We are launching primary research to determine whether Datadog's 32% revenue reacceleration reflects a durable AI workload tailwind or a cyclical bump driven by hyperscaler training contracts and concentrated exposure to a single customer.
Datadog reported Q1 2026 results before the market opened on 7 May and delivered the largest single-quarter narrative reset in its history as a public company. Revenue of $1.006 billion crossed the ten-figure threshold for the first time, growth accelerated to 32% year-over-year from 29% the prior quarter, and management raised full-year revenue guidance by roughly $240 million at the midpoint. The stock jumped 31% in a single session, its biggest one-day move since the 2019 IPO. Sell-side price targets were rebuilt in a band of $215 to $250, with TD Securities calling the print "eye-popping" and labelling Datadog a must-own stock. We are launching primary research to find out whether the reacceleration is structural.
The financial detail matters because it sits at the heart of the debate. Quarter-over-quarter revenue rose 6%, the strongest first-quarter sequential add since 2022, with a $53 million sequential increment. Annual recurring revenue passed $4 billion, with management noting that ARR growth accelerated in each month of the quarter. Non-GAAP operating margin came in at 22% and free cash flow margin at 29%. New logo annualised bookings set an all-time record and more than doubled year-over-year. Trailing twelve-month net revenue retention moved up into the low 120s.
The bull case is that this is the moment AI workloads move from experimentation into production, layered on top of a non-AI customer cohort that itself reaccelerated into the mid-twenties percent year-over-year, up from 23% the prior quarter and 19% a year ago. The bear case, championed by Guggenheim and Truist, has spent months arguing that the largest customer, OpenAI, is internalising observability tooling and could cut $150 million or more from 2026 revenue. Management's continued language around "applying a higher degree of conservatism to our largest customer" keeps the debate alive even after the print appears to refute it.
The catalyst window is compressed. The DASH user conference runs on 9 and 10 June in New York, with new AI observability and security products expected. Q2 earnings in August will be the first proof point on whether the 30%-plus trajectory holds and whether the largest-customer caveat was real sandbagging. Consensus models have not yet incorporated FedRAMP High, the two hyperscaler superintelligence-lab training wins Pomel disclosed on the call, or the doubling of new logo bookings. The window for differentiated primary research closes when the sell-side fully internalises these data points.
Key Insights
The reacceleration is broad, not concentrated in AI-native customers. Non-AI customer revenue growth accelerated into the mid-twenties percent year-over-year, up from 23% the prior quarter and 19% in the year-ago period. That detail directly undercuts the bear thesis that Datadog's print is entirely a function of OpenAI and a handful of AI-native logos. It points to cloud migration and platform consolidation tailwinds running alongside the AI workload narrative.
The guidance raise is one of the largest in the company's history. Full-year 2026 revenue guidance moved to $4.30 billion to $4.34 billion, up from a prior range of $4.06 billion to $4.10 billion, against a consensus of roughly $4.12 billion. Q2 revenue is now guided to $1.07 billion to $1.08 billion, well above the $994 million Street estimate, with non-GAAP EPS of $0.57 to $0.59 versus consensus around $0.50. Management explicitly retained its conservatism framework, including a higher haircut on the largest customer, which suggests further upside if that customer stabilises.
Two hyperscaler training wins reframe the competitive moat. On the call, CEO Olivier Pomel disclosed that Datadog has signed two large hyperscaler customers for training workloads inside their superintelligence laboratories. The signal matters because hyperscalers have historically built observability tools internally. Pomel attributed the shift to the "urgency" of the AI race, with even the largest technology companies prioritising speed over internal tooling. The economics and durability of these contracts are unmodelled in sell-side numbers.
Platform breadth is widening, raising switching costs. 56% of customers now use four or more Datadog products, up from 51% a year ago. 35% use six or more, up from 28%. 20% use eight or more, up from 13%. Of 26 shipping products, five generate over $100 million in ARR and three sit between $50 million and $100 million, with 18 still early in their lifecycle. In an enterprise market focused on vendor rationalisation, every additional module increases switching costs and weight in future budget decisions.
