#753: The Economy Is AI Now with Jordi Visser

— AI demand is genuinely exponential, not speculative; physical infrastructure bottlenecks (energy, chips, data centers) are the real constraint, not economics or technology limitations.

— The agentic economy launched in November 2024 with Claude Opus 4.5; this represents a shift from pre-training to deployment, effectively adding 7.5 billion "digital consumers" overnight and explaining parabolic growth in token demand and compute spending.

— Portfolio construction is being rewritten; traditional 60/40 portfolios are obsolete as AI-driven equities now dominate GDP growth and stock market gains disproportionately benefit Americans through direct and indirect wealth channels.

— Incumbent skepticism stems from psychological ego protection—admitting being wrong on AI threatens public credibility, making it harder for established voices to pivot narratives despite clear evidence of exponential progress.

— Tokenization and cryptocurrency will absorb financial volumes the traditional system cannot handle; stable coins, real-world asset tokens, and Bitcoin's infrastructure are necessary plumbing for the agentic economy.

— SaaS businesses face compression unless they transition from seat-based models to API-native, agent-friendly architectures; companies that fail this transition will face margin pressure and multiple compression during buildout.

— Use AI for concrete skill-building and personal productivity: Whether learning options theory, building custom memory systems for your business, or deploying agents, AI compounds advantage fastest for those who use it daily; waiting for regulatory clarity or perfect understanding is a losing trade.

— Diversify away from illiquid traditional assets: Private equity, private credit, and real estate have been safe havens from volatility precisely because they lacked price discovery; tokenization will end that protection. Consider small allocations (1–3%) to digital assets and real-world asset tokens as traditional buckets face repricing.

— Watch infrastructure constraints, not valuations: The real risk to AI upside is not a bubble but a bottleneck—power generation, chip availability, or narrative disruption from politicians slowing data center builds. Monitor gigawatt-scale data center delays and energy pricing as leading indicators; parabolic charts are normal in exponential worlds.