# Groundcover raises $100 million to challenge Datadog's observability model > Source: > Published: 2026-07-29 16:23:36+00:00 [Shahar Azulay](https://www.groundcover.com/about) and [Yechezkel Rabinovich](https://www.groundcover.com/writer/by-yechezkel-rabinovich) have raised $100 million for Groundcover, giving the observability provider fresh capital to pursue an explicit target: replacing [Datadog](https://www.datadoghq.com/pricing/) inside cloud infrastructure and AI workloads. The Series C was led by [One Peak](https://www.onepeak.tech/about), with [Morgan Stanley Expansion Capital](https://www.morganstanley.com/im/en-us/individual-investor/strategies/private-credit-and-equity/expansion-capital.html) joining alongside existing investors [Zeev Ventures](https://www.zeevventures.com/), [Angular Ventures](https://angularventures.com/how-we-invest), [Heavybit](https://www.heavybit.com/) and [Jibe Ventures](https://www.jibevc.com/), according to [CTech's July 29th report](https://www.calcalistech.com/ctechnews/article/r19aupvrme). Axios separately confirmed the round size, lead investor and participants. The financing values Groundcover at around $500 million and brings total funding since 2021 to $160 million, CTech reported. Azulay and Rabinovich met while working in the cyber division of Israel's Prime Minister's Office, where they developed and managed security and infrastructure systems. Their technical histories map directly onto the problem Groundcover is trying to solve. Azulay later managed machine-learning work at Apple after research and engineering roles at DayTwo and CYMOTIVE Technologies. Rabinovich became chief architect at healthcare security provider CyberMDX before starting Groundcover. Azulay has said his experience as a buyer shaped the founding thesis. Even organizations with large infrastructure budgets had to choose between collecting enough telemetry to debug production systems and controlling the resulting observability bill. The founders settled on eBPF, which can inspect activity at the Linux kernel layer without requiring engineering teams to add instrumentation throughout an application. ### A pricing attack built into the architecture Groundcover's challenge to Datadog starts below the dashboard. Its [eBPF-based platform](https://docs.groundcover.com/architecture/overview) collects logs, metrics, traces and Kubernetes events, then stores that telemetry inside the customer's cloud environment through a Bring Your Own Cloud, or BYOC, deployment. That structure changes Groundcover's cost base. Groundcover manages the software while customers retain the data plane and pay their own underlying storage and compute costs. Groundcover can therefore charge by monitored host rather than tying the software bill directly to ingestion volume, cardinality or the number of events collected. Groundcover's [public pricing](https://www.groundcover.com/pricing) lists its Pro plan at $30 per host per month, Enterprise at $35 and a fully on-premises deployment at $50. The pricing page says monthly bills use the average number of monitored hosts, limiting the impact of short-lived infrastructure peaks. Those list prices exclude the cloud resources customers use to run the BYOC deployment, a material cost that buyers still have to model. The trade gives Groundcover a clear sales argument. Engineering organizations can retain larger amounts of telemetry without sending every byte to Groundcover or paying Groundcover an ingestion fee. Groundcover still has to prove that lower and more predictable data charges can offset the operational and procurement advantages of buying from an established platform with a broad product catalog and integration base. Groundcover has built [migration software](https://www.groundcover.com/blog/escaping-datadog-how-we-built-an-automated-observability-migration-tool) that imports Datadog dashboards, monitors and associated data. That tooling is central to the replacement strategy. Observability platforms sit inside incident response, developer workflows and production operations, making the cost of changing vendors extend well beyond a software subscription. ### AI creates a larger telemetry bill The Series C arrives as Azulay broadens Groundcover's original cloud-monitoring pitch to cover AI agents and large language model applications. Agent systems can produce traces across model calls, tool use, intermediate steps and infrastructure, adding another high-volume data source to an already expensive monitoring stack. Groundcover has responded with an [MCP server](https://www.groundcover.com/ai-observability/mcp-server) that exposes production context to AI tools and Agent Mode connectors that bring telemetry into engineering workflows. A June 29th product release added connections for Slack, Linear, Jira, Claude Code, Codex, Cursor and incident-response tools, allowing an agent to investigate errors and act within the customer's existing software. Groundcover's [Agent Mode connectors](https://www.groundcover.com/blog/launching-connectors-to-agent-mode) run against telemetry kept in the customer's environment. That is the same architectural bet applied to a new buyer concern: companies want AI systems to use detailed production data without copying sensitive infrastructure records into another vendor-controlled cloud. Azulay told CTech that most Groundcover customers introduce telemetry and observability early in their AI adoption. That claim is self-reported, yet it explains why Groundcover is raising a round almost three times the size of its April 2025 Series B. Groundcover raised $35 million in that earlier financing, bringing total funding at the time to $60 million. ### Growth capital brings an operating test Groundcover told CTech that annual recurring revenue tripled over the past year, its workforce doubled to 140 employees and its paying customer count passed 250. Groundcover also said it signed multiple seven-figure contracts during the period. Groundcover did not provide the starting ARR figure or exact revenue, limiting the usefulness of the growth percentage. Azulay described annual revenue as being in the tens of millions of dollars and set a goal of reaching roughly $100 million in annual sales over the coming years. The headcount data gives a clearer view of what the financing buys. Groundcover said 80 employees are in Israel, with the remainder in Boston and San Francisco. The new capital is earmarked for North American expansion, engineering, AI observability products, cloud-provider partnerships and possible acquisitions of technical teams. Azulay had already identified organizational scaling as a risk. In [a post earlier in 2026](https://www.linkedin.com/posts/shahar-azulay-54156bb4_groundcover-will-2x-in-size-in-2026-that-activity-7415359612046278656-8oCY), he wrote that Groundcover had about 40 employees at the beginning of 2025 and expected to reach roughly 160 by the end of 2026. He called the pace "scary" and focused on whether Groundcover could preserve its speed, transparency and culture while building offices in Boston and San Francisco. That is the founder-operator problem behind the round. Groundcover's early differentiation came from a technical architecture and a direct cost argument. Replacing Datadog at enterprise accounts requires a larger sales organization, customer engineering, round-the-clock support and a widening integration catalog. Those functions can slow product development and raise costs even when customer telemetry stays off Groundcover's infrastructure. One Peak specializes in growth-stage software investments and says it typically writes checks of $15 million to $150 million into businesses with at least $5 million in run-rate revenue. Its involvement puts Groundcover into a different operating phase from the one backed by inception-stage investors Angular and Heavybit. The valuation now rests on Groundcover turning a credible infrastructure wedge into a broad enterprise platform before Datadog and other incumbents neutralize its pricing and AI-observability arguments.