In every tech supercycle, investors are good at identifying disruptive technologies, but bad at picking the winners. AI is no different, except for the pace and velocity of growth, investment, and hype. And this time the incumbents are leading the charge. This episode works through how to separate the hype from who actually captures the enterprise value, walking through four layers of the AI investment stack, and which of these layers will exist independently or get absorbed. Joined by Kai Wu of Sparkline Capital, Jerry Neumann, retired VC investor and writer, and David Haber of a16z, we synthesize views and research to explore this very complex topic. Guests: Kai Wu, Founder & CIO, Sparkline Capital Jerry Neumann, retired VC investor and writer David Haber, General Partner, a16z Episode Sources Key Points From This Episode: ●[00:00:00]Kai Wu on the current stage of the AI infrastructure buildout and adoption cycle. ●[00:04:13]Introduction to investing across the AI stack and the challenge of capturing value from technological revolutions. ●[00:11:18]Historical technology cycles, hype cycles, and lessons from past paradigm shifts. ●[00:15:57]The “historical autopsy” of technology booms: capital misallocation, demand assumptions, and timing risks. ●[00:20:31]Comparing AI infrastructure spending with railroad and fiber-optic buildouts across history. ●[00:22:44]The AI prisoner's dilemma and why competition drives aggressive capital expenditures. ●[00:25:38]Jerry Neumann on value capture, competition, and why great technologies do not always produce great investments. ●[00:27:51]The dot-com fiber buildout, overcapacity, and how later innovators benefited from subsidized infrastructure. ●[00:33:48]Why AI may differ from previous cycles due to the dominance of well-capitalized incumbents. ●[00:36:44]David Haber on compute demand, data center utilization, and the economics behind current AI investment. ●[00:38:33]Vertical integration, competitive advantages, and how major technology companies are positioning for AI leadership. ●[00:44:06]Whether application-layer companies can survive alongside powerful infrastructure and model providers. ●[00:47:06]David Haber on enterprise AI applications, vertical software opportunities, and context-driven value creation. ●[00:49:53]Revenue growth, valuation expectations, and the sustainability of AI business models. ●[00:57:00]Open AI growth projections, demand assumptions, and the risks of extrapolating future adoption. ●[01:00:05]Aaron’s framework for analyzing the AI investment stack and its four primary layers. ●[01:06:39]The shipping container analogy and where value ultimately accumulates in transformative technologies. ●[01:12:30]Platform companies, hyper scaler investment strategies, and the defensive motivations behind AI spending. ●[01:46:13]Final investment principles, frameworks, and key takeaways for evaluating opportunities across the AI ecosystem.