Overview
Sqreen put attack detection and blocking inside production web applications. Founded in Paris in 2015 by former Apple security colleagues Pierre Betouin and Jean-Baptiste Aviat, it joined YC's Winter 2018 batch and was acquired by Datadog on April 12, 2021. Its useful idea survives in Datadog's application security products. [1] [6]
The acquisition connected application-level security with observability: performance traces and service context used to investigate production behavior. Datadog presented the combination as a way to extend its application-performance monitoring into security. [6]
Datadog initially disclosed approximately $260 million in cash and stock, subject to adjustments, with about 25% deferred. Its 2021 annual filing later recorded approximately $219.4 million consideration for the April security-platform acquisition. These are different transaction/accounting descriptions; neither establishes investor returns. [7] [8]
Image 1 / 1
Founding Story
Betouin had spent nine years at Apple, leading security assessments for its Internet Services department. He and Aviat brought offensive-security experience to a product aimed at the people building and operating applications. [1]
The problem was practical. A suspicious network request alone provides limited evidence about what the application actually executed. Sqreen's microagents added information from within the application, connecting requests with code activity and security controls. [17]
Alven recalls meeting a small team in January 2016, with a beta used by fewer than a dozen non-paying customers. Its retrospective also describes Lucas, a former Alven employee, becoming Sqreen's founding sales representative. The company needed commercial learning alongside its technical work. [3]
Betouin said the first two years emphasized product. Before YC, the company had roughly $3 million raised, 15 employees and 100 customers. He treated YC as a way to accelerate US entry, and described establishing a San Francisco office in 2019. The account describes product readiness and deliberate US entry. [2]
Timeline
- 2015: Founded in Paris; YC records the founding year as 2015. [1]
- January 2016: Alven meets the early beta team. [3]
- September 2016: Sqreen reaches the TechCrunch Disrupt Startup Battlefield final. [19]
- Winter 2018: Sqreen joins YC. [1]
- April 2, 2019: Sqreen announces a $14 million Greylock-led Series A, more than $18 million raised in total and over 500 companies protected. Sarah Guo joins the board. [4]
- February 11, 2021: Datadog announces a definitive agreement to acquire Sqreen. [5]
- April 12, 2021: Datadog confirms closing. [6]
- April 28, 2022: Datadog launches generally available Application Security Monitoring and says Sqreen is fully integrated into its Cloud Security Platform. [9]
- October 2026 check: Datadog's documentation calls the successor product App and API Protection, formerly ASM. [10]
What They Built
Sqreen combined two controls. Runtime application self-protection observes execution inside an application to detect and act on attacks. An in-app web application firewall inspects requests within the application's runtime. Datadog's acquisition account describes following suspicious requests toward code execution and helping responders investigate the cause. [17]
The company sold Application Security Management as a unified production-security platform. Its 2019 announcement describes microagents embedded in applications, rather than a new network appliance. Le Monde's CTO praised the attack insights; Swell Investing's CTO described straightforward integration and rule setup. These are attributed customer testimonials, not independent efficacy measurements. [4]
In 2018, Security Hub introduced configurable security plugins. Reported customers included Front, Mindbody, BlaBlaCar, Triplebyte, Toptal and Algolia. [18]
The architecture creates a tradeoff. Application instrumentation can connect a request to its execution context, but installation, runtime support, policy configuration and performance still matter. A detected attempt is not automatically a successful exploit; a recorded block does not prove every attack was prevented.
