Published analyst estimates for the observability market range from $2.4B to $5.4B, but these figures are difficult to reconcile with vendor revenue. Datadog alone reported $3.43B in FY2025; Dynatrace reported $1.70B; and Splunk’s observability-adjacent revenue, Elastic’s observability segment, Grafana Labs’ $400M+ ARR, and cloud vendor observability services (AWS CloudWatch is estimated at ~$1.6B alone) push the combined total well past any narrow estimate.
The gap exists because analyst firms scope narrowly. MarketsandMarkets defines the market at $2.4B (2023) → $4.1B (2028). Grand View Research uses a broader scope at $5.4B by 2030. But both exclude significant vendor revenue that falls outside their taxonomy – security products, log analytics, cloud-native monitoring bundled into IaaS, and CI/CD observability.
A more credible figure comes from a bottom-up analysis by Dash0 CEO Mirko Novakovic, who aggregated revenue from 28 observability vendors (public filings, disclosed ARR, and conservative estimates for private companies) and arrived at approximately $12B in 2024, growing at ~20% annually. A top-down cross-check – observability spending as 15–20% of the $261B global cloud infrastructure market – suggests a total addressable market of $40–50B, of which $12B is captured by commercial software vendors. Gartner estimates a 12% CAGR for the observability sector within infrastructure software through 2027.
For perspective on where observability sits relative to adjacent markets: cybersecurity is ~$220B (Fortune Business Insights), CRM is ~$90–113B (Mordor Intelligence), cloud infrastructure services (IaaS+PaaS) exceed $400B (Statista), and the DevOps tools market is $15–18B (Mordor Intelligence). Observability at ~$12B sits within and alongside DevOps, with a higher growth rate and AI observability’s 36% CAGR among the fastest in enterprise infrastructure software.
Growth is driven by cloud-native complexity, Kubernetes adoption, the proliferation of microservices, and the emerging requirement to monitor AI workloads. Observability spend has become a top-three cloud cost line item for many enterprises, which simultaneously fuels demand and creates pricing sensitivity – a tension visible across the vendor landscape.
The 2025 Gartner Magic Quadrant for Observability Platforms evaluated 20 vendors and named six Leaders: Datadog (fifth consecutive year), Dynatrace (fifteenth consecutive year, highest on Ability to Execute), Elastic, Grafana Labs, IBM, and Splunk. Grafana Labs appeared as a Leader for the first time, reflecting the mainstreaming of open-source observability.
2. Vendor Revenue
The public vendors reveal a market where the leader is pulling away on absolute revenue while an open-source challenger grows fastest by percentage.
Datadog
FY2025 revenue $3.43B, up 28% YoY. Q4 revenue $953M, up 29%. Record Q4 bookings of $1.63B included two deals above $100M TCV; one described as an eight-figure land with a leading AI model company. 4,310 customers above $100K ARR representing ~90% of total ARR. Guided FY2026 at $4.06–4.10B. (earnings release; 10-K)
Dynatrace
FY2025 (ending March 2025) revenue $1.70B, up 19% (20% constant currency). Subscription revenue $1.62B. ARR $1.73B. FY2026 guidance raised to $1.99–2.00B revenue and $2.01–2.03B ARR. (FY2025 results; Q2 FY2026)
Grafana Labs
Surpassed $400M ARR in September 2025, up from $250M in August 2024 (~60% growth). 7,000+ customers including 70% of the Fortune 50. Raised $250M Series E at ~$9B valuation in February 2026. (press release; SiliconANGLE)
Splunk / Cisco
Acquired by Cisco for $28B (March 2024). Pre-acquisition ARR ~$4.2B, heavily weighted toward SIEM/log analytics. AppDynamics merged into Splunk’s observability unit. (Cisco IR)
New Relic
Taken private by Francisco Partners and TPG for $6.5B (November 2023). Last reported ARR ~$1.0–1.1B. Limited visibility since. (press release; CNBC)
3. Consolidation
The mid-tier of the market is being absorbed. Private equity and strategic acquirers have executed landmark transactions that reshaped the competitive landscape. The pattern bifurcates the market into large platforms (Datadog, Dynatrace, Cisco/Splunk) and open-source or OTel-native challengers (Grafana Labs, Dash0, Chronosphere). The middle ground – proprietary but sub-scale – is increasingly untenable.
