Datadog

Datadog, Inc.

Type	Public
Traded as	Nasdaq: DDOG (Class A) Nasdaq-100 component S&P 500 component
Industry	Software
Founded	June 2010 (2010-06)
Founders	Olivier Pomel Alexis Lê-Quôc
Headquarters	New York City, U.S.
Area served	Worldwide
Key people	Olivier Pomel (CEO) Alexis Lê-Quôc (CTO)
Services	System monitoring
Revenue	US$3.43 billion (2025)
Net income	US$108 million (2025)
Total assets	US$6.64 billion (2025)
Total equity	US$3.73 billion (2025)
Number of employees	~8,100 (December 2025)
Website	datadoghq.com
Footnotes

Datadog, Inc. is an American software company that provides a SaaS-based observability and security platform for cloud applications. Its products collect and correlate metrics, traces and logs from servers, databases, applications, networks and cloud services, and are sold on a subscription basis. The company was founded in 2010 by Olivier Pomel and Alexis Lê-Quôc, is headquartered in New York City, and has been listed on the Nasdaq stock exchange (ticker symbol: DDOG) since September 2019.

For the fiscal year ended December 2025, Datadog reported revenue of $3.43 billion, approximately 8,100 employees across 35 countries, and about 32,700 customers. The company was added to the S&P 500 index in July 2025.

History

Founding

Datadog was incorporated in Delaware in June 2010 (the state of incorporation was later changed; see below). Its founders, Olivier Pomel and Alexis Lê-Quôc, both received master's degrees in computer science from École Centrale Paris.

Before founding Datadog, the two were colleagues at Wireless Generation, an education-technology SaaS company: Pomel served as vice president of technology from 2002, and Lê-Quôc worked there from March 2004 to December 2010 as director of Live Operations. Wireless Generation was acquired by News Corporation in 2010; according to press reports, News Corp acquired 90 percent of the company for approximately $360 million. Both founders previously held technical roles at IBM's research division, among other employers.

Regarding the origin of the company name, Pomel explained in a December 2025 podcast appearance that neither he nor his co-founder had ever owned a dog, but that at their previous employer servers were named after animals: production servers after dogs, staging servers after birds, and development servers after cats. The production databases were collectively called "datadog", and one of them, "datadog 17", was an Oracle database that doubled in size every year and that everyone dreaded operating. Pomel described it as a name that represented the old world they were trying to leave behind; their project was initially referred to as "datadog 17", and because the name proved memorable they dropped the number and adopted it as the company name, commissioning a designer to create the puppy logo.

According to the company's IPO prospectus, Datadog was founded with the goal of breaking down the divide between development and operations teams; it began building a unified data platform in 2010 and launched infrastructure monitoring, its first use case, in 2012. Pomel has served as chief executive officer and Lê-Quôc as chief technology officer since the company's founding, and both have been directors since June 2010.

Growth

Datadog at Google Cloud Summit

Between 2012 and 2016 the company raised four rounds of venture capital funding from investors including Index Ventures, OpenView Venture Partners and ICONIQ Capital (see below). Its platform's coverage expanded to multiple cloud providers, including Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform, Red Hat OpenShift and OpenStack. As of January 2015, the platform integrated with almost 100 commonly used technologies.

On 7 October 2015, the company opened a research and development office in Paris, in the Silicon Sentier district. According to the company, its Manhattan headquarters at the time had more than 100 employees, and it served more than 1,500 corporate customers. In 2016, Deloitte's Technology Fast 500 ranked Datadog ninth in North America, based on five-year revenue growth of 10,303 percent. The same year, the company moved its New York headquarters to The New York Times Building at 620 Eighth Avenue.

On the product side, Datadog announced a beta of its application performance monitoring (APM) product in September 2016 and made it generally available on 15 February 2017. In September 2017 it acquired the log-management company Logmatic.io, bringing infrastructure metrics, APM and log management onto a single platform. In February 2018 it released version 6.0.0 of its monitoring agent, rewritten in Go.

In February 2018 the company announced its annual DASH conference, first held on 11–12 July 2018 at Spring Studios in New York City. DASH has been held annually since; according to the company, the June 2026 edition was the ninth.

