Artificial intelligence (AI) has become the defining investment story of the decade. It promises to reinvent industries, boost productivity, and create a new generation of trillion-dollar firms.

Yet beneath the optimism lies a growing unease. Are investors once again mistaking technological progress for guaranteed profits?

At a recent BizNews Conference, Shaun Peche, founder of Ranmore Fund Management, which is ranked in the top one percent of asset managers globally by performance, raised precisely that question. Peche argued that AI’s financial trajectory is beginning to resemble a classic asset bubble. In his view, the current wave of enthusiasm is driven less by real cash flows than by narrative and cheap capital.

From easy money to infinite optimism

To understand the backdrop, it helps to recall the decade of near-zero interest rates that followed the global financial crisis. For years, investors searching for yield moved beyond the traditional 60/40 split (where 60% of capital is allocated into equities and 40% into bonds) into: “alternative” assets such as private equity, venture capital, and hedge funds.

In this environment of cheap debt, private equity firms borrowed freely to buy companies, inflate valuations, and engineer rapid exits. Venture capitalists pursued the next big story, from software to space to crypto.

AI, Peche suggested, may be the newest chapter in this same pursuit of returns. He explained: “When you see those hockey-stick growth charts, it doesn’t matter what the asset class is. The returns go to zero.”

The arithmetic of hype

Peche’s concern is not about AI’s potential but about its financial logic. Consider Anthropic, one of the leading AI labs, recently valued at about $183 billion on annualised revenue of only $5 billion.

Then there is the issue of infrastructure. The world’s major technology firms are spending about $400 billion per year on data centres and chips to power AI models. Depreciated over a decade, that equals roughly $40 billion in annual costs, before electricity, labour, and software development. Peche observed that even if AI is transformative: “the math doesn’t yet work.”

Competition is another problem. The market is crowded with players such as OpenAI, Anthropic, Mistral, Grok, and DeepSeek, all offering similar services. If one raises prices, customers can easily switch. The monopoly profits many investors anticipate may never arrive.

Another challenge is that most AI firms remain private. Ordinary investors, such as those saving through pension funds or exchange-traded funds, cannot access them. Any gains are confined to a narrow circle of private investors.

The question of sustainability

None of this dismisses AI’s potential. The technology is real, and its applications, from automating code to drug discovery, are expanding. However, Peche’s warning is that technological revolutions and financial booms are not the same thing. Markets often confuse innovation with profitability, just as they did with railroads in the 19th century or dot-coms a century later.

If AI investment continues at its current pace, the question becomes whether the expected returns can ever justify the scale of spending. Most firms adopting AI tools report limited productivity gains so far. And while the major technology companies have the balance sheets to fund the build-out, their cash reserves are shrinking under the weight of buybacks and capital expenditure.

Lessons from history

Peche’s message was not alarmist but cautious. He reminded the audience that: “greed builds slowly; fear is instant.” The time to think critically, he argued, is before confidence collapses.

The question remains unresolved: is this the beginning of a genuine industrial revolution, or another case of financial exuberance disguised as innovation?

If history is any guide, the technology will endure. The valuations, perhaps not.