20VC: NVIDIA Bonanza: Buys Poolside & Invests in Mercor and Perplexity | Anthropic's $30TRN Revenue Assumption & OpenAI Confirms IPO | Why Customer Service, Defence and Robotics are Overinflated
Summary
Jason Lemkin, Rory O'Driscoll, and Harry Stebbings analyze NVIDIA’s $12B Poolside acquisition and investments in Mercur and Perplexity, highlighting the capital intensity of frontier AI. They discuss OpenAI’s planned 2027 IPO, noting its pressure to compete with Anthropic’s growth and the strategic shift toward open-weight models. The trio emphasizes that while seed returns face higher bars, executing in a hyper-growth market can still yield significant exits even if the original thesis doesn't fully materialize.
Topics discussed
Intro: Seed investing realities and sponsor ads
Analysis of US open source model funding failures and 15x returns
Cohere's $20B valuation and NVIDIA's vendor financing strategy
Frontier model training budgets and gross margin dynamics
OpenAI's IPO timeline, competition with Anthropic, and market position
OpenAI's mission, Sora's ROI, and consumer business challenges
Hugging Face acquisition rumors and the open-weight model ecosystem
Public market volatility, leverage risks, and semiconductor margins
Silicon Valley housing costs and wealth concentration trends
The 'token addiction' cycle and societal reliance on AI agents
Corporate budget backlash and the pace of AI diffusion in enterprise
Stripe's private market strategies and stock buybacks
AI agent security risks and the limits of personal productivity tools
Venture returns in defense, robotics, and the future of CX
Listen ad-free on Castria