From AI Investor Panel | Chase Coleman, Brad Gerstner, Andrew Homan, Josh Wolfe | Robin Hood 2025 · · Free Engineering Courses
“All businesses in America are shifting from general purpose compute to accelerated compute, which will require significantly more compute power and benefit chip players like Nvidia and AMD.”
On , Chase Coleman, Founder & Managing Partner at Tiger Global, spoke about computing infrastructure during AI Investor Panel | Chase Coleman, Brad Gerstner, Andrew Homan, Josh Wolfe | Robin Hood 2025 on Free Engineering Courses.
At a November 2025 Robin Hood conference panel, Chase Coleman moderated a discussion on AI investment. He stated that AI capital expenditure contributes approximately 1% to U.S. GDP growth, which he described as half of the country's total GDP growth. Coleman also characterized the $3 to $4 trillion buildout of compute over the next four to five years as about 10 times the Manhattan Project, and said this private funding would significantly boost the economy. He argued that all U.S. businesses are shifting from general-purpose to accelerated compute, which he said would benefit chip companies like Nvidia and AMD. Coleman expressed the view that the semiconductor layer would capture a very significant portion of economic rent over the next decade, and described a "layer cake" analogy where semiconductors, large language models, and applications sit in tiers, with semiconductors capturing the majority of value. He also said large language models are expected to absorb or subsume the entire SaaS app ecosystem. On wealth disparity, Coleman noted that AI could exacerbate the issue but pointed to the Invest America Act, which he said would provide prosperity accounts to every child born in the U.S. starting in 2026, as a way to democratize wealth creation.