AI Agent Capital

The firm targets software that autonomously executes multi-step tasks by combining large language models with tool use, memory, and planning. This excludes copilot-style interfaces designed merely to suggest text or code. Its investment threshold requires that the product can act independently on behalf of a user or system, handling edge-case failures without real-time human gating, which the firm views as the demarcation between the generative AI wave of 2023–2024 and the agentic systems maturing in 2025 and beyond.

AI Agent Capital relies on a technical sourcing model centered on relationships with researchers and engineers at frontier AI labs, particularly OpenAI, DeepMind, and Anthropic. The firm's partners are typically present at agent-specific academic workshops and developer conferences, and they maintain active participation in open-source agent framework communities. A substantial portion of its pipeline is founder-referred within the AI research diaspora, giving it early access to spinouts and new company formations.

The firm invests across both infrastructure and vertical applications, provided they are specific to the agentic stack. Infrastructure investments include evaluation and observability platforms for agent performance, memory-retrieval systems, and multi-agent orchestration layers. On the application side, it seeks vertical SaaS products where an autonomous agent replaces an entire human-staffed service category rather than incrementally augmenting an existing enterprise workflow.

The firm concentrates its investments in three primary hubs: San Francisco, London, and Tel Aviv. It does not maintain formal offices in all three locations but partners near the major AI research centers in each region. The geographic selectivity is driven by the conviction that foundational AI talent remains clustered in these cities and that early-stage agentic companies require dense access to compute providers and academic collaborators.

The firm's single-strategy mandate is its primary structural differentiator. While many early-stage AI funds have broadened to cover everything from foundation models to AI-enabled biotech, AI Agent Capital restricts its portfolio to companies where autonomous agents are the core product thesis. This creates a narrow, high-conviction portfolio that avoids dilution into adjacent AI categories and allows the firm's partners to provide specialized technical due diligence and community access specifically within the agentic ecosystem.