As a public company

Datadog listed its Class A common stock on the Nasdaq on 19 September 2019 (see below). In November 2019 it established a Japanese subsidiary, Datadog Japan GK.

During the 2020s the company expanded into security and artificial intelligence (AI) products, introducing Cloud SIEM in 2020, cloud security products in 2021, and LLM Observability in 2024, while continuing to make acquisitions (see below). Headcount reached approximately 8,100 and the customer base about 32,700 companies by the end of December 2025.

Datadog was added to the S&P 500 index in July 2025. In its first-quarter results for fiscal 2026, announced in May 2026, quarterly revenue exceeded $1 billion for the first time. In its second-quarter results, announced on 6 August 2026, revenue rose 36 percent year-over-year to $1.12 billion and the company raised its full-year revenue guidance from $4.30–4.34 billion to $4.45–4.47 billion.

In April 2026, the company changed its state of incorporation from Delaware to Nevada. According to its filings, the redomiciliation involved no change to its business, employees, management or headquarters.

Acquisitions

Since 2015 the company has made a series of acquisitions in monitoring, security and developer tools. According to its Form 10-K for fiscal 2025, Datadog entered into three acquisition agreements in 2023, one in 2024 and three in 2025; the aggregate consideration for the three 2025 acquisitions was $178.4 million. The report does not name the acquired companies.

Acquisitions by Datadog

Date	Company	Business	Ref.
February 2015	Mortar Data	Platform for building and operating data pipelines and applications that handle large-scale data; based in New York
September 2017	Logmatic.io	Data platform for logs and machine events; founded 2014, based in Paris
February 2019	Madumbo	Platform for testing web applications without writing additional code; based in Paris
August 2020	Undefined Labs	Testing and monitoring for developers, extending monitoring to continuous integration and continuous delivery (CI/CD) workflows
February 2021 (agreement) April 2021 (completed)	Sqreen	SaaS security platform that detects and blocks application-level attacks
February 2021	Timber Technologies	Developer of Vector, a data pipeline that collects, transforms and routes monitoring data to arbitrary destinations
November 2021	Ozcode	Live debugging with code-level visibility into applications running in production
February 2022	CoScreen	Collaboration tool for technical teams allowing multiple users to share and operate applications simultaneously
May 2022 (agreement)	Hdiv Security	Security testing software that monitors the behavior of running applications to detect known and unknown vulnerabilities
August 2022	Seekret	Automated discovery and management of APIs across environments
November 2022	Cloudcraft	Visualization service that analyzes relationships between services on Amazon Web Services and automatically generates and updates architecture diagrams
April 2023	Codiga	Static code analysis across the stages of development
November 2023	Actiondesk	Cloud-based spreadsheet application that connects to live data sources
January 2025	Quickwit	Open-source distributed search engine for logs that can run in customers' own cloud environments or on-premises
April 2025	Metaplane	Data-quality monitoring platform using machine-learning-based monitoring and column-level data lineage; founded 2020, based in Boston
May 2025	Eppo	Warehouse-native experimentation and feature-flag management platform; based in San Francisco
January 2026	Propolis	Autonomous quality-assurance testing platform that tests based on user goals and outcomes
June 2026	Adaptive ML	Developer of a reinforcement learning operations (RLOps) platform; the team was announced as joining Datadog AI Research

Metaplane and Eppo continued to be offered to existing customers under the names "Metaplane by Datadog" and "Eppo by Datadog" after the acquisitions. The purchase prices of both were not disclosed, although TechCrunch noted reports valuing the Eppo deal at $220 million.

Products

According to the company's Form 10-K for fiscal 2025, its products are modular and span infrastructure monitoring, application performance monitoring (APM), log management, digital experience monitoring, network monitoring, cloud and application security, developer-facing observability, service management, and product analytics. Capabilities such as dashboards, alerting and workflow automation are shared across products. The report states that because many products function on a standalone basis, customers can adopt capabilities incrementally or deploy several together. The products gather data through integrations that connect to the software and cloud services being monitored; the fiscal 2025 report states the platform offers more than 1,000 integrations.

