How does Datadog make money in 2027?

Direct Answer

Datadog makes money by metering observability and security telemetry across 10+ named modules, with per-host APM as the anchor SKU, per-GB Logs ingestion as the volume printer, and per-event Cloud SIEM / Cloud Security Management as the margin layer. The motion is *land-with-APM, expand-with-Logs+Security+AI* — a textbook consumption flywheel where every new container, every new log line, and every new LLM trace meters automatically against an enterprise commit. By FY27 the model is producing ~$4.2-4.5B in subscription revenue (~25% YoY off an ~$3.4-3.5B FY26 guide), with 96%+ of revenue subscription and effectively zero pro-services drag. The cash is concentrated in the $1M+ ARR club (~340 customers in 2026, on track for ~450 by FY27 close) — those accounts carry 8-12 modules each and drive the 115%+ NRR that Wall Street prices off. The four module groups that print most of the cash — APM, Logs, Infrastructure, Security — together represent ~85%+ of revenue, with Bits AI per-investigation and LLM Observability per-trace the emerging 2027 lines analysts are watching for the first standalone breakouts.

The Revenue Stack (How Each Dollar Is Metered)

• APM (Application Performance Monitoring) — Per-host, per-month. The anchor SKU since 2017. ~$31-36/host/month list, discounted to $18-25 at enterprise commit. Every Kubernetes pod, every EC2 instance, every Fargate task = a billable host.
• Logs Management — Per-GB ingested + per-million events indexed. Two-tier: ingest cheap, index (search/retain) expensive. The volume printer — customers routinely 5-10x their log spend in year 2.
• Infrastructure Monitoring — Per-host, per-month. The original 2010 product. ~$15-23/host/month. Often bundled free-with-APM at enterprise tier to drive land.
• Cloud SIEM + Cloud Security Management — Per-event analyzed + per-GB scanned. Highest gross margin module group (~85%+) because compute is incremental on already-ingested logs.
• Application Security Management (ASM) — Per-host, per-month — sits on top of APM agents, so zero deployment cost for customer = high attach rate.
• RUM + Synthetic Monitoring — Per-session (RUM) + per-test-run (Synthetic). Front-end telemetry layer — smaller revenue but high attach in e-commerce + media verticals.
• LLM Observability — Per-trace + per-token-monitored. Brand-new 2025 GA, expected to be the fastest-growing line on the price sheet through 2027.
• Bits AI (Investigations + Dev Agent + SRE Agent) — Per-investigation consumption + per-seat for Dev Agent. The 2026-2027 monetization pivot — Datadog's AI agents bill against the same enterprise commit as the rest of the platform.
• Database Monitoring + Network Performance Monitoring + Serverless — Per-host / per-database / per-function adders. Mid-cycle expansion modules.

The "$1M+ Customer" Math

• ~30,000 total customers as of Q1 FY26; ~3,800 spending $100K+ ARR; ~340 spending $1M+ ARR.
• The $1M+ club is <2% of customer count but ~50%+ of revenue — classic enterprise SaaS concentration.
• Named anchor accounts publicly disclosed: Atlassian, Toyota, Activision Blizzard, Samsung, Comcast, Peloton, Whole Foods. Atlassian and Samsung are believed to be in the $20M+ ARR tier based on case-study scale.
• Average $1M+ customer carries 8-12 modules vs. 2-4 for sub-$100K accounts — every additional module compounds gross margin because the agent + backend are already deployed.
• FY27 trajectory: ~450 $1M+ customers, ~5,000 $100K+ customers, with the largest single account believed to be approaching $50M ARR by year-end.

The Module Mix (FY26 Estimate → FY27 Trajectory)

• APM + ASM — ~35-40% of revenue. The flagship anchor — slowing in growth (~20%) but still the largest line.
• Logs Management — ~20-25% of revenue. Growing 30-35% — the volume tailwind from microservices + AI agents emitting torrents of logs.
• Infrastructure Monitoring — ~15-20% of revenue. Mature, growing ~15% — the loss-leader-bundled-with-APM dynamic compresses standalone growth.
• Cloud SIEM + CSM + ASM (Security group) — ~10-15% of revenue, growing 40%+. Crossing 15% of revenue mix is the 2027 milestone analysts watch.
• RUM + Synthetic Monitoring — ~5-8% of revenue. Steady 20% growth.
• Bits AI + LLM Observability — currently sub-1%, expected to be the first revenue line *broken out separately* in earnings releases by mid-2027.

