
Dynatrace Closes $915M Arize Deal—Open-Source Projects Stay Supported

Dynatrace completed its acquisition of Arize on October 1, 2026, closing a deal announced at $915M, according to The Inference’s account of the close. Dynatrace now owns a business focused on tracing and evaluating AI applications and agents. For customers and developers, the immediate question is which products they can keep using while the companies work toward a connected offering.
In Steve Tack’s closing note, Dynatrace’s product chief wrote, “If you’re building with Phoenix, keep building,” and specified that Phoenix remains open source, OpenInference remains open and OpenTelemetry-compatible, and Arize AX and Dynatrace remain available as standalone offerings. The acquisition puts their teams under one owner; the connections between their products are still being planned.
The agreed price and the business Dynatrace acquired
The August 13 investor announcement valued the transaction at $915 million, subject to customary adjustments, comprising approximately $815 million in cash plus replacement equity awards for Arize employees joining Dynatrace. The announced funding plan allowed for cash on hand, Dynatrace’s existing credit facility, or both. The $915 million figure describes the agreed transaction value, rather than a separately disclosed final cash payment.
Arize co-founders Jason Lopatecki and Aparna Dhinakaran have joined Dynatrace. Lopatecki continues to lead the Arize team and reports to Dynatrace chief executive Rick McConnell; Dhinakaran joined from her role as Arize’s chief product officer. Their move gives the combined business leadership familiar with Arize’s products, while the current product commitments give existing users continuity during integration planning.
How the AI and production workflows could connect
Arize’s tools examine what an AI application does. A trace can expose an agent’s trajectory through model calls, retrieved context and tool use, while evaluations and experiments help engineers assess output quality and compare changes. Phoenix supplies open-source workflows for tracing, investigating failures, curating datasets and running evaluations. Arize AX offers managed workflows across development and production.
Dynatrace observes the environment in which those applications run: services, infrastructure, user experiences and business processes. Those signals answer different questions from an AI evaluation. An application may respond quickly while its agent selects the wrong tool or gives an inaccurate answer; an agent failure may also coincide with a service or infrastructure problem. Connecting the two views could help AI engineers and SRE teams trace an affected user interaction through both the agent’s behavior and the systems supporting it.
Consider a hypothetical deployment after which an agent starts retrieving stale context. Its trace could identify the retrieval step, and an evaluation could expose the poorer answer. Production telemetry could then show whether the service supplying that context changed or slowed at the same time. A connected investigation of this kind is the aim of the integration work, rather than a combined feature described as available at closing.
The proposed scope reaches beyond traces alone. Evaluations, datasets and experiments could gain context from application performance, infrastructure health and user outcomes, allowing an issue found in production to inform the next development cycle. The benefit depends on how those records are linked: a quality score by itself does not explain which service supplied an agent’s input, and healthy infrastructure by itself does not establish that an answer was useful.
What continuity means for Phoenix and OpenInference
Phoenix remains available as an open-source project. Its role extends from collecting and inspecting AI traces to evaluating behavior and comparing experiments, so continued support matters to teams using it throughout development. Arize AX remains a separate managed offering, and the Dynatrace platform remains available on its own. Ownership has changed, but those offerings have not been presented as a single released product.
OpenInference addresses a different layer of the workflow: instrumentation and semantic conventions for AI observability. Phoenix is built on OpenInference and OpenTelemetry. OpenTelemetry has accepted a code grant of OpenInference GenAI instrumentation, which is being incorporated incrementally into its GenAI instrumentation work, while OpenInference continues as an open, compatible project. That relationship matters because teams can instrument AI behavior through open conventions even as product interfaces evolve.
The continuity watchlist is therefore specific. Phoenix users can watch whether tracing, evaluation and experiment workflows keep advancing as open-source work; OpenInference users can watch its ongoing compatibility with OpenTelemetry and the progress of the donated instrumentation. Teams considering a future Dynatrace connection will also need to see how existing traces and evaluations participate without losing the workflows they already use. Continued support is the commitment at closing; the details of those connections belong to the forthcoming roadmap.
The next integration milestone
With the acquisition complete, the product teams can decide which links between Arize’s AI workflows and Dynatrace’s operating context deserve priority. That choice has practical consequences: connecting a trace to application telemetry serves incident investigation, while connecting evaluations and datasets to production outcomes serves the improvement cycle. Neither path follows automatically from the change in ownership, and a release order for the planned connections has yet to be set out.
The next substantive milestone is a product roadmap that names the first integrations and explains how Phoenix, OpenInference, Arize AX and existing Dynatrace deployments will take part. Until those details arrive, users have a clear current position: the acquisition is complete, the named offerings remain available, and the open projects retain the support described at closing.
Also read:
Related articles

Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.