newsfilter
← Interviews
Bloomberg Television7 min

Menlo Makes First-Ever IPO Investment With Anthropic

Matt Murphy, Ed Ludlow, Romaine Bostick

Investor Day, Panel
  • Menlo Ventures secured approximately 50% of their available allocation for Anthropic's IPO, marking the firm's first-ever investment in an initial public offering despite its typical focus on early and growth-stage private equity.
  • Anthropic achieved capital efficiency by building foundational models for roughly 1/100th the cost of OpenAI at the time, leveraging "compute multipliers" that increased compute efficiency by 4x year-over-year.
  • The company demonstrated exceptional market traction immediately post-launch, soaring in revenue with model benchmarks that were at or better than OpenAI on nearly every dimension despite raising significantly less capital.
  • Forward-looking market expansion targets include:
    • Enterprise Replatforming: Potential to replace legacy mainframe systems by refactoring old applications to run on modern infrastructure.
    • Vertical Penetration: Achieving near 100% developer adoption for tools like "Claude Code" overnight, expanding into life sciences for accelerated drug discovery.
    • Diverse Applications: Extending beyond text generation to robotics (via foundation models like those at Skill), video production, and consumer applications without the vertical impediments of past technology waves.
  • Strategic differentiation exists in the application layer, where specialized partners like Legor (focused on legal) utilize Anthropic's models but add forward-deployed engineering and organizational workflow adaptation that the base model alone does not provide.
  • Anthropic maintains a foundational commitment to "trust and safety" as a day-one principle, with the current strategy prioritizing safety and responsible AI deployment over short-term shareholder returns, exemplified by the Pentagon example and public statements from leadership.
  • Management anticipates the company will evolve from selling tokens to selling "intelligence" and innovation capabilities, with the belief that the current enterprise usage represents only the early innings of a broader AI-driven replatforming wave.