Partner at Craft Ventures
Check size: No personal check-size authority is published. Disclosed Craft financings involving Kevin range from seed and Series A through Series C and later growth rounds; the financing totals are company-round sizes, not Kevin's individual checks.
Kevin backs early- and late-stage software companies where a major platform shift creates a new security or infrastructure control point. His recent work emphasizes runtime-first cloud security, autonomous offensive and defensive systems, non-human identity, agent access, endpoint visibility, and control of agentic software. He looks for category-defining technical architectures, rapid product velocity, deep customer pull, and teams capable of expanding a focused wedge into a large enterprise platform.
Define the new attack surface or infrastructure bottleneck, why incumbents cannot extend into it, and the architectural control point you own. Bring technical benchmarks, deployment time, telemetry advantage, false-positive and prevention metrics, customer references, retention and expansion, product-release cadence, enterprise readiness, and a credible path from the initial wedge to a multi-product platform. For growth rounds, add durable revenue quality and category-leadership evidence.
Experienced technical founders with a precise view of an emerging control point, deep product architecture, unusually fast execution, obsessive customer engagement, strong enterprise references, and a route from a sharp initial product to a durable platform. His disclosed investments repeatedly favor teams that anticipated a category before consensus and then proved the thesis with usage and expansion.
No explicit personal pass list is public. His investment writing implies weaker fit for crowded AI-security products without a distinct control layer, reactive tools that cannot operate at machine speed, shallow wrappers without proprietary telemetry or architecture, point products with no platform path, and teams lacking enterprise-grade reliability or direct customer evidence.
Identify the new technical layer created by a platform shift, verify that incumbents cannot cover it through feature extension, and test whether the startup owns the telemetry and enforcement point needed to become a platform.
Track whether an initially non-consensus architecture compounds into customer adoption, rapid product expansion, strong references, and repeated financing rather than relying on the originality of the thesis alone.
“A new layer of agentic software deserves a new layer of control.”
— https://www.craftventures.com/articles/partnering-with-neo-to-secure-agentic-software
Agentic software creates a new enterprise control layer, favoring endpoint-native platforms that understand intent, create policy, and enforce guardrails in real time.
On-device AI can build a contextual model of human and agent work, enabling real-time prevention of insider risk, data loss, and unsafe actions without burdensome endpoint overhead.
The explosion of non-human identities and agent access requires a purpose-built privileged-access and governance platform backed by fast execution and strong enterprise validation.
A runtime-first cloud-security architecture, deep customer obsession, and relentless product velocity can displace established competitors and expand from an initially contrarian category thesis.
+ 106 more investments. View fund →