By admin 
The AI agent observability space is witnessing an unprecedented surge in growth, prompting enterprises to critically evaluate their needs for advanced observability products and solutions. In this dynamic environment, observability startup groundcover has announced a significant milestone: securing $100 million in a Series C funding round led by One Peak. This latest infusion of capital brings the company’s total funding to an impressive $160 million, signaling strong investor confidence in its vision for the future of operational intelligence.
groundcover claims to have amassed over 250 paying customers, a testament to its rapidly expanding market presence. The company also reports a remarkable tripling of its annual recurring revenue over the past year, indicating substantial traction and a growing demand for its offerings. More notably, groundcover is increasingly displacing established observability platforms within enterprise environments, a bold claim that underscores its disruptive potential in a fiercely competitive sector. While these figures are company-reported, they collectively point to a powerful momentum in one of enterprise software’s most entrenched markets.
Historically, the observability market has been dominated by industry giants such as Datadog, Dynatrace, New Relic, Splunk, and Grafana. These established players command billions in annual revenue and boast years of product maturity and extensive customer bases. Breaking into this established order has always been an arduous challenge. However, groundcover’s core argument is that the advent of artificial intelligence has fundamentally disrupted the foundational assumptions upon which these traditional observability platforms were built. Rather than engaging in a feature-by-feature competition, the four-year-old company is advocating for a paradigm shift in observability architecture itself. This shift is necessitated by the increasing autonomy of AI systems, their prodigious generation of telemetry data, and their growing involvement in software operations. While the ultimate success of this thesis remains to be definitively proven, it offers a compelling and insightful lens through which to examine the profound evolution of observability in tandem with the accelerating adoption of enterprise AI.
AI is Transforming Telemetry into a Core Infrastructure Challenge.
Traditionally, observability has been largely confined to a post-production discipline. The established workflow involved engineers deploying applications, diligently monitoring logs, metrics, and traces, subsequently investigating any identified incidents, and then iteratively improving system reliability. However, this well-trodden path is rapidly changing. The integration of AI-assisted software development has dramatically accelerated deployment cycles. AI-powered coding assistants are generating code at an unprecedented rate, infrastructure is evolving with unparalleled speed, and organizations are deploying increasingly sophisticated distributed systems that amalgamate microservices, Kubernetes clusters, APIs, and large language models. Concurrently, enterprises are beginning to operate sophisticated AI agents capable of executing multi-step workflows, invoking external tools, and directly interacting with production systems.
Each of these emergent activities generates a substantial volume of telemetry data. The cumulative effect is an explosion of operational data that organizations are increasingly reluctant to discard. AI applications introduce additional, novel layers of observability that extend beyond traditional infrastructure monitoring. These new layers encompass the intricacies of prompt execution, model latency, token consumption, retrieval pipeline performance, tool invocations, and the nuanced behavior of AI agents. As enterprises venture further into the realm of autonomous systems, this rich telemetry becomes exponentially more valuable, providing the critical context necessary to understand precisely what an AI system has done and, more importantly, why it acted in a particular manner.
For many organizations, this burgeoning data volume creates significant tension with existing pricing models that often charge based on the sheer amount of data ingested. Historically, engineers have frequently resorted to cost-saving measures such as sampling traces, drastically shortening data retention periods, or selectively limiting the scope of data collection. While these approaches effectively reduce costs, they simultaneously diminish visibility, precisely at a time when AI-driven systems demand more comprehensive operational context. Shahar Azulay, co-founder and CEO of groundcover, articulated this frustration during a recent media briefing, stating, "We’ve seen telemetry exploding. Users are frustrated by not getting all the value from Datadog and similar platforms. They’re limiting the data, siloing it, sampling it." Whether this widespread frustration will be sufficient to fundamentally reshape the observability market remains an open question, but the underlying trend is undeniable. AI is fundamentally shifting the focus of observability from simply collecting "enough" data to meticulously gathering everything an organization might potentially need.
groundcover Advocates for Architectural Overhaul, Not Just AI Additions.
In response to the evolving demands of AI-driven operations, many traditional observability vendors have rapidly introduced AI-powered assistants, AI-driven root cause analysis capabilities, and specialized AI observability features over the past two years. Leading platforms like Datadog, Dynatrace, New Relic, and Grafana have all announced new products designed to assist enterprises in monitoring AI applications and automating various operational tasks. groundcover, while acknowledging these developments, contends that they do not address what it perceives as the more fundamental, architectural issues: the location of telemetry data and the prevailing pricing models.
