Datadog

Datadog shares are down despite second-quarter earnings beating management's forecasts on both the top and bottom lines. The company guided to a future slowdown in spending from its largest artificial intelligence-native customer—OpenAI—which we suspect is the reason for the selloff.

Why it matters: Customer concentration risk, deceleration in the revenue growth outlook, and crowded positioning driving up the stock price recently are likely hurting the shares. But the underlying business is healthy. Customers (AI-native and non-AI), multiproduct adoption, and incremental sales and marketing expenses are converting to revenue at an impressive clip—all indicative of a network-effect moat. Nearly all quantifiable key performance indicators—from growth to profitability to lifetime value of customers—sit in the top decile of companies in our software universe. This commands a premium as investors seek out enablers of an agentic future. We love the business case for Datadog, but the valuation is a hurdle. Datadog has traded more richly than Palantir for much of 2026.

The bottom line: We maintain our wide-moat rating and our $200 fair value estimate as we see a fantastic product fit for a growing end market, but we also see a demanding valuation that leaves very little margin for error. Shares are overvalued, in our view, but we would happily invest in this observability leader at a lower share price. Long-term secular drivers of observability demand like cloud migration, agentic workflows, and GPU monitoring are all strong, providing a healthy backdrop for our modeled 25% average annual growth rate over the next five years, but customer concentration risk is a key risk factor to monitor. Mitigating this somewhat is increasing diversification among AI-native customers.

The indiscriminate selling of anything software from the first-quarter SaaSpocalypse has transformed into a sorting mechanism—as we predicted in our first-quarter Technology Observer—that rewards infrastructurally positioned software firms (cybersecurity, and so on) that sit at complex customer choke-points—but sometimes that rewarding mechanism can go too far. In these scenarios, a wide-moat company like Datadog can be disproportionately punished on any hints of nondurability of growth. Should we see another double-digit move to the downside or a persistent grind-down over the coming months without any material change to the long-term trajectory, we would advocate sizing into what we believe is a phenomenal company.

Our $200 per-share fair value estimate implies an enterprise value of 15 times our 2026 sales estimate.

The main drivers of Datadog’s revenue are the enterprise cloud migration movement, the adoption and development of applications, increasing complexity of technological infrastructures, growth in unstructured data, enterprise demand for efficiency-enhancing software, and Datadog’s ability to penetrate these markets and continually develop high-value products.

To develop our revenue forecasts, we estimate Datadog’s penetration of the IT market by compiling and analyzing various market size estimates. We also analyze the growth trajectory of data generation and the role that telemetry will play within an influx of AI workflows. We then create a baseline projection based on the attractiveness of Datadog’s core offerings. This results in an average annual growth rate of roughly 25% over five years, decreasing to 13% by 2035.

Our sizing estimate for the IT operations market is approximately $65 billion in 2026, and we expect an average annual growth rate of approximately 10%, resulting in $80 billion by 2028. We anticipate that as IT workflows mature, this growth rate will slow to 9% beyond 2029, eventually reaching a $130 billion market by 2034.

In 2020, the IT operations market was about $35 billion, and Datadog had only a 2% penetration. However, due to effective execution, strong product-led growth, and rising adoption of cloud workflows, Datadog has grown faster than the overall market. We currently estimate its penetration at roughly 7%, and we believe that switching costs and network effects will help Datadog continue capturing market share, ultimately reaching a 14% penetration of the $120 billion market by 2034.

Cost of revenue, R&D, and sales and marketing are the largest expenses, primarily consisting of salaries. Cost of revenue also includes third-party cloud hosting costs. Cost of revenue has decreased from 25% of revenue in 2019 to 20% in fiscal-year 2025, and we project it will decrease to 19% by 2034, thanks to better volume discounts from hyperscaler cloud providers and internal efforts to optimize data compression. This amounts to a gross margin of 81% by 2034—placing Datadog near the 80th percentile of 241 software companies since 2015. R&D as a percentage of revenue has remained consistently around the mid-40% range since 2022. Although we expect R&D spending to increase in dollar terms, the fixed salary component should begin to scale well as the company adds more customers, leading us to estimate R&D at 33% of revenue by 2034.

We assign Datadog a Morningstar Economic Moat Rating of wide, based on switching costs and network effects.

Strong switching costs tied to Datadog's core software solution arise from cloud-based applications and the vast amounts of data those services generate. Switching observability platforms is challenging. Whether it is an in-house or third-party application, the applications businesses run compose their technological stacks. If anything goes wrong anywhere in this stack and companies lack visibility into what’s happening, the business can lose millions of dollars in seconds or minutes. Switching costs stem from the significant time investments to deploy observability solutions. Switching to a new platform can be disruptive and expensive because customers need to reconfigure, relearn, and redeploy various observability solutions. Switching costs also come from high customer risk aversion, high cost of failure, and the mission-critical nature of the systems supported by the observability platforms. These switching costs keep customers from chasing cheaper options.