Cisco / Splunk
$28B (March 2024). Created the largest combined observability + SIEM revenue base. Cisco merged AppDynamics and ThousandEyes into Splunk’s unit. (Network World)
New Relic
$6.5B take-private (November 2023). Consumption-pricing pivot had gained traction but created revenue volatility that public markets punished. (CNBC)
Observe
~$1B acquisition by Snowflake (announced January 2026). Snowflake’s largest deal; signals data platform companies see observability as a natural extension. (TechCrunch)
4. Cost Pressure
Observability cost has become a first-order concern. Datadog’s multi-dimensional pricing model – per host for infrastructure, per GB for logs, per span for APM, per custom metric, per RUM session – creates billing complexity that compounds at scale. Reports of bills growing 30–50% year-over-year are common. Teams have reported optimizing their infrastructure architecture to minimize Datadog host counts rather than for operational efficiency – a sign the billing model is distorting technical decisions.
This dynamic is a direct tailwind for three categories: telemetry pipelines (Cribl, which claims 30–90% cost reduction by filtering data before it reaches backends), open-source backends (Grafana Labs, whose LGTM stack removes licensing cost), and OTel-native startups (Dash0, Groundcover, Coralogix) that position on transparent pricing. Datadog responded at DASH 2025 with Flex Logs, a flexible storage tier for cost optimization (DASH 2025), acknowledging the pressure without changing the underlying model.
Switching costs remain the moat: dashboards, monitors, and team knowledge built over years don’t export cleanly. Most teams split workloads across vendors rather than fully migrating – keeping Datadog for APM while moving logs and metrics to cheaper alternatives.
OpenTelemetry
5. The OpenTelemetry Project
OpenTelemetry is the second-highest-velocity project in the Cloud Native Computing Foundation, behind only Kubernetes. Formed in 2019 from the merger of Google’s OpenCensus and the CNCF’s OpenTracing, OTel defines a vendor-neutral standard for generating, collecting, and transmitting telemetry data. The project reports contributions from over 10,000 individuals across 1,200 companies. Approximately 900 developers from 200 companies contribute monthly – an 18% year-over-year increase in developer participation and a 22% rise in company involvement (CNCF Mid-Year 2025).
The top organizational contributors – Splunk, Microsoft, Lightstep (now ServiceNow), Google, and Dynatrace – account for roughly 60% of contributions. The remaining 40% comes from startups, individual contributors, and smaller vendors. OTel’s governance is structured to prevent any single company from dominating: no company can hold a majority of Governance Committee seats, and the specification evolves through an open RFC process.
OTel achieved CNCF Graduated status on January 31, 2024 (CNCF), joining Kubernetes, Prometheus, and Envoy at the highest maturity level. The project processes an estimated 4 quadrillion telemetry signals per week across adopters. The CNCF launched the OpenTelemetry Certified Associate (OTCA) certification in November 2024 – a further indicator of ecosystem maturity and enterprise readiness.
6. Signal Maturity
OTel organizes telemetry into signals, each at a different stage of maturity (status page). The three core signals have reached stable/GA status. Two newer signals are in earlier stages.
Traces
Stable. The most mature signal, with broad production adoption since 2021–2022.
Metrics
Stable. GA across major language SDKs (Java, .NET, Python, Go, JavaScript).
Logs
Stable. The Logs Bridge API connects existing log frameworks (Log4j, SLF4J, Python logging) into OTel.
Profiling
Public Alpha (March 2026). Uses eBPF-based agents. Cross-signal correlation with traces via span IDs in development. Based in part on technology donated by Elastic.
Events
Experimental. Structured, named log records with well-defined semantics, built as a semantic layer on top of logs.
A notable gap: client-side instrumentation for browsers and mobile remains experimental and mostly unspecified. Core Web Vitals support is in progress, and mobile SDKs (Android, iOS) are early-stage. This is one area where proprietary agents (Datadog RUM, Dynatrace RUM) retain a clear advantage and represents a near-term ceiling on OTel’s commoditization reach.
