20VC: Why You Need a $1BN Fund To Do Series A Today | OpenAI vs Anthropic: Who Wins Enterprise | SpaceX at $2TRN and Data Centers in Space | The $20BN Groq Deal Broken Down | Jeff Bezos' $100BN New Fund

What It Covers

Harry Stebbings, Jason Lemkin, and Rory O'Driscoll analyze Anthropic capturing 73% of new enterprise AI spending versus OpenAI, SpaceX's potential $2T valuation after announcing a chip fabrication facility, the $20B Groq-NVIDIA asset deal structure, Jeff Bezos raising $100B for AI-transformed manufacturing, and why Series A funds now require $1B minimum to compete effectively.

Key Questions Answered

• •Anthropic vs. OpenAI Enterprise Lock-in: Ramp data shows Anthropic capturing 73% of new enterprise AI spending, up from 40% in December. The strategic risk for OpenAI: enterprises building production applications on Claude Sonnet and Opus 4.5/4.6 are unlikely to switch after weeks of tuning and QA investment. Token costs below 5-20% of revenue make switching economically irrational, creating durable lock-in that compounds monthly.
• •AI Product Monetization Test: The clearest signal of whether an AI feature has real value is a 50% or greater ARPU increase post-launch. Notion doubled ARPU by charging $20/month for AI versus $10 for base. If a software company cannot charge meaningfully more for its AI capabilities, the feature lacks product-market fit. Figma's Make tool failing this test while the company adds sales headcount signals deteriorating fundamentals.
• •Series A Fund Size Floor: Leading Series A rounds today requires writing $25-30M checks, maintaining 50% reserves on initial capital, and running 20-30 portfolio companies. The math produces a minimum viable fund size of roughly $1B. Funds below this threshold cannot lead competitive rounds, forcing them into follower positions with less ownership and weaker governance rights at a time when round sizes have doubled over 18 months.
• •Groq-NVIDIA Deal Structure Warning: The $20B Groq acquisition used an asset purchase structure to avoid antitrust review, triggering double taxation: corporate-level tax on the asset sale gain, then individual investor tax on distributions. The effective tax rate for founder Jonathan Ross reached approximately 60%, destroying an estimated $4-5B in value. Asset purchase structures are the only viable path for large AI acquisitions avoiding regulatory scrutiny, but the cost is severe.
• •Unicorn Exit Math Crisis: The ratio of viable acquirers to unicorn-plus companies sits at a career low. Hyperscalers will not acquire hundreds of vertical AI application companies. Legacy software incumbents cannot afford to buy AI-native replacements valued above their own market caps. This leaves IPO as the only realistic exit, yet many companies raised at $9-10B valuations that current public market fundamentals cannot support, creating a structural liquidity trap.
• •SpaceX Valuation Framework: Evaluating Elon Musk's ventures requires separating three categories: revenue-generating businesses valued on multiples, announced projects in execution, and speculative future visions. Starlink's 53%+ profit margins provide a real DCF anchor. The TerraFab announcement adds option value only if assigned a probability of completion and timeline. Musk's track record on engineering achievement is strong; his track record on timing predictions is consistently optimistic by years.