The China-US tech rivalry just escalated from chips to AI models, but the real power play? It's open-source economics and who controls LNG for the hyperscalers.
📊 11 episodes across 9 podcasts
⏱ 599 minutes of intelligence analyzed
🎙 Featuring: Chamath (All-In Podcast, LLC), Jason (All-In Podcast, LLC), David Sacks (All-In Podcast, LLC), Friedberg (All-In Podcast, LLC)
|
The Big Shift
The narrative around global AI dominance is shifting from pure compute power to the economics of open-source models and the underlying energy infrastructure. Forget just NVIDIA — the real battleground is access to cheap, abundant natural gas, which China quietly commands, alongside a strategic push towards commoditizing AI knowledge.
The problem: The US is focused on protecting proprietary AI models, while China is flooding the market with powerful, low-cost open-source alternatives. This isn't just about market share; it's about altering the global technology landscape, as highlighted by multiple hosts on The Prof G Pod with Scott Galloway and All-In with Chamath, Jason, Sacks & Friedberg.
The evidence: Chinese models like Kimi K3 offer comparable performance to leading Western models at up to 99% less cost, according to Ed on The Prof G Pod. This "AI dumping" is undercutting American alternatives and could tank the market if the US government tries to intervene, warns Jason on All-In. More fundamentally, China's favorable view toward AI (84% excited vs. 10% in the US) and its efficient use of compute due to chip constraints give it a structural advantage, notes Selina Xu on The Prof G Pod. This isn't a short-term trend; it's a long-game strategy to shift global economic value from IP to physical production.
"The price per million output tokens for OpenAI's model GPT 5.6 is $45. For Claude Fable 5, it's $50. The equivalent price for Deepseek's model is $0.87. So it is 99% lower than the American alternatives."
— Ed on The Prof G Pod with Scott Galloway
Why it matters: This isn't just about software; it's about the physical world. The energy demands of AI data centers are soaring, creating a projected natural gas deficit in the US by 2028, according to Matthew Smith (Invest Like the Best with Patrick O'Shaughnessy). China’s "east data west compute" strategy leverages its energy resources to power AI, underscoring a broader shift where physical infrastructure, not just algorithms, determines AI supremacy.
The move: Expect increased scrutiny on energy infrastructure and a growing tension between protecting IP and fostering open-source innovation. The US needs to decide if it's protecting AI software or enabling its global proliferation and the underlying energy demand that comes with it.
The Rundown
① Physical AI is a much larger play than digital AI.
Applied Intuition co-founders Qasar Younis and Peter Ludwig argue that physical AI, driving intelligent machines across industries, will be orders of magnitude larger than digital AI, impacting 3% of global GDP just from automotive. (Qasar Younis on Business Breakdowns) ▶ Listen · Apple Podcasts
→ The Opportunity: Investors and operators should look beyond software and into the industrial applications of AI, where proprietary data from real-world machines creates a significant moat and the market potential is vastly underexplored.
② NYU's endowment overhaul shows aggressive capital allocation is back for LPs.
Michelle Knudsen, CIO of NYU, radically transformed the $8 billion endowment in two years, shifting from a conservative, bottom-up approach to embracing quant, macro, venture, and emerging managers. (Michelle Knudsen on Capital Allocators – Inside the Institutional Investment Industry) ▶ Listen · Apple Podcasts
→ What This Means for GPs: LPs are increasingly willing to support bold, diversified strategies and new managers if the plan is clear and governance strong, signaling a renewed appetite for growth and active management.
③ AI literacy is the new entrepreneurial filter, bypassing traditional hurdles.
Mark Cuban states AI will enable a new wave of entrepreneurs by democratizing innovation and believes LLMs are critical truth-seeking tools that can combat political information asymmetry. (Mark Cuban on All-In with Chamath, Jason, Sacks & Friedberg) ▶ Listen · Apple Podcasts
→ The Operator Play: Companies need to prioritize "AI-literate" employees and integrate AI tools to stay competitive, as this technology significantly lowers the barrier to entry for new ventures and improves data-driven decision-making.
④ Deep customer relationships, not just price, win in a scaled market.
Chuck Surack built Sweetwater, a billion-dollar music and audio equipment company, by prioritizing exceptional customer service and deep relationships over price competition, a contrarian strategy that even avoided early e-commerce. (Chuck Surack on How I Built This with Guy Raz) ▶ Listen · Apple Podcasts
→ Key Takeaway: In crowded markets, a relentless focus on unique value proposition and service excellence can create defensible moats far more powerful than simple cost advantages, applicable even to digitally native businesses.
📖 Want the full episode breakdowns, guest details, and listen links?