48% currently using OpenTelemetry
25% planning to implement
25% evaluating
61%+ rate OTel as “very important” or “critical”
46% of adopters report >20% ROI
40% of adopters report 10–20% ROI
Complexity, cost, and resource constraints were cited as the primary barriers. These findings describe a technology that has crossed the early-adopter phase but has not yet reached ubiquity – roughly half the market is in, and nearly all of the other half is looking at it. The Grafana Labs OTel Report found similar adoption patterns, with organizations using OTel reporting improved collaboration between development and operations teams as a secondary benefit.
European regulatory pressure is an additional adoption driver. Cumulative GDPR fines have reached 7.1 billion euros since 2018. Three EU regulations now intersect with observability data: DORA (financial services, effective January 2025), the EU Data Act (non-personal/industrial data, effective September 2025), and the NIS2 Directive (critical infrastructure). These frameworks penalize structural control deficiencies including weak vendor management and inadequate logging. OTel’s vendor-neutral, self-hosted collection model – and architectures like Groundcover’s “bring your own cloud” – directly address sovereignty requirements that SaaS-only platforms struggle with.
Competitive Dynamics
8. Collection Commoditization
OpenTelemetry’s core structural effect is the commoditization of telemetry collection. Before OTel, instrumenting with one vendor’s agent meant telemetry was in that vendor’s format, sent to that vendor’s endpoints. Switching required re-instrumenting everything – a tax that functioned as lock-in.
With OTLP as a de facto wire-format standard, organizations instrument once and route data to any compatible backend by changing an exporter configuration. The OTel Collector can fan data to multiple backends simultaneously, enabling parallel evaluations and gradual migrations. Switching costs drop from months to days.
This shifts differentiation from how you collect data to what you do with it: analytics, AI-driven automation, query performance, UX, and platform breadth. As The New Stack notes, OTel has lowered the barrier to entry for observability startups – they compete entirely on backend value because the collection layer is open.
9. Vendor Positioning
Each major vendor has adopted a distinct posture toward OTel, reflecting their architectural history and business model.
Datadog
Embrace and extend. Accepts OTLP via the Datadog Agent and a direct intake API. Launched the Datadog Distribution of the OTel Collector (DDOT) at DASH 2025. Supports OTel-native metrics in dashboards via Semantic Mode. But the proprietary agent remains the recommended path – it is the upsell surface for NPM, security, profiling, and database monitoring. Features like Continuous Profiler code hotspots, Dynamic Instrumentation, and Data Streams Monitoring require dd-trace, not OTel SDKs. Financially, the strategy is working: net revenue retention held at ~120% in Q3 2025, and an eight-figure expansion deal was closed with a customer that standardized on Datadog APM using OpenTelemetry – 17 products adopted, 40% reduction in resolution times (earnings call). OTel is being absorbed into the expansion playbook, not causing churn.
Dynatrace
Governance leader and hybrid model. Contributes to the OTel Technical Committee, specifications, and Collector. Ships a production-ready OTel Collector distribution for free. Grail data lakehouse stores OTel data natively without lossy translation. DQL operates directly on OTel attributes. OneAgent + OTel hybrid positions OTel as complementary to automatic instrumentation.
Grafana Labs
Structurally aligned. Backend stack – Tempo (traces), Mimir (metrics), Loki (logs), Pyroscope (profiling) – accepts OTLP natively. Grafana Alloy (formerly Grafana Agent) is built on the OTel Collector framework. As collection commoditizes, Grafana becomes the default open-source backend.
Honeycomb
OTel-native from inception. Co-founded by OTel contributors. Strongest OTel-native query experience, but niche revenue (~$10.8M).
Splunk / Cisco
Deep OTel roots via SignalFx (whose founders co-created OTel). Splunk Observability Cloud is OTLP-native. Post-Cisco integration introduced agentic AI-powered observability (September 2025).
10. Cloud Provider Strategies
All three major cloud providers now support OTLP ingestion natively, each with a different strategic calculus.