The 10-K lists the introduction years of major products as follows: APM (2017); log management (2018); digital experience monitoring and network performance monitoring (2019); Cloud SIEM, Continuous Profiler and Incident Management (2020); Cloud Security Posture Management, Cloud Workload Security, Database Monitoring and Sensitive Data Scanner (2021); Application Security Management, Cloud Security Management, Audit Trail, Observability Pipelines, Cloud Cost Management and Universal Service Monitoring (2022); Application Vulnerability Management, Data Streams Monitoring and Workflow Automation (2023); Event Management and LLM Observability (2024); and OnCall, Product Analytics and Bits AI SRE (2025).

Monitoring and observability

Per the annual report, the infrastructure monitoring product monitors IT infrastructure in real time across public cloud, private cloud and hybrid environments, as well as container and serverless configurations, consolidating collected data in one place and correlating it. APM provides code-level distributed tracing across microservices, hosts, containers and serverless functions. Log management ingests and indexes logs and provides search, visualization and alerting; associated mechanisms include Logging Without Limits, which decouples the cost of log ingestion from processing, and Flex Logs, which separates storage from querying so retention and query capacity can be scaled independently.

Products addressing the user-facing side include Synthetics, which monitors applications and API endpoints using simulated requests; Real User Monitoring (RUM), which analyzes the performance of web browsers and mobile applications as experienced by actual users; and Product Analytics, which aggregates data about products and users. Other products include the always-on Continuous Profiler, Database Monitoring for query metrics and execution plans, and Error Tracking, which groups similar errors. Data Streams Monitoring targets event-driven applications and Data Jobs Monitoring targets Apache Spark and Databricks jobs; Cloud Network Monitoring covers network traffic and Network Device Monitoring covers network hardware.

In October 2025 the company released Updog, a free public web dashboard that displays the operational status of the APIs of SaaS services including Amazon Web Services, Cloudflare, OpenAI and Slack. The company said that detecting fine-grained patterns in telemetry allows outages to be detected earlier, citing a case in which it surfaced an Amazon DynamoDB degradation 32 minutes before AWS's own status page was updated.

Security

According to the annual report, the security products include Cloud Security, which scans customer infrastructure without an agent to detect vulnerabilities, misconfigurations, permission risks and compliance violations; Code Security, which prioritizes vulnerabilities using runtime context; and Cloud SIEM, which covers threat detection and investigation and the storage and analysis of security logs. Threat management comprises Workload Protection, which analyzes host and container workloads in the kernel, and App & API Protection. Sensitive Data Scanner detects, classifies and redacts sensitive data, which the report says supports compliance with regulations such as the GDPR, HIPAA and the California Consumer Privacy Act (CCPA). Audit Trail, which records changes made on the platform, was introduced in 2022.

AI products

LLM Observability, made generally available in 2024, traces chains of operations that use large language models (LLMs), recording the inputs and outputs, errors, token usage and latency of each step. In 2025 the company introduced Bits AI SRE, which the annual report positions as an always-on AI agent for troubleshooting, developed against thousands of real incidents to help identify root causes and shorten service recovery.

At its DASH conference in June 2026, the company announced what it described as more than 100 new features and enhancements, including Bring Your Own Cloud (BYOC), which deploys the platform in the customer's own environment and processes and indexes data on the customer's cloud object storage; Bits Agent Builder, for customers to build their own AI agents; AI Guard, which protects AI applications from prompt injection and data exfiltration; and Agent Console, which centrally monitors coding-assistance agents such as Claude Code, Cursor and GitHub Copilot. German IT publication heise online noted that the company positions Bits AI SRE as a model-agnostic control layer above individual LLMs and does not disclose which foundation models it uses.

The company operates an AI research organization, referred to in its fiscal 2025 annual report as the AI Research Lab. In its second-quarter 2025 results the company presented the lab's first two outputs: Toto, a foundation model for time-series data trained solely on the company's own internal telemetry, with openly released weights; and BOOM, a time-series benchmark dataset specialized for observability metrics. According to a March 2026 report in The Register, Toto has 151 million parameters and was pre-trained on more than two trillion time-series data points; the report described the company's strategy of building domain-specific models rather than relying on general-purpose LLMs.

