AGI’s Economic Paradox

While many foresee artificial general intelligence leading to deflation, a Harvard economist argues its massive energy and infrastructure needs could paradoxically fuel inflation. This debate has profound implications for central banks and the future of interest rates.

The economic crystal ball, perpetually hazy, seems particularly clouded when it comes to the advent of artificial general intelligence.

For months, perhaps even years, a prevailing narrative has taken root: AGI, with its promise of unprecedented efficiency and automation, will be a potent force for deflation.

It is a vision painted by tech titans and macroeconomists alike, suggesting a future where goods and services become ever cheaper, driven by the tireless, cost-effective algorithms of an intelligent machine age.

Indeed, OpenAI CEO Sam Altman has championed this view, describing AI’s deflationary impact as underappreciated and misunderstood.

Investor Raoul Pal went further, labeling the technology a deflationary nuclear bomb, predicting a future where electricity is the other one to become astonishingly cheap.

But a powerful dissenting voice has emerged from the hallowed halls of academia, challenging this widely accepted gospel.

Ken Rogoff, the esteemed Harvard professor and former Chief Economist of the International Monetary Fund, offers a starkly contrarian prognosis.

Speaking on the Dwarkesh Patel Podcast, Rogoff warns that the very power that makes AGI so transformative — its insatiable hunger for energy and infrastructure — could paradoxically push inflation higher, forcing interest rates to remain elevated for longer than many anticipate.

AGI and AI are upward pressures on interest rates, Rogoff stated unequivocally.

His reasoning is rooted in a fundamental understanding of capital investment.

The deployment of AGI, he argues, isn’t merely about software; it is about a colossal, unprecedented build-out of physical infrastructure.

Think hyperscale data centers stretching across continents, demanding dedicated power grids, advanced cooling systems, and a vast network of fiber optics.

This isn’t just a digital revolution; it is an intensely material one.

With the huge energy needs and the capital investment, you’re going to see even more spending, not less, he explained, dismantling the notion that efficiency inherently leads to lower overall costs.

Rogoff’s perspective directly confronts the simplistic view that automation always leads to cheaper outputs.

He draws on historical precedents, suggesting that major technological shifts, particularly those requiring significant upfront capital, often involve a period of intense investment that can stimulate demand and, consequently, inflation.

He cites research by MIT economist Daron Acemoglu, which demonstrates that automation doesn’t necessarily depress wages or reduce investment.

Instead, it can redirect economic activity toward capital-intensive growth paths.

This isn’t a story of a silent, cost-cutting machine slipping seamlessly into existing structures; it is a wholesale re-engineering of the global economy, demanding immense resources.

The implications of Rogoff’s warning are profound, particularly for central banks.

For years, these institutions have grappled with the specter of inflation, often battling it with higher interest rates – a blunt instrument that can stifle economic growth.

If AI, the very technology hailed as a panacea for productivity, actually fuels inflationary pressures, central bankers face an even more complex and unenviable task.

The traditional playbook, which might anticipate lower rates as technology drives down prices, would be rendered obsolete.

This could actually make life harder for central banks, not easier, Rogoff concluded, underscoring the potential for significant policy challenges.

Consider the sheer scale of the energy demands.

Training and running advanced AI models consume staggering amounts of electricity, comparable to the output of small nations.

As more companies and sectors adopt AGI, the demand for power will skyrocket, placing immense strain on existing grids and necessitating massive investments in new energy generation and transmission infrastructure.

This isn’t cheap.

Furthermore, the specialized chips, cooling systems, and physical space required for these computational behemoths represent significant capital expenditures.

These are costs that trickle through the supply chain, affecting everything from raw materials to skilled labor, ultimately contributing to upward price pressures.

The divergence of opinion between figures like Rogoff and the deflationary proponents highlights the inherent uncertainty in forecasting the economic impact of truly disruptive technologies.

Is AI a force that will relentlessly drive down costs across the board, or is it a capital-intensive beast that will necessitate a massive reallocation of resources, leading to a period of inflationary investment?

The answer will shape monetary policy, investment strategies, and ultimately, the financial landscape for decades to come.

The stakes are undeniably high, and as the digital age hurtles forward, the debate over AI’s true economic signature promises to be one of the most critical of our time.