Google Cloud
Instrumental in creating OTel (via OpenCensus/OpenTracing merger). Native OTLP ingestion in Cloud Trace via telemetry.googleapis.com. Strategic interest: OTel reduces switching costs, helping Google compete against AWS lock-in. (InfoQ; Google Cloud Blog)
Microsoft Azure
Recommends the Azure Monitor OpenTelemetry Distro as primary instrumentation path. Actively migrating from legacy Application Insights SDKs. Added AI agent monitoring via OTel from Foundry, Copilot Studio, LangChain, and OpenAI Agents SDK.
AWS
Ships ADOT (AWS Distro for OpenTelemetry). Built CloudWatch Application Signals – APM-like service maps and SLO monitoring – on OTel data. Strategy: “good enough, deeply integrated, and included.”
For third-party vendors, this is double-edged. It validates OTel as the standard and makes multi-cloud practical – positive for Datadog. But it enables “good enough” built-in observability that compresses the low end of the market. The marketplace channel (Datadog is a top seller on AWS Marketplace, with purchases counting toward committed cloud spend) provides commercial cushion but also creates dependency.
11. OTel Beyond APM
OTel’s expansion beyond application tracing into infrastructure, Kubernetes, and fleet management creates new competitive vectors.
Kubernetes
The OTel Operator manages Collectors and auto-instrumentation via CRDs. Receivers for kubelet stats, cluster state, and k8s objects provide infrastructure telemetry. Supports DaemonSet, sidecar, and deployment patterns. (docs)
Infrastructure
Host metrics receiver provides CPU, memory, disk, and network collection. Integration breadth lags proprietary agents (Datadog’s 750+ integrations vs. ~100+ OTel receivers), but the gap narrows each release.
Fleet Management
OpAMP (Open Agent Management Protocol) enables remote configuration, health monitoring, and updates for Collector fleets. Nike presented an enterprise-scale implementation at KubeCon NA 2025. Once mature, OpAMP makes Collectors as manageable as proprietary agents.
CI/CD
Semantic conventions for CI/CD pipelines (pipeline runs, task runs, VCS attributes) enable treating build systems as distributed traces. The Jenkins OTel Plugin and Gradle/Maven OTel plugins are production-ready. Emerging but meaningful for platform engineering teams.
Architectural Implications
12. eBPF and OTel Convergence
A significant development in 2025 was the convergence of eBPF-based instrumentation with OpenTelemetry. The OTel eBPF Instrumentation project (OBI) – originally Grafana Beyla, donated by Grafana Labs – provides zero-code, kernel-level auto-instrumentation for Linux HTTP/S and gRPC services. The first alpha release was a collaboration between Grafana Labs, Splunk, Coralogix, and Odigos.
OBI captures distributed traces and RED (Rate, Errors, Duration) metrics without any application code changes, operating at the kernel level via eBPF. Database instrumentation covers PostgreSQL, MySQL, MongoDB, Redis, and others. The OTel eBPF SIG was launched in May 2025 and is targeting a stable 1.0 release with expanded protocol and language support.
This matters because it addresses the primary friction point of OTel adoption – manual instrumentation. If eBPF can deliver “Dynatrace-like” zero-code visibility through an open standard, the value proposition of proprietary agents narrows further. Grafana Labs articulated the architecture clearly: eBPF for automatic baseline visibility, SDKs for application-specific depth.
13. Data Model Translation
OTel uses a hierarchical attribute model with three layers: resource attributes describing the entity (service, host, container), scope attributes identifying the instrumentation library, and signal-level attributes on individual spans, metrics, or logs. Attributes use dot-notation semantic conventions:
Vendors with flat tag models must collapse this hierarchy. Datadog’s key:value tags lose the resource/span distinction – everything becomes a flat tag. Datadog maintains a semantic mapping layer, but the mapping is lossy: typed OTel attributes (int, double, bool, array) are coerced to strings, and resource identity – a first-class OTel concept – has no Datadog equivalent.
A second friction point: OTel SDKs default to cumulative metric temporality in many configurations. Datadog’s OTLP intake requires delta temporality, forcing SDK or Collector-level conversion.
