Artificial intelligence is poised to usher in a new wave of entrepreneurship, while simultaneously reshaping competitive advantages across industries, according to Citadel founder and CEO Ken Griffin.

Speaking on a recent Goldman Sachs’ Exchanges podcast, Griffin said the rapid development of agentic AI systems is lowering barriers to entry for startups, allowing small teams to compete with established companies at unprecedented speed.

At the same time, the soaring cost of AI infrastructure is creating a new class of competitive moat for companies with the financial resources to invest heavily in computing power, he said.

“The competitive moats of our corporate society are all being filled at breathtaking rates. We’re likely to see a golden age of entrepreneurial activity,” Griffin said.

He illustrated AI’s capabilities with an internal project at Citadel designed to automate the replication of academic finance research.

Traditionally, reproducing a published finance paper required six to eight weeks of work by master’s or PhD-level researchers, he said. Citadel built an agentic AI system capable of reading a paper, reproducing its findings, verifying the published results, testing them out-of-sample and completing the entire process in just two to three hours.

The experience highlighted how quickly AI is moving beyond routine automation into highly specialized knowledge work.

“This is not just a white-collar job. This is a master’s or PhD-level job,” Griffin said.

Despite the productivity gains, Griffin said Citadel has not reduced staffing levels.

Instead, the firm has redirected employees toward tackling a broader range of investment opportunities.

“I’ll take every single productivity gain I can get because with the talented people we have, we just have more to go after,” he said.

Griffin argued that while some professions, such as language translation, may face significant disruption, many companies will use AI to expand capacity rather than simply eliminate jobs.

The broader implication, Griffin said, is that AI will weaken traditional competitive advantages enjoyed by established firms.

Entrepreneurs can now build businesses with dramatically smaller teams by deploying agentic AI systems to perform work that previously required dozens of employees, he said.

Griffin described one startup that would ordinarily have needed 30 to 40 staff but is operating with only a handful of people by relying heavily on AI agents.

“We’re going to see a lot of these stories come to light over the next couple of years as entrepreneurs embrace this technology to really take on some very interesting opportunities,” he said.

As AI handles more operational and analytical tasks, Griffin expects innovation cycles to accelerate and incumbents to face increasing competitive pressure from leaner, AI-enabled challengers.

Compute becomes the new strategic asset

While AI may reduce barriers to launching businesses, Griffin believes access to computing infrastructure will increasingly separate winners from losers.

Citadel and Citadel Securities have used machine learning for roughly a decade, beginning with TensorFlow and later incorporating transformer-based models to improve pricing, trading and risk management.

Griffin argued that available compute is effectively fully utilized, reflecting the rapid adoption of AI tools across the economy.

“We compete for compute access with everybody,” Griffin said.

The result has been sharp inflation in computing costs, with large market-making firms now spending hundreds of millions of dollars annually on compute infrastructure.

“The price of compute has certainly gone up per unit of compute beyond where people would have reasonably projected it two or three years ago,” Griffin said.

High-margin firms can absorb these expenses, he noted, while businesses with thinner margins may struggle to keep pace, effectively creating a new competitive moat built around AI infrastructure rather than software alone.

Data centers become a strategic priority

Meeting the surging demand for AI will require a significant expansion of data center capacity, Griffin argued, and he believes the United States should ensure those facilities are built domestically.

With AI workloads driving unprecedented demand for electricity, Griffin said new data centers should be paired with dedicated power generation rather than placing additional strain on existing electricity grids.

“Build the corresponding required power generation. Tie the generator to the grid…but build that data center in America,” he said.

Griffin also advocated for a renewed focus on nuclear power, including small modular reactors, as a long-term solution for powering AI infrastructure, while describing abundant US natural gas resources as an important near-term advantage.

Failing to build sufficient domestic AI infrastructure, he warned, could leave the US dependent on foreign providers for one of the most strategically important technologies of the coming decades.

“They’re going to get built. Do you want them built here in America? Or do you want them built abroad?” Griffin said.