Datadog is widening the gap on its closest pure-play peer. Dynatrace reported revenue of $515.5 million growing 18.2% year-over-year, with fiscal 2026 total revenue guidance of $1.95 billion to $1.965 billion. Datadog's ARR alone is now over $4 billion, more than double Dynatrace's $1.73 billion. Cloudflare grew 34% but is cutting more than 1,100 jobs as AI reshapes its workforce. The competitive threat that matters medium-term is Palo Alto Networks' announced acquisition of Chronosphere, which introduces pricing pressure as Datadog pushes further upmarket.
FedRAMP High is the sleeper catalyst the sell-side has not modelled. Datadog for Government received FedRAMP High certification during the quarter, opening expanded sales into US federal agencies and high-security public sector accounts. The federal observability total addressable market is essentially absent from current sell-side models. First announced agency wins will be a discrete data point to watch over the next two quarters.
Participation Opportunity
Woozle Research is inviting professional investors to sponsor or co-sponsor this primary research. Participation is collaborative. All funds receive full access to research outputs including interview summaries, transcripts, and the final synthesis report.
Launch: 13 May 2026
Delivery: 25 May 2026
Participation: Limited to 5 funds
Catalyst: Q1 2026 print, $1 billion revenue threshold, guidance raise, DASH conference window
Research: 30+ enterprise CIO and SRE interviews across AI-native and traditional enterprise cohorts, 15+ OpenAI and hyperscaler infrastructure channel checks, 15+ federal systems integrator and FedRAMP buyer interviews, 10+ competitive win/loss interviews against Dynatrace, Chronosphere, and Splunk
Deliverables: raw data, transcripts, synthesis report, analyst access
This research will proceed with a minimum of one fund and is limited to a maximum of five.
The Catalyst
Datadog's print has done something unusual in software this year. It has reset a debate rather than continued one. For most of 2025 and into the early weeks of 2026, the consensus story on Datadog was about concentration risk. Guggenheim had downgraded the stock to Sell in July 2025 with a $105 price target, warning of a $150 million or greater revenue hole if OpenAI moved observability in-house. Truist had argued that management's framework implied no incremental contribution from the largest customer in 2026. The Q4 2025 print in February triggered a sell-off that took the stock to roughly $111. Sentiment going into the May print was depressed, even with 42 of 46 analysts at Strong Buy, because the implied upside was conditioned on a customer the market believed was leaving.
Pomel's response on the call was to widen the lens. The non-AI cohort accelerated. New logo bookings doubled. Hyperscalers, the most credible internal-build candidates in the entire software market, signed training-workload contracts. Pomel described the average enterprise observability stack as "4, 6, 7, 15, 25 different things" spread across organisations, calling it "a huge mess". That framing is doing analytical work. It positions Datadog as the consolidation play, with AI as an additive layer rather than the primary growth engine. If the framing holds, the largest-customer debate becomes a sideshow rather than the central question.
The more troubling narrative for bears is the inflection language itself. Pomel told the call that the company is seeing an inflection point in customer consumption as AI workloads move into production environments. He pointed to increasing data volumes at every layer of the platform, AI products getting into production and finding users, and the same dynamic playing out at AI-native and non-AI companies. Tone matters here. As recently as February, management was caveating revenue volatility in the AI-native cohort. The shift from "we may see volatility" to "we see an inflection point" inside one quarter is not a small editorial change. It is a deliberate narrative reframe.
The human element is worth surfacing. When Pomel and his co-founder started Datadog in 2010, they spent the first six months without writing a single line of code, listening to operators and developers describe the gap between their teams. That obsessive listening became the cultural foundation of a company that has now compounded to over $4 billion in ARR. Pomel's own candour on the call was notable. Asked about the long-term shape of the business under agentic AI, he acknowledged it is hard to tell where the company will be in four or five years, noting that two years ago he would not have predicted that most engineers would return to coding in the console. That candour cuts both ways. It is refreshing relative to the typical software CEO script. It also tells you the pricing model question, in particular how per-host and per-user economics behave when AI agents replace human SREs, is genuinely unresolved at the company level.