Instead of operating a conventional Software-as-a-Service (SaaS) platform that stores customer telemetry within vendor-managed infrastructure, groundcover employs a distinct approach it terms "bring-your-own-cloud" (BYOC) architecture. In this model, customers retain complete control over the data plane, encompassing telemetry storage and processing, within their own cloud environments—whether that be AWS, Microsoft Azure, or Google Cloud. groundcover, in turn, provides a managed control plane and a user-friendly experience. Furthermore, a fully self-hosted deployment option is also available for organizations requiring maximum control.
While some competitors, including Datadog and a select few other observability vendors, offer limited hybrid or customer-controlled data residency options, these generally do not equate to a comprehensive BYOC model. In most conventional scenarios, telemetry data continues to be processed and stored within the vendor’s managed infrastructure, with customers typically afforded only partial controls such as regional data residency, private links, or selective log forwarding. This fundamental architectural divergence significantly influences nearly every facet of groundcover’s strategic direction.
Because customers already bear the cost of their own cloud infrastructure, groundcover posits that it can bypass the practice of charging based on telemetry ingestion volumes. Instead, its pricing is primarily predicated on the number of monitored hosts, irrespective of the telemetry volume generated. The company believes this pricing strategy fundamentally alters customer behavior. Rather than making difficult decisions about which logs or traces are too expensive to retain, organizations can, in theory, maintain complete telemetry data sets, leveraging them for comprehensive operational analysis, regulatory compliance, and sophisticated AI-assisted troubleshooting.
"We don’t price by data volume," Azulay explained. "We price by the size of the infrastructure." This distinction is crucial because AI workloads tend to generate telemetry data at a far greater rate than the underlying infrastructure itself. This does not necessarily translate to universally cheaper pricing. Organizations with relatively light workloads distributed across a large number of hosts might find different economic advantages compared to dense Kubernetes environments generating immense volumes of telemetry. The company’s own briefing materials acknowledge that its per-host pricing model is most advantageous for organizations with high telemetry density and may be less compelling for lightly utilized fleets. Nevertheless, the broader strategic argument extends beyond mere cost savings to encompass predictability. Enterprise infrastructure teams frequently grapple with observability bills that fluctuate unpredictably with application growth. groundcover’s model aims to align pricing more closely with infrastructure planning and resource allocation rather than the often volatile nature of data generation.
eBPF Forms the Technical Cornerstone of groundcover’s Differentiation.
The second critical pillar underpinning groundcover’s strategic approach is its deep integration and utilization of eBPF (extended Berkeley Packet Filter). eBPF is a transformative Linux kernel technology that has rapidly emerged as one of the most vital building blocks for modern cloud observability solutions. Instead of compelling developers to undertake the often complex and time-consuming task of manually instrumenting applications, eBPF enables software running directly within the operating system kernel to observe network traffic, system calls, and application behavior with minimal or no code modifications. This capability facilitates significantly faster deployment cycles and broader, more comprehensive visibility across entire infrastructures.
For organizations operating Kubernetes clusters and cloud-native applications, reducing instrumentation complexity can translate directly into shorter deployment times while simultaneously enhancing telemetry coverage. Azulay argues that this capability becomes particularly indispensable as AI systems increasingly generate complex, multi-faceted interactions across distributed services. "Our sensor allows us to observe systems very deeply, from infrastructure to application to AI workloads, without developers needing to instrument code," he elaborated during the briefing.
While eBPF technology itself is not exclusive to groundcover, with many observability vendors now incorporating it into their platforms, groundcover contends that its unique differentiation lies in the synergistic combination of automatic eBPF collection with customer-controlled data storage, OpenTelemetry compatibility, and unified pricing, all delivered within a single, cohesive platform. The company’s own research briefing acknowledges that none of these individual technologies constitute an insurmountable competitive moat. The claimed differentiation arises from the integrated approach: an eBPF-first collection strategy, a managed BYOC architecture, host-based economics, and comprehensive full-stack observability delivered as a unified solution.
AI Agents Emerge as Both Customers and Sophisticated Users of Observability.
Perhaps the most forward-looking and intriguing aspect of groundcover’s strategy extends beyond the traditional paradigms of system monitoring. The company increasingly frames observability not just as a diagnostic tool but as fundamental infrastructure for autonomous software development. Historically, observability platforms have primarily served human operators tasked with investigating production incidents and performance issues. groundcover envisions a future where observability platforms will increasingly serve AI agents as well.