What Datadog actually integrated
The 2021 acquisition announcement described the intended combination. The 2022 launch supplies the later product milestone: distributed tracing used for application-security monitoring, presented within Datadog's platform. Acquisition intent and a shipped integration are separate evidence. [9] [17]
Current AAP extends that account with API discovery, posture and protection. Datadog describes an in-app WAF, IP/user blocking and attack qualification using execution context. Qualification can remain unknown when evidence is insufficient. Blocking depends on compatible tracing libraries and configuration; support must be checked for the actual service. [10] [11]
Market Position
Sqreen's developer-oriented deployment gave engineering teams a route into production security. That did not eliminate security-team ownership: the original announcement explicitly described collaboration between developers and security teams. [4]
Betouin saw observability vendors extending into security while traditional security vendors consolidated. His account connects that market observation to Datadog's adjacent product philosophy. A larger platform offered a distribution and shared-context route, but public material does not quantify how much standalone growth this displaced. [2]
The present market still includes distinct alternatives:
| Offering | Observed approach | Buyer must check |
|---|---|---|
| Datadog AAP | Application tracing, threat investigation and configured blocking | Runtime support, required Infrastructure Monitoring and billed hosts |
| Contrast Protect | In-runtime monitoring and defensive controls | Specific language/component support and deployment modes |
| Falco | Kernel-system-call events and rule evaluation | Host privileges, event loss, rule maintenance and application-context gaps |
The tools inspect different evidence and require different deployment work. Contrast documents Protect separately by language. Falco documents kernel events and warns that raw system-call arguments cannot easily be correlated with other events. [11] [14] [15] [16]
Business Model
Sqreen was a SaaS security platform. Its 2019 funding release reported more than 500 protected companies and more than $18 million in funding. YC's later profile says more than 800 organizations trusted the product. Those statements use different dates and definitions; they do not disclose churn, revenue growth or annual contract value. [4] [1]
Standalone audited revenue, gross margin and burn were not established. Financing raised is not a company's sale value or a shareholder payout. Datadog's announced transaction description and later accounting consideration should retain their own labels. [7] [8]
Today's Datadog pricing page lists AAP starting at $31 per host per month, billed annually, or $36 on-demand. It says AAP can be sold separately but requires Infrastructure Monitoring. A comparison must include that prerequisite and the actual billable-host definition; it cannot assume buyers must purchase every observability product. [12]
Post-Mortem
The founders chose a platform combination
The strongest causal evidence is Betouin's account: discussions began during Series B fundraising, the product visions were adjacent, and the founders expected to build faster together. This explains the choice they described. It leaves independent runway, rejected financing terms and alternative offers unknown. [2]
Shared application telemetry supplies a plausible mechanism. Engineers can investigate suspicious requests alongside application behavior, and an existing instrumentation footprint can reduce a separate deployment task. Datadog's later ASM launch supports the integration direction. It does not measure sales savings, switching costs or the counterfactual outcome had Sqreen stayed independent. [9]
Integration preserved the job and changed its setting
Sqreen's standalone chapter ended with acquisition; application-security work continued inside Datadog. The current product's tracing-library prerequisites and explicit unknown attack qualifications show why the outcome is more than a checkbox. Instrumentation, evidence coverage and review still constrain useful protection. [11]
For a rebuild, the relevant opening is a testable workflow problem: helping a team establish which services are covered, what an event proves, and whether a narrowly scoped policy is safe to enforce. A new vendor must compare that job with current Datadog, Contrast and Falco workflows. A cheaper interface or a kernel probe alone does not establish better protection or demand.
Key Lessons
- Distribution can shape a strategic exit. Existing instrumentation and shared customer workflows can make a platform combination attractive.
- Follow the acquisition through integration. An agreement describes intent; a later launch shows what reached customers.
- Separate attempts, evidence and outcomes. Unknown qualification and unsupported services must remain visible.
- Compare the complete buying requirement. Include prerequisite products, supported deployments and actual pricing units.
- Keep economics tied to their measure. Customer counts, funding and accounting consideration cannot establish burn or investor returns.
Sources
- YC company profile
- Betouin interview, June 14, 2021
- Alven acquisition reflection
- Sqreen Series A announcement, April 2, 2019
- Datadog acquisition agreement
- Datadog acquisition closing
- Datadog 2020 10-K, subsequent events
- Datadog 2021 10-K, Note 7
- Datadog 2022 ASM launch
- Current App and API Protection
- Datadog protection mechanics
- Datadog current pricing
- Contrast Protect
- Contrast supported languages
- Falco kernel events
- Falco dropped-event limitations
- Datadog acquisition product account
- Security Hub launch coverage
- Startup Battlefield 2016 finalists

