← Home
The Twenty Minute VC · June 25, 2026 · 1h 24m

Deepseek Raises $50BN | Wall St's $725BN AI Question | The Rise of Open Source & How it Threatens OpenAI & Anthropic | OpenAI Builds it's Own Chip: Jalapeno | The Death of Moats & The New AI Software Winners

In this solo episode of 20VC, Harry Stebbings breaks down the biggest shifts in AI: Anthropic poaching Google talent, DeepSeek's $50B raise, Apple's memory supply warning, the $725B return question from Wall Street, experiments with AI finance agents, the erosion of defensive moats, and why open source now threatens OpenAI more than Anthropic ever did.

This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.

Novel

The 20VC team created an AI finance vice president that outperforms human counterparts on speed, accuracy, and cost.

Highlights

Google Loses Two AI Legends as Anthropic Wins the Talent War00:00
Anthropic poached two top AI researchers from Google, signaling a shift in where elite AI talent wants to work.
China's $50B DeepSeek Bet Changes the AI Power Balance14:45
Chinese AI lab DeepSeek raised $50 billion, a sum that could reshape global AI competitiveness.
AI's Memory Crisis Has Begun — Apple Warns of a '100-Year Flood'27:15
Apple warned that demand for high-bandwidth memory (HBM) is creating a supply crunch as severe as a 100-year flood, threatening AI training capacity.
Databricks, ServiceNow & the New AI Software Winners58:30
Established platforms like Databricks and ServiceNow are poised to capture outsized AI value by embedding models into existing enterprise workflows.
Per-seat pricing becomes obsolete when AI agents can do the work of many humans, collapsing the unit economics of traditional SaaS.
OpenAI's Custom Models Could Rewrite Enterprise Software01:12:00
OpenAI's development of bespoke, fine-tuned models for enterprises could replace large chunks of traditional SaaS.
OpenAI's Biggest Threat Isn't Anthropic Anymore01:17:00
Open-source models—not Anthropic—now pose the greatest competitive threat to OpenAI’s closed-source dominance.

Editorial

Wall Street Finally Asks the $725 Billion Question: Who Pays for AI?30:00
Analysts are now openly questioning how the massive $725 billion in AI infrastructure spending will ever generate a commensurate return.
Traditional business moats are eroding under rapid AI iteration, making defensibility a misleading concept for startups.
Was this useful?