Vendors with schema-on-read backends preserve the hierarchy natively. Dynatrace’s Grail and Grafana’s Tempo store OTel data without translation, making resource vs. span distinction queryable.
14. Query Fragmentation
There is no OTel-standard query language; OTel explicitly scopes itself to generation, collection, and transport. Each vendor fills this gap differently, and OTel-native query capability is becoming a meaningful differentiator.
Datadog
Multiple syntaxes: a DSL for metrics, Lucene-style for logs and traces, and DDSQL (SQL dialect). Semantic Mode combines equivalent OTel and Datadog metrics. No structural trace queries. Facet declaration required for logs.
Dynatrace
DQL: single pipe-based language across all signals. Operates natively on OTel attributes. No translation layer. Unified across logs, traces, metrics, events, and business data.
Grafana
Signal-specific: PromQL, LogQL, TraceQL. TraceQL was designed for OTel – supports resource. vs. span. distinction and structural parent/child queries.
An OTel-native query experience would support: dot-notation attributes as first-class citizens, typed values without facet declaration, resource/span separation, semantic convention version awareness, and structural trace queries. No vendor fully delivers all of these. Dynatrace (DQL) and Grafana (TraceQL) are closest.
15. Security Convergence
The boundary between security and observability is blurring. OTel can bridge the gap by sending enriched traces and logs to SIEM platforms alongside observability backends, giving security analysts application context – which user performed an action, what service handled it, the duration and outcome – without leaving their tools. As Elastic notes, observability pipelines increasingly facilitate runtime detection, anomaly detection for lateral movement, container integrity signals, and network policy telemetry.
Datadog has leaned into this with Cloud SIEM, Application Security Management, and Cloud Security Management – all flowing through the Datadog Agent. Elastic positions across both observability and security natively. Splunk’s value to Cisco was largely the SIEM + observability convergence. This trend makes the OTel Collector strategically important for security teams as well – a single collection pipeline for both operational and security telemetry.
OTel and AI
16. OTel GenAI Conventions
OpenTelemetry has published semantic conventions for generative AI systems, covering model spans, agent spans, events, and metrics. Technology-specific conventions exist for Anthropic, OpenAI, Azure AI Inference, and AWS Bedrock.
Work on agent framework conventions – covering CrewAI, AutoGen, LangGraph, IBM Bee Stack, and others – is ongoing. The community is defining a standardized approach for tracing multi-step tool use, RAG, and multi-agent orchestration. Datadog natively supports OTel GenAI Semantic Conventions (v1.37+), auto-mapping attributes to its LLM Observability schema.
If these conventions achieve the same adoption trajectory as OTel’s HTTP or database conventions, they will standardize AI workload telemetry across providers – repeating the pattern that transformed APM from a proprietary-instrumentation market to an open-standards one.
17. AI for Observability
Every major vendor now ships AI-assisted features. Two distinct architectural approaches have emerged – one topology-aware, one correlation-based.
Datadog
Bits AI suite (2025): SRE Agent (autonomous troubleshooting), Dev Agent, Security Analyst. LLM-driven, correlation-based. In 2026, launched an MCP Server enabling coding agents (Claude Code, Cursor, Codex) to access telemetry during development. (DASH 2025)
Dynatrace
Davis AI: causal, topology-aware reasoning using the Smartscape dependency graph. Longer track record in production AI – 15 years of the Davis engine precede the current LLM wave.
Cisco / Splunk
Agentic AI-powered Splunk Observability (September 2025). AI agents across the full incident response lifecycle.
Microsoft
AI agent monitoring in Application Insights, with OTel-collected telemetry from Foundry, Copilot Studio, LangChain, and OpenAI Agents SDK. (docs)
18. Observability for AI
The inverse: monitoring AI/ML workloads themselves. LLM inference latency, token costs, hallucination rates, GPU utilization, model drift, and AI agent execution traces. This is the category driving the highest growth rates in observability.
Datadog’s LLM Observability provides end-to-end tracing across AI agents. The June 2025 expansion added agentic AI monitoring, LLM Experiments, and an AI Agents Console. A Q4 2025 eight-figure deal with “a leading AI model company” (earnings call) signals that AI infrastructure operators are emerging as a material revenue segment.
