Three stories landed within hours of each other this week and they form an uncomfortable triptych. Databricks hit a $188 billion valuation. Index Ventures co-founder Neil Rimer predicted that the historic AI wealth being generated in Silicon Valley will have to be redistributed outward. And The Atlantic argued that the business of AI companies now looks more like oil refining than software development: capital-intensive, infrastructure-bound, geographically fixed. Put these three together and you get a portrait of an industry that has quietly stopped being tech.
From Code to Capital Stack
The old Silicon Valley mythology was frictionless: a few smart people, a laptop, a dorm room, infinite scale. What Databricks, Microsoft Azure, and the data center buildout represent is the death of that myth. You cannot run a frontier AI model without gigawatts of power, rare earth materials, and industrial-scale cooling. The Atlantic's framing, tech as oil refining, isn't a metaphor so much as a structural diagnosis. And Rimer's redistribution thesis follows directly: when you're running a capital-intensive industry, the windfalls don't stay local. They ripple. TurboFund's live VC intelligence has been tracking exactly this shift in where AI investment is flowing across sectors.
The Second Act Problem
Databricks is specifically described as AI's 'favorite second act,' a data infrastructure company that successfully rebranded. This is the pattern now: legacy enterprise tech finds an AI frame and gets revalued. The question Rimer raises, where does the money actually go, is partly answered by the Australian startup story running in parallel. Stripe's Startup Battlefield in Sydney is a direct bet that the redistribution flows outward geographically, that the next layer of AI value capture happens outside the Valley. Whether that's redistribution or just expansion is the question nobody wants to answer directly.