The acquisition adds code-level performance intelligence for AI and cloud-native workloads, and brings a world-class database engineering team that will help power Dash0's next-generation backend
NEW YORK & BERLIN, August 17, 2026 Dash0, the OpenTelemetry-native agentic observability platform built on the promise of letting customers Ship More and Break Less, today announced it has acquired Polar Signals, the company widely regarded as the best continuous profiler on the market, and unique production-safe GPU profiler for CUDA-based NVIDIA architectures across inference and training workloads. Financial terms of the transaction were not disclosed.
The acquisition brings continuous profiling natively into SignalStore, giving Dash0 customers a new, deeper layer of visibility, down to the line of code, across CPU, memory, and GPU, and now connected directly to the agents that build and run that code. Adding this new context layer to Agent0 will enable autonomous optimization use cases to improve performance and reduce costs without manual effort.
Why Polar Signals
Profiling has quietly become one of the highest-leverage signals in modern infrastructure. It shows exactly where CPU, memory, and now GPU cycles are actually spent, at the level of an individual function or kernel. Polar Signals built the deepest implementation of this in the industry and is uniquely able to continuously profile CUDA workloads on NVIDIA GPUs in production, with minimal overhead, for both training and inference. As more of the software Dash0 customers run is written and operated by AI agents, that level of ground-truth insight into what code is actually doing becomes essential, not optional.
Profiling has historically been an expert's tool: powerful, but so complex to run and interpret that only a small group of performance specialists ever touched it. AI changes that equation. By putting profiling data directly into the agentic development loop via MCP, any developer and any coding agent can now automatically draw on that same expert-level insight as a routine part of building and shipping software. This is a core part of what Ship More. Break Less. means in practice: giving every team, not just performance specialists, the tools to move fast without flying blind.
What this means for customers
With profiling in SignalStore, Dash0 unlocks several new capabilities:
- Code-level insight, for developers and agents alike. Developers get line-of-code performance visibility across their entire pipeline. Through MCP, coding agents get the same data, grounding their suggestions in real profiling evidence rather than guesswork.
- Continuous, automated optimization. Agent0, along with a new capability, AutoTune, will continuously analyze customer profiles and proactively create pull requests to optimize CPU and memory usage, helping teams cut infrastructure spend and improve performance without manual profiling.
- Smarter data collection with SignalControl. SignalControl automatically manages profiling granularity based on system conditions. When an anomaly occurs, it automatically steps up to capture all relevant profiling data, so teams have the depth they need to diagnose an issue without paying the overhead of that depth all the time.
A team and a backend, not just a product
Beyond the product, the acquisition brings Dash0 a team of world-class database engineers. Their work on “Great Lakes”, Polar Signals' purpose-built storage engine for profiling data, will become the foundation for replacing ClickHouse in SignalStore, giving Dash0 a more efficient, more scalable backend, and opening the door to deployment and security architectures for customers that weren't previously possible. Great Lakes is the successor to FrostDB and was built on the lessons and technological advancements of the AI era.
Polar Signals is simply the best team in the world at this problem. Software is increasingly written and run by agents, and agents need ground truth about what code is actually doing, not just traces and metrics, but the exact line, the exact kernel, the exact cost. Profiling used to be something only a handful of experts could use. Now it's something every team and every agent can use every day. Bringing profiling natively into SignalStore and bringing this team in to help rebuild our backend moves Dash0 a full step ahead. Welcome, PolarBears, let's build something great together.Mirko Novakovic, CEO and Co-Founder, Dash0
Joining Dash0 gives us the platform and reach to bring continuous profiling, including what we've built for GPU and CUDA workloads, to a much broader set of teams, and to connect it directly into how AI agents build and operate software.Frederic Branczyk, Founder and CEO, Polar Signals
About Dash0
Dash0 closes the loop from code to production in the AI era. Built natively on OpenTelemetry, Dash0 tracks what AI builds, observes everything running in production, and fixes problems autonomously - all powered by a single, open data foundation. Dash0 helps engineering teams ship more and break less in a world where code moves at machine speed.
Founded in Germany in 2023 and headquartered in New York, Dash0 serves 750+ customers, including Zalando, Taco Bell, and The Telegraph.
The company raised $110M in Series B funding at a $1B valuation, with participation from Balderton Capital, Accel, Cherry Ventures, DTCP, DIG Ventures, July.fund, and T.Capital. Learn more at dash0.com.
About Polar Signals
Founded in 2020 in Berlin by Frederic Branczyk, Polar Signals builds Polar Signals Cloud, an always-on, zero-instrumentation continuous profiler for CPU, memory, and GPU workloads, and is the company behind the open-source Parca project. Polar Signals was built by engineers previously of Red Hat, CoreOS, and the Prometheus and Kubernetes maintainer communities.
Media Contact
Becki Howland