Coralogix acquired Aporia (AI observability and guardrails platform) in January 2025 and launched the Coralogix AI Center three months later (blog). The pattern is clear: traditional observability vendors are buying AI monitoring capabilities rather than building them organically.
19. AI Observability Market
The AI observability sub-market is the fastest-growing segment in observability, with growth rates two to three times the overall market.
Growing by $2.9B between 2025–2029, 22.5% CAGR. Broader scope including AIOps. (Technavio)
These figures are driven by two forces: every enterprise deploying AI needs to monitor it, and AI workloads generate novel telemetry requirements (token economics, prompt/response quality, agent decision traces) that existing platforms were not designed to handle.
Emerging Players
20. Emerging Areas
Six distinct areas have emerged where startups are clustering and capital is concentrating. Each represents a gap in the incumbent platforms, an expansion of what “observability” means, or both.
OTel-native backends compete directly with Datadog and Dynatrace on the premise that open-standard collection plus a purpose-built backend delivers comparable value at lower cost and without lock-in. Dash0’s $1B valuation in under three years is the proof point for this thesis.
eBPF zero-code instrumentation attacks the primary friction point of OTel adoption – that SDK-based instrumentation requires developer effort. If kernel-level eBPF can deliver baseline visibility through an open standard, the value proposition of proprietary agents narrows. The OTel OBI SIG (Grafana, Splunk, Odigos, Coralogix) is the community effort; Groundcover is the funded commercial play.
AI agent tracing is the most novel category. Existing observability tools were designed for request/response architectures. AI agents that run for hours, use tools, retrieve documents, and make branching decisions require new tracing primitives – session replay, decision-path visualization, trajectory anomaly detection. Laminar, InfiniteWatch, and Helicone are early entrants.
LLM evaluation and monitoring is adjacent to but distinct from traditional observability. Arize and Braintrust focus on model quality (hallucination rates, evaluation benchmarks, prompt/response scoring) rather than infrastructure metrics. This is largely a greenfield market – incumbents are building but startups are further along.
Telemetry pipelines sit between producers and consumers of observability data. They exist because observability costs have become untenable – Cribl’s entire $3.5B valuation rests on the fact that enterprises pay too much to store and query telemetry they don’t need.
K8s-native + AI SRE platforms combine Kubernetes-specific telemetry with AI-driven troubleshooting. The bet is that Kubernetes environments are complex enough and distinct enough from traditional infrastructure to warrant purpose-built tools rather than general-purpose platforms with K8s add-ons.
21. Recent Acquisitions
Seven observability-related acquisitions in the twelve months preceding this report – an unprecedented pace – reveal which capabilities buyers are prioritizing and which company types are acquiring.
Observe
Snowflake, ~$1B (January 2026). Snowflake’s largest acquisition. Data platform companies see observability as a natural extension of their storage layer. (TechCrunch)
Langfuse
ClickHouse (January 2026). Open-source LLM tracing (YC W23, MIT-licensed, 2K+ paying customers including 19 Fortune 50). Database infrastructure companies are buying AI observability. (GitHub)
Traceloop
ServiceNow, $60–80M (March 2026). Creators of OpenLLMetry, the open-source OTel-native LLM tracing library. Raised only $6.1M – a ~10–13x return for investors. OpenLLMetry remains open source. IT platform companies are buying OTel-native AI observability.
Helicone
Mintlify (March 2026). LLM gateway and observability (YC W23). 16K organizations, 2B+ LLM interactions. Developer tooling companies are acquiring AI observability. (Helicone)
Lumigo
Dash0 (February 2026). Serverless/AWS-native observability. Brought Lambda and LLM observability into Dash0’s OTel stack. (Dash0)
Highlight.io
LaunchDarkly (April 2025). Session replay + backend traces ($8M raised). Feature flag companies are acquiring observability for the developer workflow. (Highlight.io)
Aporia
Coralogix (January 2025). AI observability and guardrails. Led to the Coralogix AI Center launch three months later. (Coralogix)
The buyer profile is striking: a data warehouse (Snowflake), a database engine (ClickHouse), an IT workflow platform (ServiceNow), a developer docs company (Mintlify), a feature flag vendor (LaunchDarkly), and two observability incumbents (Dash0, Coralogix). Observability is being absorbed into adjacent platform layers, not just consolidated within its own market.
