Skip to content

Depthfirst

Extremely fast-moving, well-capitalized AI security lab (two large rounds within 90 days) founded by DeepMind/Databricks/Faire alumni; still early enough that independent efficacy data is limited.

Visit Website ↗ + Add to Compare
50/100Incremental Innovator

Innovation Matrix Assessment

Innovation Velocity 7/10

Raised an $80M Series B (announced 2026-03-31) led by Meritech Capital less than 90 days after a $40M Series A, bringing total funding to $120M since a 2024 founding.

Operational Value 4/10

Enterprise adoption is described as an active scaling priority for the new funding; specific named customer deployments were not disclosed in coverage reviewed.

Market Momentum 6/10

Series B closed 2026-03-31 with Meritech Capital, Forerunner Ventures, and The House Fund joining existing investors Accel and Box Group.

Category Disruption 5/10

Building proprietary, in-house security-specific AI models (dfs-mini1) rather than relying on general-purpose LLMs is a differentiated technical bet.

Real-World Efficacy 3/10

False-positive-reduction and remediation-quality claims are company-described; no independent benchmark was located.

Enduring Relevance 5/10

Reducing false positives and providing actionable fixes addresses a long-standing, real pain point in application and infrastructure security.

Why CISOs Should Care

Depthfirst is an applied AI lab building security models that understand code, business logic, and infrastructure together to identify real vulnerabilities, cut false positives, and hand developers actionable fixes rather than another alert queue.

What Makes It Different

Trains its own in-house security models (starting with dfs-mini1) purpose-built for vulnerability detection and remediation, rather than wrapping general-purpose LLMs around existing SAST/DAST tooling.

The Matrix Verdict

50/100 — INCREMENTAL INNOVATOR

Extremely fast-moving, well-capitalized AI security lab (two large rounds within 90 days) founded by DeepMind/Databricks/Faire alumni; still early enough that independent efficacy data is limited.

Editorial Note: Claims vs. Verified Findings

Funding history and founding-team background verified via SecurityWeek and BusinessWire. False-positive-reduction and fix-quality claims are vendor-described.

Sources