A New York City software startup valued at $1 billion has bought a Berlin engineering team specializing in a niche but increasingly critical corner of tech: figuring out exactly which lines of code are slowing down or draining money from a company's computer systems. Dash0, founded in 2023, announced it has acquired Polar Signals, a Berlin-based startup that specializes in continuous profiling, with the goal of feeding that data directly into an AI agent designed to write and apply its own code fixes while software is still running.
Dash0 develops what's known as an observability platform, software that monitors IT systems, identifies errors, and analyzes application performance for engineering teams. As reported by Business Insider, the acquisition of Polar Signals and its engineering team will let Dash0 trace the causes of slow systems down to individual functions and lines of code. Dash0 CEO Mirko Novakovic, who started the company with former colleagues from his previous venture, said software is increasingly being written and operated by agents rather than humans alone.
What Continuous Profiling Actually Does
Polar Signals was founded in 2020 by Frederic Branczyk, a former Red Hat senior principal engineer and Kubernetes tech lead, who built the open-source continuous profiling tool Parca after being inspired by Google's internal performance-profiling practices, according to Grafana. Continuous profiling shows where code consumes computing power and memory, and it identifies where applications become slow or expensive to run. The technology relies on Linux eBPF, which samples stack traces directly from the kernel without modifying application code or requiring developers to install software development kits, incurring less than 1% CPU overhead, per an InfoQ report.
That kernel-level visibility is exactly what Dash0 says its AI agents need. Agentic capabilities are meant to enable AI systems to write, monitor, and improve software while it is running, but the company argues those systems first need precise information about what code does down to individual lines, the kernel, and the costs it generates. Under the deal, Dash0's Agent0 will receive Polar Signals' profiling data, and a planned feature called Autotune is designed to continuously search for inefficient code, generate suggestions to reduce CPU or memory usage, and prepare pull requests for developers to review.
Merging the Storage Engines
Polar Signals had also built a storage engine called Great Lakes, designed to handle large volumes of profiling data, and Dash0 plans to make Great Lakes the foundation of its own SignalStore, replacing the ClickHouse database it previously relied on. Ahead of the deal, Polar Signals' development team had already voted in August 2024 to merge its open-source Parca-Agent project into the OpenTelemetry-eBPF-Profiler initiative, according to a Polar Signals blog post, aligning its technology with the same open standards Dash0's platform is built on.
That alignment matters because OpenTelemetry has grown into the second-largest Cloud Native Computing Foundation open-source project behind Kubernetes, with production adoption reaching over 40% among cloud-native organizations trying to avoid getting locked into a single vendor, per research from Onfire AI. Legacy observability incumbents have historically relied on proprietary agents that tie customers to one platform, and Dash0's pitch leans on the opposite approach.
Backers and a Familiar Playbook
Dash0's $110 million Series B round, which closed in March 2026, valued the company at $1 billion and brought its total capital raised to $155 million across three rounds, following a $35 million Series A in October 2025 and a $9.5 million seed round in November 2024, according to Balderton Capital, one of the startup's investors alongside Accel, Cherry Ventures, DTCP, Dig Ventures, July Fund, and T.Capital. Dash0 says it now counts more than 750 customers, including Taco Bell and The Telegraph, which is owned by Axel Springer. In its first nine months of commercial availability, the company had already signed more than 600 paying enterprise clients, including e-commerce platform Zalando, per a report from SiliconANGLE.
Novakovic has walked this path before. He previously co-founded application performance monitoring startup Instana in 2015, raising $57 million in venture funding before selling the business to IBM in November 2020 for an estimated $250 million, according to CFO.com. Polar Signals, by comparison, had raised a more modest $10.8 million in total venture funding, capped by a $6.8 million round in March 2024 led by Spark Capital alongside Google Ventures and Lightspeed Venture Partners, as reported by SiliconANGLE.
The deal lands amid a broader boom in enterprise observability spending. The global market for IT observability tools and platforms is projected to grow from $11.91 billion in 2026 to $22.99 billion by 2031, a compound annual growth rate of 14.1%, according to MarketsandMarkets. Dash0 is betting that as software increasingly optimizes itself, the companies that can hand AI agents the deepest, most precise runtime data will be the ones that win enterprise customers away from established players.
