Series A & B · Lead investorVertical AI · Infrastructure · Security · Dual-useSan Francisco, CA
Portfolio

Teams that replace the work.

Every company here sits in the critical path of its customers’ operations. Each one gets better with use and is built on advantages a general-purpose model can’t easily copy.

Series A
2025

Ledgerline

Autonomous month-end close for mid-market finance teams. Reconciles, accrues and flags exceptions before the controller logs in.

Vertical AIFinance
Series B
2025

Sentinel Grid

Identity and permissions for AI agents. Every agent action is scoped, logged and revocable across the enterprise stack.

AI SecurityIdentity
Series A
2026

Halyard

Agent orchestration and evaluation for regulated enterprises. Production reliability with a full audit trail.

InfrastructureEvals
Series B
2024

Clearwater Claims

An AI adjuster for commercial property insurers. Takes first notice of loss through settlement recommendation in hours instead of weeks.

Vertical AIInsurance
Series A
2025

Meridian Works

Autonomous planning and quality inspection for advanced manufacturing lines serving aerospace and defense primes.

Dual-useManufacturing
Series A
2026

Pallas Health

Revenue-cycle agents for health systems that handle prior authorizations, denials and appeals end to end.

Vertical AIHealthcare
Series A
2024

Vantage Labs

An autonomous SOC analyst that triages, investigates and closes alerts, escalating only what needs human judgment.

AI SecuritySecOps
Series B
2026

Foundry Data

Turns unstructured enterprise documents into governed, permission-aware context for agents in production.

InfrastructureData
Series A
2026

Atlas Orbital

Geospatial intelligence that combines satellite, sensor and logistics data into decisions for commercial and public-sector operators.

Dual-useGeospatial
Series A
2025

Brief & Bench

A litigation operations agent that runs discovery review, privilege logs and deadline management for AmLaw 200 firms.

Vertical AILegal
Series A
2025

Keel Inference

Routing and cost control for enterprise inference. Picks the right model for each task and cuts spend without losing quality.

InfrastructureInference
Series B
2026

Obsidian Shield

Runtime protection for models and agents. Detects prompt injection, data exfiltration and unsafe tool calls in real time.

AI SecurityRuntime
12
Portfolio companies
100%
Rounds led or co-led
4
Focus areas
A–B
Entry stage
Join them

Building the next one? Tell us.