The Cost Structure (Where The Dollars Go)

• R&D ~25% of revenue — ~2,200+ engineers across NYC, Paris, Boston, Tel Aviv. Module velocity is the moat — Datadog ships 3-5 net-new product lines per year.
• Sales & Marketing ~30% of revenue — efficient by SaaS standards (peers run 40-50%). PLG-assisted enterprise motion: agents go viral inside an account before sales arrives.
• Cost of Revenue ~20% — of which ~12-14% is cloud infrastructure (AWS primary, Azure + GCP for multi-cloud customers). The single biggest line item is AWS data transfer + S3 + EC2 for ingest + indexing.
• Gross Margin ~80-81% — the religious line. Olivier Pomel has stated publicly that ~80% is the floor he will defend; AI-inference workloads (Bits AI, LLM Obs) are priced explicitly to maintain that floor.
• Free Cash Flow Margin ~30%+ — best-in-class SaaS efficiency. The Rule-of-50+ profile (growth + FCF margin) is what earns the premium multiple.

What Changes In 2027

• Bits AI consumption breakout — Datadog likely begins disclosing AI-agent revenue as a separate line, the way Snowflake disclosed Cortex.
• Cloud SIEM crosses 10%+ of revenue standalone — graduating from "emerging" to "named segment" in earnings.
• LLM Observability becomes a real revenue line — every customer running production LLMs needs prompt + token + hallucination monitoring; Datadog's bet is to own that telemetry layer.
• Named-vertical solutions — Datadog for Financial Services, Datadog for Healthcare, Datadog for Public Sector — pre-bundled module SKUs at vertical-specific price points.
• The Flex Logs tier matures — cold-storage log retention at <$0.10/GB/month, designed to recapture the workloads customers were sending to S3 + Athena to escape Datadog Logs sticker shock.
• Per-employee Dev Agent seat pricing — the first time Datadog charges by *seat* rather than by *infrastructure unit*, opening a TAM the consumption model never touched.

The 80% Gross Margin Guard-Rail

• Olivier Pomel (CEO) treats 80% gross margin as a non-negotiable floor, publicly defended on every earnings call since 2021.
• AI-inference workloads (Bits AI, LLM Obs) are priced with a cost-pass-through buffer — Datadog's gross margin on AI features is structured to be no worse than 75% even at peak GPU spot prices.
• Multi-cloud commits (AWS + Azure + GCP) give Datadog negotiating leverage — they publicly disclosed an $800M+ AWS commit in 2024, in exchange for marketplace co-sell + EDP credits.
• Reserved-instance arbitrage on the ingest tier — Datadog buys 3-year reserved compute against a workload they can forecast within ±5%, capturing 30-40% savings vs. on-demand.
• The compression risk is AI inference cost volatility — if frontier-model API prices spike, the LLM Obs and Bits AI margins are the soft spot; Pomel's hedge is to push customers toward smaller, fine-tuned in-house models monitored *by* Datadog rather than served by it.

Revenue Stream Table

Money Flow Diagram

Bottom Line

Datadog's 2027 P&L is a consumption-pricing flywheel disguised as enterprise SaaS — per-host APM lands the account, per-GB Logs prints the volume revenue, per-event Security carries the margin, and Bits AI + LLM Observability are the new TAM-expanding lines that justify the premium multiple. The ~$4.2-4.5B revenue target is built on ~450 $1M+ customers carrying 8-12 modules each, defended by an 80% gross-margin floor Pomel will not let slip. The structural bet is that every dollar of new AI infrastructure spend in 2027 throws off telemetry, and Datadog owns the meter on that telemetry.

See also: [q1669](/answers.html?q=q1669) (Datadog's competitive moat) - [q1671](/answers.html?q=q1671) (Datadog vs. Splunk vs. New Relic) - [q1673](/answers.html?q=q1673) (Datadog's AI strategy)