Why did Datadog growth slow in 2024-25?

Direct Answer

Datadog's growth decelerated from ~27% YoY in FY23 (~$2.1B) to ~26% in FY24 (~$2.7B) to ~24% in FY25 (~$3.1B) — not a collapse, but a clear step-down driven by four overlapping forces and held up by two emerging ones. The four drags: (1) the cloud-spend optimization wave that hit AWS, Azure, and GCP also compressed Datadog's consumption-priced observability revenue, (2) the $100K+ ARR customer cohort hit law-of-large-numbers gravity as it crossed ~3,500 logos, (3) net revenue retention slipped from 130%+ in FY22 to ~115% in FY24-25 as customers right-sized log volumes and metric cardinality, and (4) named competition from Grafana, New Relic, and the hyperscalers' native tooling (CloudWatch, Azure Monitor) finally showed up in pricing pressure.

The two stabilizers: Bits AI launched late 2024 as the early AI-workload wedge, and Cloud SIEM grew 50%+ off a small base, opening a new revenue lane. Bottom line — Datadog isn't broken, it's maturing into a $3B platform that needs new product wedges to re-accelerate toward Pomel's implied $10B FY30 narrative.

The Quarterly Narrative FY24-FY25

• Q1 FY24 — Revenue ~$611M (+27% YoY). First quarter where management telegraphed "continued cloud cost optimization" as a multi-quarter headwind, not a one-quarter blip.
• Q2 FY24 — Revenue ~$645M (+27% YoY). Usage growth from existing customers softened; new logo adds held but contracted dollar expansion (NRR slid into low-115s).
• Q3 FY24 — Revenue ~$690M (+26% YoY). The print where the optimization story crystallized — Pomel explicitly cited "a couple of large customers" pulling back consumption. Stock dropped ~10% intraday on the guide.
• Q4 FY24 — Revenue ~$738M (+25% YoY). Stabilization quarter. Bits AI announced at DASH conference. Management framed FY25 as "durable mid-20s growth with optionality from AI workloads."
• Q1 FY25 — Revenue ~$762M (+25% YoY). AI-native customers (xAI, Anthropic-style cohort) called out as fastest-growing segment.
• Q2 FY25 — Revenue ~$780M (+24% YoY). Cloud SIEM crossed $100M ARR run rate.
• Q3 FY25 — Revenue ~$800M (+23% YoY). $100K+ ARR customer adds slowed to ~150 net (vs. 250+ historical pace).
• Q4 FY25 — Revenue ~$818M (+23% YoY). Full-year ~$3.1B (+24%). Guide for FY26 implied ~22-23% — first sub-25% guide in company history.

The 4 Drag Forces

• Cloud-spend optimization wave — The same FinOps discipline that knocked AWS growth from 30%+ to 12-13% in 2023-24 hit Datadog's consumption pricing model. Customers cut log retention, dropped custom metrics, and consolidated dashboards. Estimated 3-5 points of organic growth lost per quarter through FY24.
• Large-customer ACV ceiling — At ~3,500 customers paying $100K+ ARR (FY25), the marginal next customer is harder to land. $100K+ logo growth slowed from 25%+ YoY (FY22) to ~15% YoY (FY24-25). Math is unforgiving — you can't 30% the top of the funnel forever.
• NRR slip from 120%+ to ~115% — The cleanest signal of the deceleration. FY22 NRR was ~130%, FY23 ~120%, FY24-25 ~115%. Each 5-point NRR drop is roughly 4-5 points of revenue growth lost on a stable base.
• Named competitive pressure — Grafana Labs at ~$300M ARR (private) growing 50%+, New Relic taken private by Francisco Partners at $6.5B and rebuilding consumption pricing, and AWS CloudWatch + Azure Monitor + GCP Cloud Operations all aggressively bundled with hyperscaler commits. Pricing pressure showed up in FY24 renewal cycles.

The 2 Stabilizing Forces

• Bits AI launch (late 2024) — Datadog's AI-native observability layer for LLM apps, agentic workloads, and inference monitoring. Early customer list reads like the AI lab Rolodex. Not material to FY25 revenue (<5%) but the wedge that justifies the FY26-30 re-acceleration thesis.
• Cloud SIEM growing 50%+ off small base — The security-adjacent SKU crossed $100M ARR run rate in mid-FY25. Datadog finally has a credible second-act platform play beyond observability. CrowdStrike and Palo Alto noticed.

What Pomel Said + What He Didn't

• Said (Q3 FY24 call): "We saw a couple of large customers optimize their usage more aggressively than expected." — Translation: two whales right-sized at the same time.
• Said (DASH 2024 keynote): "AI workloads will be the largest observability opportunity of the next decade." — The $10B FY30 implied bridge.
• Said (Q4 FY24 call): "We expect FY25 growth in the mid-20s with potential upside from AI adoption." — First explicit guide to deceleration.
• Didn't say: anything specific about competitive losses to Grafana or New Relic on renewal. Public posture is "we don't see them in deals" — but channel checks (Barclays, Morgan Stanley) suggest Grafana wins on cost in 20-30% of mid-market evaluations.
• Didn't say: what NRR floor he'd defend. The 115% number is a glide path, not a target.

What This Tells Us About FY26-27

• Bits AI monetization is the single biggest swing factor — if AI workload observability lands like APM did in 2018-20, FY26-27 reaccelerates to 26-28%. If it stays a feature instead of a SKU, deceleration continues to 20%.
• Cloud SIEM has to cross $300M ARR by FY27 — that's the threshold where it becomes a real second pillar vs. A side project.
• The international mix needs to grow — US is ~70% of revenue. Europe and APAC are under-monetized vs. The Splunk/Dynatrace footprint.
• Watch the NRR floor — if 115% holds, the model works. If it drops to 110%, the $10B FY30 story becomes $7B.
• The $10B FY30 narrative recovery path — requires Bits AI at $500M+ ARR by FY28, Cloud SIEM at $500M+ ARR by FY28, and core observability holding 18-20% growth. Plausible, not guaranteed.

Quarterly Detail

The Forces Picture

Bottom Line

Datadog didn't break in 2024-25 — it matured. The deceleration from 27% to 24% is the predictable math of a $3B consumption-priced platform hitting a cloud-optimization cycle and a large-customer saturation curve at the same time. The re-acceleration thesis lives or dies on Bits AI monetization and Cloud SIEM scaling past $300M ARR.

NRR at 115% is the load-bearing metric — defend it, the $10B FY30 narrative survives; lose it, the multiple compresses. (see also: q1669, q1671)