The competitive picture sharpens the urgency. Dynatrace continues to grow at 18%, roughly half Datadog's pace, and has acknowledged the competitive gap to peers with broader portfolios. Cloudflare grew 34% but is reshaping its workforce. The medium-term risk that matters is Palo Alto Networks' acquisition of Chronosphere, announced in November, which puts a well-capitalised competitor into the observability category just as Datadog pushes upmarket. The risk is not that Chronosphere wins the next twelve months. The risk is that pricing pressure shows up in renewals over the next eighteen to twenty-four months, particularly at the largest enterprise accounts where Datadog has been most aggressive on price.
The forward trajectory hinges on three discrete catalysts inside a sixty-day window. DASH on 9 and 10 June will set the product narrative for the second half of the year, with major AI observability and security launches expected and a roughly $15 million cost impact already inside Q2 guidance. First federal agency wins post-FedRAMP High should land in the same window. Q2 earnings in August will be the first read on whether the largest-customer conservatism was real sandbagging. The sell-side has not yet rebuilt models to incorporate any of these. That is the window for differentiated primary research.
Key Intelligence Questions
The research will focus on the commercial and operational dynamics that determine whether Datadog's reacceleration is durable, where the largest-customer haircut actually sits, and whether new revenue surfaces such as federal and hyperscaler training are sized correctly. Each question targets a specific input to the investment model.
OpenAI Workload Trajectory: Real Haircut or Sandbag?
The single most contested input into Datadog's 2026 model is the trajectory of its largest customer. Guggenheim's bear case rests on the assertion that OpenAI is building in-house log management and metrics tooling and will materially reduce its Datadog spend over the year. Truist has read management's "core business excluding the largest customer grows at least 20%" framing as a signal that no incremental growth is being assumed from that account in 2026. Management's continued conservatism language in the Q1 release supports the bear interpretation literally, even as the headline numbers refute it.
The public data cannot resolve this. Datadog does not disclose customer-level revenue. OpenAI's infrastructure spend mix is not public. Recent estimates of OpenAI's Datadog ARR sit in a wide $170 million to $300 million range and predate the Q1 print. What looks like sandbagging from the outside could equally be an honest read on a customer actively internalising workloads. The difference between those two interpretations is worth several hundred million dollars of 2026 revenue and a meaningful slice of the current valuation.
Key Intelligence Question: What percentage of OpenAI's observability spend has been internalised versus retained at Datadog over the past two quarters, and what is the directional trajectory through the remainder of 2026? Are the internal tools OpenAI is building intended to replace Datadog or to complement it on specific workload types, and how do OpenAI infrastructure and SRE teams describe the renewal posture?
Hyperscaler Training Deals: Contract Structure and Durability
Pomel's disclosure that Datadog has signed two large hyperscaler customers for training workloads inside their superintelligence labs is one of the most consequential single data points in the print. Hyperscalers have historically built observability internally. If they are now buying externally on training workloads, the addressable market for AI-specific observability widens materially. But the economics and durability of these deals are unknown. Are they committed multi-year contracts, or consumption-based bursts tied to specific training runs? Are they tied to specific superintelligence projects with finite duration, or are they platform-level relationships that extend into inference?
The distinction matters because training-workload revenue is structurally different from steady-state production observability. A training run that completes is a revenue event, not an annuity. If the bulk of the hyperscaler revenue is consumption-tied to specific training campaigns, the durability question becomes acute when those campaigns wind down or shift internally.
Key Intelligence Question: Which hyperscaler labs are the two unnamed customers Pomel referenced, and what is the contract structure of each deal? Are these committed multi-year platform agreements, consumption-based training contracts tied to specific superintelligence projects, or hybrid arrangements, and what is the renewal probability beyond the initial training campaigns?
Non-AI Cohort Durability: Catch-Up or Structural?
