AI: the race for power – and its limits

In Le Temps, Daniel Steck, Analyst at Piguet Galland, deciphers the frenzy of investment around AI and the risks of overheating.

An investment frenzy that raises questions

Artificial intelligence is no longer just making headlines – it’s now attracting colossal investment flows. Tech giants are locked in a race for computing power, fuelling an unprecedented wave of infrastructure spending.

Data centres, electronic chips, cooling systems… this “construction phase” is mobilising not only technology specialists but also players in the energy and industrial equipment sectors.

“We are currently in the phase of building the tools for AI, with massive investments in infrastructure and a race for power through data centres and semiconductors,” explains Daniel Steck, technology analyst at Piguet Galland.

When “picks and shovels” become strategic

As in previous industrial revolutions, the first winners are not necessarily those who use AI, but those who provide the tools to develop it.

Energy suppliers, thermal equipment manufacturers and electronics specialists are directly benefiting from this investment boom.

“The utilities sector, traditionally defensive, has become a growth industry over the past two years,” adds Daniel Steck.

In Switzerland, Belimo – active in heating and air-conditioning systems – has seen its results surge, driven by growing demand from data centres. In other words, the AI wave is now spilling over into vast areas of the economy.

A risk of overheating not to be underestimated

While the potential of AI is undeniable, its current valuation nonetheless raises concerns.

The “Magnificent Seven” – the seven tech giants that together account for over a third of the S&P 500 – are posting impressive performances, yet their capital expenditure forecasts are soaring. For some investors, this evokes memories of the excesses of the dot-com bubble in the early 2000s.

“There is a risk that the momentum could fade if corporate earnings fail to grow fast enough to justify these investments,” warns Daniel Steck.

A consolidation phase could therefore emerge, with more realistic expectations and valuations better grounded in fundamentals.

Profitability: the ultimate litmus test

At the heart of the matter lies one key question: can companies turn these massive investments into tangible revenues?

“The central issue is whether the massive investments we’re seeing will prove profitable in the long term. To achieve this, companies will need to generate billions in revenue from their end products,” notes Daniel Steck.

While the paid version of ChatGPT remains relatively marginal, Microsoft is already succeeding in monetising its AI solutions, notably through Copilot. Meanwhile, Meta and Alphabet could enhance the profitability of their advertising platforms thanks to AI, thereby justifying higher pricing.

Between promise and prudence

Artificial intelligence is opening a new technological chapter – one whose scale surpasses that of cloud computing or mobile internet.

But like every revolution, it comes with cycles of euphoria and correction.

For investors, the challenge is to identify the sustainable winners of this transformation — those who, beyond the media buzz, will build the economic foundations of the AI era.

📖 Read the full article in Le Temps (french only)