Large e-commerce firms provide a good example of the switching costs. Managing hundreds of thousands of sales on their websites, the firms must track how a customer moves from browsing to checkout, then to payment processing. Each stage generates unique telemetry (logs, metrics, traces) that must be analyzed, correlated across various services, and displayed clearly so users can see four signals: latency (how long requests take), traffic (system load), errors (failure rate), and saturation (resource limits). If any of these signals fail and go unaddressed, errors in customer requests can accumulate, risking unhappy customers and lost sales. This mission-critical aspect of monitoring, and its link to revenue, is what observability platforms address, which in turn increases the cost of switching from a proven vendor.

Switching to a new platform can be disruptive and expensive because customers need to reconfigure, relearn, and redeploy various observability solutions. Lightweight "agents"—software processes that run throughout the company's tech infrastructure—must be reintegrated with individual components of the tech stack to handle telemetry collection. Even if the initial setup is quick, customizing and fine-tuning the observability platform still locks in resources after significant engineering hours are invested. In a hypothetical switch, a customer would likely need to run parallel observability systems for months to ensure full-stack visibility isn’t degraded. Effectively, there is a time cost, risk, and financial burden with switching.

The lower half of Datadog’s 30,000 customers contribute only 1%-2% of total revenue but contribute substantially to the firm's network effects. The smaller small- and medium-sized business customers provide valuable insights into diverse workloads that shape Datadog’s product roadmap. The land-and-expand and SMB-focused model offers a different perspective because smaller clients often experiment with new technology before larger enterprises do. Datadog can analyze usage patterns and requested integrations while also identifying what drives positive observability outcomes versus what doesn’t. Datadog can use customer feedback to enhance its offerings. More customers with diverse use cases lead to more feedback and data, which in turn creates a better product roadmap for Datadog, ultimately fostering stronger customer entrenchment.

When a large company requests an observability feature or integration, Datadog is often already familiar with the request, aware of potential technical challenges, and prepared with a solution. This happened with Amazon’s initial launch of AWS Lambda, its serverless product offering, a decade ago. Early adopters were smaller businesses that didn’t want to spend time and money managing servers for small pieces of code that only needed to scale during high demand. This AWS innovation created new observability challenges because applications became dominated by short-lived processes and fragmented, high-volume telemetry. Traditional observability solutions couldn’t monitor processes that only lasted 200 milliseconds. Recognizing this shift, Datadog developed serverless monitoring in 2017—well before other observability platforms, including AWS. When larger enterprises eventually adopted serverless solutions, Datadog benefited disproportionately.

AWS, Microsoft Azure, and Google Cloud are significant threats in the software world. They could cut out the need for specialized software and take over the market. However, we believe observability buyers naturally gravitate toward neutral, nonbiased platforms that operate seamlessly across clouds and markets. The large tech firms tend to act as walled gardens that optimize for their own ecosystems and aim to lock customers in. AWS, Microsoft, and Google all offer some version of an observability tool. Still, these are designed around the parent company’s own naming and language conventions and its own dependency map. Using AWS’ observability tool to ingest, query, or display telemetry from an application running on Microsoft Azure would run into a complex mapping of filters, alert adjustments, and data indexing rules to make sense of the telemetry in one cloud and display it in another. The large firms are unlikely to dedicate resources to fixing this cross-cloud problem because they make more money by keeping customers locked into their environments. Datadog builds unified, cloud-agnostic observability platforms that effectively handle complex technology infrastructures.

Increasingly complex technological infrastructures, cloud migration, and more third-party applications increase demand for observability platforms like Datadog’s that make sense of telemetry data.

Datadog’s cloud-neutral approach and unified taxonomy and topology reduce cloud vendor lock-in, making it an attractive value proposition to the majority of firms that prefer multicloud services.

Datadog’s telemetry orchestration creates a unified environment that allows AI agents to operate freely within an organization. We expect the enablement of AI agents to be a major revenue driver.

Competition in observability is intense, forcing Datadog to invest heavily in research and development to be ahead of the curve on new features and integrations.

Datadog’s attempts to move upstream and win over large customers involve significantly more investment in sales and marketing. Success is not guaranteed when competing more directly against Dynatrace.

Datadog’s consumption-based pricing model can occasionally surprise customers with large bills if the customer is not paying attention to usage metrics.