Portfolio

AI character platform for interactive roleplay and creative storytelling.

Janitor AI is a textbook power-user product: a highly engaged community of users who push the platform to its limits and pull the mainstream behind them. We backed the founders because they understood that AI-native character interaction is a new content category, not a feature.

AI matchmaking for professionals — connecting people based on what they actually need right now.

Professional networking is broken because it is symmetric and passive. Sitch's agentic approach to matching makes it asymmetric and active. We invested because the founders have deep distribution insight into how professionals actually form high-value relationships.

Authorization layer for AI agents — enabling agents to act on behalf of users safely and with consent.

As AI agents proliferate, the permission and authorization layer becomes critical infrastructure. Arcade is building the trust fabric that lets agents operate in the real world. We invested because this is foundational agentic software with a clear path to becoming a standard.

AI-powered professional network that helps people present their best professional self.

Professional identity is ripe for AI-native reinvention. SuperMe's founders have a sharp view of how AI changes the way people communicate their value and build their careers. We invested because they are building for power users of professional networks first.

AI-powered voice dictation built for professionals who think faster than they type.

Voice input has failed repeatedly because it was built for average users. Willow is built for power users — people who need speed, accuracy, and deep integration with their workflows. We invested because the founders have a technical edge in speech that most teams cannot replicate.

AI lab building world models for interactive and generative experiences.

World models are the next frontier of AI-native content. Moonlake's founders are among the few teams with the technical depth to build at this layer. We invested because they are building infrastructure that will power an entirely new category of agentic and interactive products.

Agentic map — a spatial interface for AI agents to understand and navigate the physical world.

Maps have not been reimagined for the agentic era. Gobi is building the spatial layer that lets AI agents reason about place, movement, and context. We invested because the founders have a unique combination of mapping expertise and agentic software vision.

AI intelligence layer for private credit operations — automating the data-intensive work of credit analysis.

Private credit is a massive, underserved market where the work is highly repetitive and data-intensive. Ellis is building an AI-native tool for the power users of this market: credit analysts and portfolio managers who need speed and accuracy at scale. We invested because the founders have deep domain obsession and a clear wedge.