Business model

According to the annual report, revenue is generated from subscription agreements, with transaction prices consisting of fixed fees for committed usage and variable consideration for usage purchased in addition. Usage above committed amounts is billed incrementally — monthly under ratable agreements, or once committed amounts are exhausted under usage-based agreements. Billable products include infrastructure monitoring, APM, log management, Synthetics, security monitoring, Continuous Profiler, serverless monitoring, network monitoring, RUM and incident management, together with ancillary services such as custom metrics in dashboards, monitoring of Docker containers and indexed spans.

Market position

The fiscal 2025 annual report describes the company's principal market as IT operations management and, citing Gartner research published in December 2025, sizes that market at $82 billion in 2029, of which the observability-related portion is $39 billion; the company estimates its four addressable segments (IT operations management, security software, application development and analytics platforms) at a combined $187 billion in 2029. The report notes that the Gartner material represents research opinion rather than statements of fact, as of its December 2025 publication. The company has stated that it was named a Leader in Gartner's Magic Quadrant for Observability Platforms for the sixth consecutive time in the 2026 edition.

The sector has seen consolidation among competitors: New Relic agreed to go private in a $6.5 billion deal in July 2023, and Cisco's acquisition of Splunk was completed on 18 March 2024.

Technology

High-level architecture

Monitoring agent

Hosts being monitored run software called the Datadog Agent, which collects data and sends it to the company's service. The agent is written in Go; major version 6.0.0, released in February 2018, was a complete rewrite of the earlier Python implementation. According to the release notes, the rewrite aimed at improved performance, reduced resource consumption and concurrent execution of checks, while checks and plugins written in Python remained usable without modification. Major version 7 started with 7.16.0 and supports only Python 3.

The agent's source code is public, dual-licensed with user-space components under the Apache License 2.0 and BPF code under the GNU General Public License 2.0. Versions 5 and earlier were maintained in a separate repository (DataDog/dd-agent), written in Python and released under a simplified BSD license; that repository's license file carries copyright notices for Boxed Ice (2009) and Datadog (2010–2016).

Agentless monitoring is also available; for AWS, the service periodically calls the Amazon CloudWatch API under a delegated IAM role.

Backend

The backend data stores have been rebuilt over several generations. According to the company's engineering blog, time-series data was originally stored in Apache Cassandra, followed by implementations based on Redis, MDBM, a Go implementation of B+ trees, and RocksDB. Because the Go implementation handled scalar metrics while the RocksDB-based implementation handled distribution metrics, two time-series databases ran in parallel until they were consolidated into a single engine, "Monocle", written in Rust and structured as a log-structured merge-tree.

Logs and other events are stored in "Husky", a third-generation event store introduced in 2022. According to the company, Husky is a schemaless, distributed, column-oriented store that keeps data on commodity object storage and supports both analytical aggregation and full-text search. It separates roles among Writers, Compactors (which merge small files) and Readers, with FoundationDB serving as the metadata store that tracks which files are visible to each tenant. Apache Kafka is used in the ingestion path.

Sites

The service is offered across multiple independent "sites" that do not share data with one another: US1, US3 and US5 in the United States, EU1 in Europe, UK1 in the United Kingdom, AP1 in Japan, AP2 in Australia, and US1-FED and US2-FED for U.S. government customers.

The Japanese data center began operating in Tokyo in April 2023; according to the company it was its first data center in Asia, adding to existing locations in the United States, Europe and AWS GovCloud, and addressing requirements to store data within Japan. The Australian site launched in July 2025 on AWS's Asia Pacific (Sydney) region.

For the U.S. government, the US1-FED site is offered as "Datadog for Government"; the company announced FedRAMP High baseline authorization in May 2026. Its Trust Center also lists conformance with SOC 2 Type 2, ISO/IEC 27001, ISO/IEC 27017:2015, ISO/IEC 27018:2019, ISO/IEC 27701, ISO/IEC 42001:2023 for AI management systems, PCI DSS and Australia's IRAP, among others.

