Anthropic, Glean & OpenRouter: How AI Moats Are Built with Deedy Das of Menlo Ventures

What It Covers

Menlo Ventures partner Deedy Das discusses Anthropic's meteoric rise to $60B valuation, enterprise AI moats, and investment strategies across model versus application layers.

Key Questions Answered

• •Enterprise Search Moats: Glean's success stems from solving unglamorous technical problems like handling multiple Slack workspaces and building custom integrations that competitors won't prioritize long-term.
• •AI Market Share Shift: Enterprise LLM API spend shows OpenAI dropping from 50% to 25% market share while Anthropic rose from 12% to 32% in two years.
• •Model Layer Advantage: Building foundational AI models requires harder technical work than applications, creating stronger defensive moats since apps can't easily replicate model capabilities.
• •Research Investment Framework: Follow talented teams with clear competence toward likely future outcomes, like mechanistic interpretability for AI transparency as models make critical societal decisions.