Идеи
AI demand drives unprecedented infrastructure buildout
Tom Brown states AI capabilities have been increasing rapidly and consistently, AI demand is exponential and multiplying every year, and this is creating an industrial buildout larger than the 1800s railroads. That supports a broad long-term investment case for AI compute and data center infrastructure.
Data center power shortage supports infrastructure
Tom Brown says compute is clearly the bottleneck for AI progress, with a huge shortage of power and construction labor. Countries that permit land and support data center construction can capture jobs, tax revenue, and additional grid power, supporting data center infrastructure and power/utility investments.
AI costs collapse, adoption and value accelerate
Tom Brown notes the economic value from frontier AI labs is doubling about every six months while the cost to access last year's best model has fallen roughly 20x. This cost collapse should cause AI models to proliferate into more use cases and support broad AI software and services adoption.