Open-source involvement

Through acquisitions, the company is involved in the development of several open-source projects. Timber Technologies, acquired in February 2021, developed Vector, a data pipeline that collects, transforms and routes monitoring data; Vector is written in Rust and released under the Mozilla Public License 2.0. Quickwit, acquired in January 2025, developed an open-source distributed search engine for logs. The company also publishes instrumentation libraries for various languages, such as dd-trace-py for Python.

Finances

Revenue grew from $362.78 million in fiscal 2019 to $3.43 billion in fiscal 2025.

Revenue (US$ millions)

Fiscal year	Revenue
2019	362.8
2020	603.5
2021	1,028.8
2022	1,675.1
2023	2,128.4
2024	2,684.3
2025	3,427.2

(Source: Form 10-K filings for the respective years)

By region, fiscal 2025 revenue was $2.43 billion in North America and $994.08 million elsewhere; based on customer billing addresses, regions outside North America accounted for about 29 percent of revenue, and the United States was the only single country exceeding 10 percent.

As of 31 December 2025, the company had about 32,700 customers in more than 160 countries. Approximately 4,310 customers had annual run-rate revenue (ARR) of $100,000 or more, together representing about 90 percent of total ARR, and about 603 customers had ARR of $1 million or more.

As of the same date the company had approximately 8,100 employees across 35 countries; about 44 percent of full-time employees were located outside the United States, 34 percent of whom were in France. Research and development employed about 3,900 people and sales and marketing about 3,600. The company's headquarters occupy leased space at 620 Eighth Avenue (The New York Times Building) in New York City, with additional leased offices in Boston, Denver, San Francisco, Paris, Dublin, Amsterdam, Sydney, Tokyo, Singapore and Seoul; it owns no real estate.

Funding

Before going public, Datadog raised several rounds of venture funding: a $6.2 million Series A co-led by Index Ventures and RTP Ventures in November 2012, a $15 million Series B led by OpenView Venture Partners in February 2014, a $31 million Series C led by Index Ventures in January 2015, and a $94.5 million Series D led by ICONIQ Capital in January 2016, bringing total funding at that point to approximately $148 million.

Initial public offering

The company filed a registration statement (Form S-1) with the U.S. Securities and Exchange Commission on 23 August 2019 and listed its Class A common stock on the Nasdaq Global Select Market on 19 September 2019, at an IPO price of $27.00 per share for an initial offering of 24 million shares. The stock opened at $40.50 and closed its first day at $37.55, 39 percent above the IPO price. CNBC reported that Datadog was the fourth cloud software company to go public in 2019 and reach a market capitalization of at least $10 billion, after Zoom, Slack and CrowdStrike. The offering closed on 23 September; after the underwriters exercised their option to purchase additional shares on 25 September, the total came to 27.6 million shares and net proceeds after fees and expenses were $705.9 million. The joint lead underwriters were Morgan Stanley, Goldman Sachs, J.P. Morgan and Credit Suisse.

In September 2019, ahead of the listing, Bloomberg reported, citing people familiar with the matter, that Cisco had offered more than $7 billion to acquire Datadog a few weeks before the IPO, and that Datadog declined in favor of going public.

Share classes and index membership

The common stock is divided into Class A shares carrying one vote each and Class B shares carrying ten votes each; only Class A is listed. As of 31 December 2025, outstanding Class B shares represented about 43 percent of total voting power, with holders including the company's officers and directors. As of 5 February 2026, 328,274,648 Class A and 24,301,433 Class B shares were outstanding.

Datadog was added to the Nasdaq-100 index in the annual reconstitution announced on 10 December 2021, effective before the market open on 20 December. On 2 July 2025 it was announced that Datadog would join the S&P 500, replacing Juniper Networks following its acquisition by Hewlett Packard Enterprise, effective 9 July.

External links

• Official website
• Business data for Datadog, Inc.:
  • Google
  • Reuters
  • SEC filings
  • Yahoo!