22. OTel-Native Startups
A cohort of startups built from the ground up on OpenTelemetry, with no proprietary agent or format. Their existence is a direct consequence of collection commoditization.
Dash0
$1B valuation, $155M total (Series B $110M, March 2026, Balderton). Founded 2023 by the ex-Instana team (Mirko Novakovic). 600+ paying customers including Zalando and Taco Bell. Doubled customer count in five months between rounds. Acquired Lumigo. Fastest-growing company on this list relative to age.
Coralogix
$1B+ valuation, $115M Series E (June 2025, TechCrunch). 4,000+ customers. Acquired Aporia for AI observability. Launched agentic AI assistant “Olly.” OTel-compatible with strong Kubernetes support. Targets enterprises wanting unified observability + security with AI-powered analysis.
Chronosphere
~$1.6B valuation (2023), $343M total (Series C $200M, General Atlantic). Ex-Uber engineers who built M3DB. Cost control differentiator – 100% trace retention without sampling. Slower recent news raises questions about growth trajectory; valuation may be stale.
Honeycomb
~$150M total (Series D $50M, 2023, Headline). Pioneered “observability vs. monitoring.” 160%+ NRR but ~$10.8M revenue. Revenue-to-funding gap makes standalone path difficult. More likely an acquisition target.
SigNoz
~$6.5M total (TechCrunch). Open-source OTel-native on ClickHouse. Direct Datadog alternative. Small and acquirable – Langfuse/ClickHouse pattern.
23. Kubernetes & Infrastructure
Groundcover
$60M total (Series B $35M, April 2025, SiliconANGLE). eBPF-based, K8s-purpose-built. “Bring your own cloud” model – telemetry stays in customer infra. 500%+ ARR growth. Fortune 100 customers. Zero-instrumentation via eBPF. (Network World)
Odigos
$13M raised (led by Venture Guides, angels from Chronosphere, Honeycomb, and Lightstep founders). eBPF + OTel auto-instrumentation. Key contributor to the OTel eBPF Instrumentation SIG. Targets the “zero-code distributed tracing” use case. (US Tech Times)
Metoro
YC-backed, $500K seed. K8s-native AI SRE. Three-person team generating $330K revenue (September 2025). Combines telemetry with AI-assisted troubleshooting specifically for Kubernetes. Early but capital-efficient.
Cribl is the dominant independent telemetry pipeline, positioned at the strategic chokepoint between data producers and observability backends. Cribl Stream routes, filters, and transforms data before it reaches the backend, enabling 30–90% cost reduction. $319M Series E (August 2024, GV) at $3.5B valuation. $200M ARR as of January 2025. (Crunchbase News; TechTarget)
At this scale, Cribl is an IPO candidate rather than an acquisition target. The pipeline layer’s strategic significance – controlling data flow, reducing costs, enabling multi-vendor strategies – makes it a company that Datadog competes against rather than acquires.
25. AI Observability Startups
Pure-play AI observability companies are drawing disproportionate capital relative to current revenue, reflecting a bet on the 36% CAGR forecast.
Arize AI
$131M total (Series C $70M, February 2025, Adams Street). Customers: Uber, Booking.com, Duolingo, PepsiCo, Wayfair. Datadog is a strategic investor – signaling acquisition interest or partnership hedge. Largest pure-play AI observability round at announcement.
Braintrust
$800M valuation, $80M Series B (February 2026, Iconiq with a16z and Greylock). “The observability layer for AI.” Evaluation, testing, and monitoring. Rapid valuation growth suggests strong demand for AI-specific tooling independent of APM vendors.
26. Early Stage & Long Tail
Below the growth-stage companies, a long tail of seed and pre-seed startups reflects where the next wave of innovation and acquisition targets is forming.
