings with a newly detailed unified product.

Dynatrace says it intends to preserve the workflows AI builders already value while adding context from the broader environment. It also plans to continue supporting Arize’s open, builder-first approach. Phoenix is built on OpenTelemetry and OpenInference, open-source foundations for recording and describing activity in AI applications.

Arize CEO Jason Lopatecki, chief product officer Aparna Dhinakaran and the Arize team are joining Dynatrace. The company says customers and builders will help shape integration priorities. Its closing announcement describes a product and platform roadmap still being developed, not a completed connection between all the capabilities it outlines.