How the New Administration Will Impact Crypto, AI & Tech Globally w/ Ben Horowitz & Salim Ismail

Summary

• Horowitz’s central call is that Washington’s shift from “complete war with the industry” to “aggressive pro-tech” marks a genuine regime change for US technology. Biden-era M&A blocks, biotech constraints and crypto enforcement suppressed exits and development; Salim reported that more than two dozen companies were pushed out of banking. Regulatory relief could now release the backlog. “It’s so night and day.”
• Stargate’s scale is not established as funded fact. The conversation first calls it a quoted $500 million private-infrastructure commitment and later describes $500 billion of capital deployment; Horowitz says “it’s a little unclear what it is.” Elon Musk’s claim that Masayoshi Son had at best $10 billion exposes the financing gap. The physical bottleneck is energy, making nuclear generation near data centers and distributed power markets strategically important.
• America’s winning AI strategy is to preserve its decentralized startup engine, not imitate China’s state-directed system. China integrates AI rapidly across companies, government and the military, while GPU restrictions acted as a Darwinian force for algorithmic efficiency: DeepSeek was described as near GPT-4 performance with roughly “100x lower” cost per token. Horowitz’s conclusion: “We’ll get to the breakthroughs first. We probably won’t stop them from stealing them.”
• AI turns biology from drug discovery into drug design, although animal and human testing remain stubbornly physical bottlenecks. Diamandis says historical development costs more than $10 billion per drug and succeeds only 10% of the time; Horowitz compares AI’s role in biology to calculus giving physics a modeling language. Yet AI “doesn’t quite get rid of” in-vivo trials, and limited access to health data still restricts progress.
• Moonshot investing depends on more than polished storytelling: founders must ship, recruit and raise capital when the company looks broken. Horowitz rejects 20-person startups with eight vice presidents, favors engineers and early revenue milestones, and prizes leaders who can raise “when there’s no reason to give them money.” Charisma without operational precision fails; wealthy teams may simply “call in rich” when the mission turns painful.
• Humanoid economics could erase labor arbitrage, while AI-native companies should automate organizations faster than incumbents can restructure them. Diamandis models a $30,000 robot at $300 monthly, $10 daily or about $0.40 an hour; Horowitz cautions that the supply chain remains overwhelmingly Chinese. Ismail expects jobs to be decomposed task by task, but newly designed companies could reach the “billion-dollar, one-employee startup” far sooner.
• Crypto’s next cycle is framed as infrastructure for an AI-saturated internet, not merely another token trade. Horowitz identifies proof of humanity, deepfake provenance, machine-to-machine payments, user-held identity keys and zero-knowledge credit proofs as urgent applications now that developers have been “let out of jail.” Ismail sticks with a $250,000-$300,000 Bitcoin prediction for this year; Horowitz declines a price call because he says he is at an RIA, a registered investment firm.

Deep dive

Not yet available upstream; scheduled sync will retry.