What Investing in AI Really Means With Eric Chu
Or listen wherever you podcast!
Episode Description
On this episode of the On The Rise Podcast, host Jeremy Dyer talks with Eric Chu, co-founder of Trade Desk Securities and a 20-year veteran of Wall Street and digital infrastructure. In the show's first deep dive on AI, Eric explains what people are actually buying when they "invest in AI" — the software layer of large language models versus the data centers and infrastructure that power all of it. He breaks down why AI's exploding power demand is driving interest in nuclear energy, whether AI mirrors the 1990s internet, what it really takes to build an AI-ready data center, and how AI is changing investing itself. Plus practical advice for individual investors on horizon and discipline.
Summary
A career built in digital infrastructure.
Trained as an engineer with an NYU MBA, Eric moved from a hedge fund and investment banking (covering telecom and data centers) into leading a Nasdaq-listed data center company's infrastructure division — deploying nearly $1 billion with Warburg Pincus across roughly ten large-scale data center projects in Asia.
"Investing in AI" means two very different things.
Eric splits it into software (the large language models like ChatGPT or Anthropic's) and the hardware/infrastructure that supports them. Almost everything in AI — training, inference, storage, transmission — happens in data centers, so you can't really invest in AI without touching digital infrastructure.
Power demand is exploding.
Ten years ago a data center rack drew 4–6 kilowatts ("5 to 7 microwaves"). Today an AI rack can pull 90–200 kilowatts, with people already discussing megawatts per rack. A single AI data center can draw hundreds of megawatts, and demand is roughly doubling every couple of years.
That's why nuclear enters the conversation.
As facilities head toward gigawatt-scale power needs, Eric argues the traditional utility grid simply won't be enough — making nuclear reactors (and fusion) a credible direction for powering the next generation of data centers.
AI resembles the early internet.
Eric sees a reasonable parallel to the 1990s internet: few grasped its scope early, then applications exploded. He believes AI is fundamentally changing how people research and consume information across every industry — and that we're in the very early innings.
Don't chase headlines.
Advice differs by investor type. Eric's firm mainly helps institutional investors and developers moving into data centers, starting with education: converting land or a crypto-mining site into an AI-ready data center is a long journey of power, interconnection, cooling, redundancy, and uptime "nines."
AI is a tool, not an advisor.
Eric compares AI to an advanced Excel spreadsheet — great for cutting through noise, researching stocks, and even monitoring a portfolio with alerts, but not (yet) a replacement for a human advisor. Used well, it boosts productivity for investors who don't have 10–12 hours a day for research.
Your horizon is your edge.
For individual investors, Eric's top advice is to think about your investment horizon. Freedom from the short-term benchmarks that pressure institutions is an advantage — keep a consistent, unemotional plan over 10+ years and lean on tools like dollar-cost averaging.
Resources
LinkedIn: https://www.linkedin.com/in/eric-chu-cfa-7159567/
Website: https://www.linkedin.com/company/tradesk-securities/home/