The most analytically important number in the print may not be the headline. It is the mid-twenties percent reacceleration in the non-AI cohort, up from 23% the prior quarter and 19% a year ago. The bull interpretation is that cloud migration and platform consolidation are running as durable tailwinds independent of the AI workload narrative. The bear interpretation is that this is a post-optimisation rebound, with enterprises that aggressively cut cloud spend in 2023 and 2024 now reinvesting on a lagged basis. Both are consistent with the reported numbers. They have very different implications for 2027 and 2028.
Public data does not distinguish between these. CIO and IT-buyer behaviour at the deal level does. Whether enterprises are expanding Datadog usage because they are migrating new workloads to cloud, consolidating multiple observability vendors, or simply normalising spend after a cycle of cuts is a question that can only be answered through direct conversations with buyers.
Key Intelligence Question: Among large non-AI enterprise customers, what is driving the acceleration in Datadog spend over the past two quarters? Is the dominant driver new workload migration, consolidation of competing observability vendors, normalisation after the 2023 to 2024 optimisation cycle, or expansion into security and adjacent product categories, and which of these is sustainable into 2027?
FedRAMP High Pipeline: How Large is the Federal Opportunity?
FedRAMP High certification is treated almost as a footnote in the Q1 release, but it opens a category of customers that is essentially absent from current sell-side models. Federal observability spend is structured differently from commercial spend. Procurement cycles are longer, but contract sizes are larger and stickier once awarded. The competitive set is also different. Splunk has historically owned the federal observability market through its security positioning. Datadog's entry with FedRAMP High puts it into direct contention for agency-level platform consolidation.
The question is whether the federal pipeline materialises in 2026 or sits as a 2027 to 2028 contributor. The answer depends on systems integrator engagement, agency-level procurement timing, and competitive displacement against incumbents.
Key Intelligence Question: How are federal systems integrators such as Booz Allen, Leidos, and GDIT positioning Datadog into agency procurement opportunities post-FedRAMP High, and which agencies are nearest to award? What is the realistic 2026 and 2027 federal revenue contribution, and against which incumbent vendors is Datadog most actively displacing?
Chronosphere and Pricing Pressure in Renewals
The competitive risk that matters medium-term is not Dynatrace, growing at half Datadog's pace. It is Palo Alto Networks' acquisition of Chronosphere, which puts a well-capitalised competitor into observability with a credible product and platform-bundling logic. The risk shows up first in renewals at large enterprise accounts, particularly where security and observability budgets sit with the same buyer.
Datadog's growth has been supported by strong pricing power, with net revenue retention in the low 120s. If Chronosphere and Palo Alto begin to compete aggressively on price for renewals in late 2026 and 2027, that retention number compresses, and the margin profile of new business changes. The question is whether this is already showing up in win/loss dynamics on large deals or remains a 2027 risk.
Key Intelligence Question: In large enterprise observability deals over the past six months, how often is Chronosphere appearing on the shortlist alongside Datadog, and what is the win/loss pattern? Are there early signs of pricing pressure in Datadog renewals at customers where Palo Alto is the incumbent security vendor, and how are procurement teams modelling the platform-bundling logic?
How to Participate
Woozle Research is inviting professional investors to sponsor or co-sponsor this primary research. Participation is collaborative. All funds receive full access to research outputs including interview summaries, transcripts, and the final synthesis report.
Launch: 13 May 2026
Delivery: 25 May 2026
Participation: Limited to 5 funds
Catalyst: Q1 2026 print, $1 billion revenue threshold, guidance raise, DASH conference window
Research: 30+ enterprise CIO and SRE interviews across AI-native and traditional enterprise cohorts, 15+ OpenAI and hyperscaler infrastructure channel checks, 15+ federal systems integrator and FedRAMP buyer interviews, 10+ competitive win/loss interviews against Dynatrace, Chronosphere, and Splunk
Deliverables: raw data, transcripts, synthesis report, analyst access
This research will proceed with a minimum of one fund and is limited to a maximum of five.
Email to confirm your interest
This document is for informational purposes only and does not constitute investment advice or a recommendation to buy or sell any security. Woozle Research conducts primary research on behalf of institutional investors. All research is conducted in compliance with applicable regulations.