Its "Agent Mode" product empowers engineers to investigate incidents using natural language queries across logs, metrics, traces, and Kubernetes events. More profoundly, Azulay articulates a vision where observability becomes the critical feedback loop that informs coding agents about the actual behavior and outcomes in production environments. Rather than merely detecting failures after deployment, observability evolves into a continuous stream of operational context that autonomous systems can leverage to rigorously evaluate changes, proactively identify regressions, and, ultimately, recommend or even implement corrective actions. "We’re seeing observability moving from being a post-production tool… to people taking context from production and feeding it back to their coding agents so they can write code better," Azulay stated.
Currently, the company emphasizes that human oversight remains paramount. While Agent Mode excels at investigating incidents and surfacing actionable recommendations, any changes made to production systems still require explicit human approval. Azulay anticipates a gradual increase in autonomy as organizations become more comfortable delegating operational workflows to AI systems. This vision mirrors a broader trend emerging across the enterprise software landscape, where AI agents are increasingly permeating development, testing, deployment, and operational phases, moving beyond their role as isolated assistants.
Strategic Considerations for Enterprises Evaluating Observability Alternatives.
groundcover is strategically entering an intensely competitive market populated by vendors with decades of accumulated enterprise experience and established market dominance. Datadog, for instance, alone generated over $3 billion in annual revenue in 2025. Dynatrace, Cisco’s Splunk business, Grafana Labs, and New Relic all maintain extensive partner ecosystems, mature integrations, and robust enterprise support organizations that newer entrants find exceptionally difficult to replicate quickly. groundcover is not attempting to achieve parity with these incumbents overnight. Instead, it positions its strategy around the argument that AI represents an architectural inflection point, analogous to the seismic shifts from on-premises infrastructure to cloud-native computing.
According to Azulay, many customers initially adopt groundcover primarily to reduce their observability costs. However, they increasingly remain loyal customers because they value unrestricted access to richer telemetry data and the promise of AI-native operational workflows. He asserts that groundcover deployments typically replace incumbent platforms rather than operating in parallel, although the company has not yet publicly disclosed detailed customer migration data or independent studies to definitively validate this claim.
The company’s journalist briefing also advises a degree of caution regarding certain performance claims. Revenue growth, customer acquisition figures, and enterprise adoption statistics are sourced directly from groundcover. Published customer case studies detailing significant cost savings are authored by the vendor and should be approached with a critical eye, requiring further independent validation. The briefing further recommends a thorough examination of precisely what metadata, if any, leaves customer environments in standard BYOC deployments, rather than making a blanket assumption that no operational data ever reaches vendor infrastructure.
These caveats are particularly important given the increasingly crowded nature of the observability market. Gartner, a leading research firm, currently tracks over one hundred distinct observability products, and nearly every major vendor now actively markets AI-powered operational capabilities. Ultimately, success in this evolving landscape will likely hinge less on whether AI is relevant—a point that appears increasingly inevitable—and more on whether enterprises conclude that their existing architectural paradigms remain sufficiently adequate to meet future demands.
The Overarching Question Investors Are Betting On.
Viewed in isolation, groundcover’s Series C funding round represents another substantial infrastructure funding event. However, when considered in a broader context, it reflects a growing and critical debate about the future evolution of observability in an era where software increasingly possesses the capability to write, test, and operate itself. If AI continues to generate exponentially larger volumes of operational data, traditional assumptions regarding telemetry collection, pricing methodologies, and data storage strategies are likely to face mounting pressure. Vendors that have built their business models around charging for data ingestion may find it imperative to evolve their economic frameworks in tandem with shifting customer expectations. Concurrently, new entrants like groundcover have a strategic opportunity to architect their solutions around these evolving assumptions from the ground up.
groundcover believes this opportunity lies in the strategic convergence of customer-controlled infrastructure, automated telemetry collection, and AI-assisted operations, all integrated into a platform specifically engineered for autonomous software rather than simply adding AI features as an afterthought to existing observability products. Whether this fundamental architectural bet proves to be durable and sustainable will ultimately depend on widespread enterprise adoption over the coming years. However, the company’s latest funding round strongly suggests that at least a segment of the investor community believes that the next major battleground in the observability space will not be fought over sophisticated dashboards or proactive alerts. Instead, it will be contested by those who successfully build the operational data layer upon which increasingly intelligent software relies to understand—and, eventually, to autonomously manage—the complex systems it operates.