Laminar
YC S24, $3M seed (March 2026, Atlantic.vc). Open-source AI agent debugging with browser session replay synced to traces. Customers include Browser Use and OpenHands. Notable angel: Ben Sigelman (OTel co-creator). Purpose-built for AI agents that run for hours and generate thousands of spans.
Pydantic Logfire
$17.2M total (Sequoia-led Series A). OTel-native observability from the creators of Pydantic, the Python data validation library used by most major AI/ML frameworks. Distribution moat: Pydantic is installed 250M+ times/month. Logfire instruments Python AI workloads with zero added dependencies. A wedge play into AI observability via Python dominance.
Langtrace
$5.3M seed (Redpoint Ventures). OTel-native LLM observability with 30+ framework integrations. Positions on open standards where Arize and Braintrust use proprietary schemas.
InfiniteWatch
$4M pre-seed (December 2025, Base10). Ex-CoverWallet team. Monitors AI agents across voice and web. Session Replay Agent and Voice Agent products. Analyzes 2M+ customer interactions monthly. Targets agentic customer interaction intelligence.
OpenObserve
$3.6M seed (2022, Nexus, Dell Capital). Rust-based open-source observability. Claims 140x lower storage costs via Apache Parquet columnar storage. 2.6TB/day ingestion on single node. Positions as cost-efficient Datadog/Splunk alternative.
Axiom
$41.4M total (Series B, February 2025). Petabyte-scale, schema-less data backend for logs, traces, and events. Angels include Nat Friedman (ex-GitHub CEO) and Adam Wiggins (Heroku founder). Positions on unlimited data retention at fixed cost – the anti-Datadog pricing model.
$13M total (Series A led by Sequoia India / Peak XV). High-cardinality metrics platform (Levitate). 20M+ cardinality per metric – purpose-built for K8s environments with dynamic labels.
Embrace
Mobile-first OTel observability. Backed by NEA and YC. Customers include New York Times, Marriott, and Home Depot. Contributed the Kotlin OTel SDK upstream – filling a gap in OTel’s mobile coverage. One of the few startups addressing the RUM/mobile signal gap noted in Section 6.
DeepEval
$2.2M seed. Open-source LLM evaluation framework, reportedly the most widely adopted. Used by teams at OpenAI, Google, and Microsoft for model testing. Evaluation tooling is the adjacent layer to LLM observability – complementary acquisition target.
The pattern across early-stage companies: nearly all are OTel-native, eBPF-based, AI-agent-focused, or some combination. No funded startup in 2025–2026 has launched with proprietary-only instrumentation. And the acquisition clock is fast: Traceloop achieved a ~10–13x return in under three years; Helicone and Langfuse were acquired within two years of YC graduation.
27. Acquisition Patterns
Three patterns are visible across the startup and M&A landscape.
OTel-native is table stakes. Every funded startup is built on OTel or at minimum OTLP-compatible. Proprietary-only instrumentation is no longer fundable for new entrants.
AI observability draws disproportionate capital. Arize ($131M), Braintrust ($800M valuation), the Langfuse acquisition, and the Aporia acquisition all reflect a bet on the AI workload monitoring opportunity. Capital is flowing ahead of revenue.
Everyone is buying, not building. Seven acquisitions in twelve months – by a data warehouse, a database engine, an IT workflow platform, a developer docs company, a feature flag vendor, and two observability startups – demonstrate that the build-vs-buy calculus has tipped across the entire software stack. The velocity of the AI market and the architectural specificity of OTel-native backends make organic development too slow. Traceloop’s ~10–13x return on $6.1M raised, achieved in under three years, sets the benchmark for what small OTel-native AI observability startups can exit for.
Dash0
Medium-term. Growing too fast to sell cheap. Likely: Cisco, IBM, ServiceNow.
Coralogix
Medium. $1B unicorn with AI + security angles. Likely: Cisco, Oracle, SAP.
Chronosphere
Medium. Growth unclear, valuation may be stale. Likely: Cisco, Oracle.
Honeycomb
High. Revenue gap makes standalone path hard. Likely: Datadog, Cisco, ServiceNow.