AI fever pushes markets to extremes but investors must keep nerve

Pic via Getty Images

Artificial intelligence has become the engine of the global equity rally. Chipmakers, cloud platforms and software giants linked to AI are shattering records, propelling major indices to all-time highs.

I admire innovation as much as anyone, but valuations are now so stretched that investors must separate genuine growth from speculative mania.

The numbers speak for themselves. Nvidia has climbed nearly 65 % year to date, giving it a market value just shy of $4 trillion, more than the combined capitalisation of the entire German DAX.

Microsoft’s market cap has surged past $4.1 trillion, trading at over 35 times forward earnings, a multiple last seen during the late-1990s dot-com boom.

Semiconductor Equipment leader ASML has gained 48 % in 2025, pushing its price-to-sales ratio to a record 18. In Asia, Taiwan Semiconductor Manufacturing Company (TSMC) now trades at 32 times forward earnings, up from 22 at the start of the year.

These are extraordinary businesses, but even extraordinary businesses cannot defy economic gravity.

History offers sobering parallels. In 1999, Cisco briefly became the world’s most valuable company on hopes it would supply the ‘plumbing’ for the internet. The tech was transformative, but the share price still fell almost 80 % when expectations outran reality.

Today’s enthusiasm rests on real breakthroughs. AI is reshaping industries from healthcare to logistics. Nvidia’s latest quarterly report showed $34 billion in data-centre revenue, up 220 % from a year earlier. Microsoft’s Azure AI bookings are running at an annualised $140 billion, double last year’s pace.

These numbers are spectacular. Yet, even if such growth persists for several years, current valuations already price in a near-perfect future.

Liquidity is the accelerant. Investors are convinced that the Federal Reserve will deliver at least three rate cuts before year-end, while the European Central Bank has already trimmed rates twice in 2025.

Cheaper money fuels appetite for risk and compresses discount rates, pushing growth-stock valuations higher. But liquidity can reverse quickly. A single hot inflation print or a hawkish Fed shift could send yields spiking and trigger a rapid re-pricing.

Meanwhile, costs are creeping up. Tariffs imposed by President Donald Trump on key Chinese technology inputs in August are starting to filter through semiconductor supply chains. If margins compress while valuations remain at extremes, earnings disappointments could be punished severely.

Investors should not assume AI leadership will remain concentrated in today’s favourites.

History teaches us that dominant platforms often face unexpected challengers or regulatory hurdles.

The European Union’s AI Act comes into force early next year with strict data-handling and transparency requirements. Washington is debating antitrust action as the biggest firms consolidate computing power and proprietary models. Such measures could slow growth or raise compliance costs.

I’m not advocating that investors abandon the sector, indeed, quite the opposite. AI is a multi-decade trend and will generate tremendous long-term value.

The question is how to capture that value without taking reckless risk.

First, diversify. Many portfolios are dangerously overweight in a handful of mega-cap US tech names. Investors should broaden exposure across geographies and sectors. High-quality industrials, healthcare innovators and select financial firms offer attractive earnings growth at far lower multiples.

Second, stress-test. Model scenarios where discount rates rise by 100 basis points, or where AI revenue growth slows to half the current consensus. If a portfolio cannot withstand those shocks, it is too concentrated.

Third, maintain liquidity. Holding a buffer of cash or short-duration bonds provides flexibility to buy when volatility creates genuine bargains. In a correction, those who can deploy capital quickly will benefit most.

Finally, seek independent advice. Market narratives can become intoxicating, and confirmation bias is powerful when everyone seems to be making money. Objective analysis helps ensure strategies remain aligned with long-term goals.

Despite the risks, I remain increasingly optimistic about innovation. Companies that combine strong balance sheets, sustainable free cash flow and genuine technological leadership will continue to create wealth.

However, investors must distinguish between owning the future and overpaying for it.

The lesson from past bubbles is clear that tech revolutions endure, but valuations do not. Those who chase momentum without discipline risk turning tomorrow’s breakthroughs into today’s losses. Extraordinary opportunities demand extraordinary prudence.

Nigel Green, is the group CEO and founder of deVere Group, an independent global financial consultancy.

The views, information, or opinions expressed in the interviews in this article are solely those of the author and do not represent the views of Stockhead.

Stockhead does not provide, endorse or otherwise assume responsibility for any financial advice contained in this article.