Investing in AI in 2026: ETFs, stocks, or a speculative bubble?

Published 5 June 2026 — Artificial intelligence is the hottest investing topic of the year. Everyone wants "their slice of Nvidia". But before you rush into an AI ETF or a star stock, one question matters: are you investing in the future… or buying at the top of a bubble? The honest take, no hype.

You're probably already invested in AI

This is the point thematic-ETF sellers avoid mentioning. A global equity ETF like the MSCI World or MSCI ACWI already includes the giants that are AI today: Microsoft (OpenAI, Copilot), Nvidia (the chips), Alphabet (Gemini), Amazon (AWS cloud), Meta, and the entire semiconductor chain.

In other words: with a simple world ETF in your PEA, you already benefit from AI's growth — diversified, low-cost, and without a concentrated bet on a single theme. For many investors, that's more than enough.

The 3 ways to get more AI exposure

1. Thematic AI ETFs

They bundle a basket of companies tied to AI and robotics. The most followed in 2026:

Ongoing charges generally run from 0.35% to 0.50%/yr: pricier than a standard world ETF. Important: most are not PEA-eligible, except the Amundi MSCI Robotics & AI (LU1861132840), which can be held in a PEA.

2. Individual stocks (Nvidia & co)

Buying Nvidia, Microsoft or TSMC directly. The potential is real: Nvidia shows revenue growth above 40% per year and operating margins exceeding 50%. But a single stock also means extreme volatility and timing risk: you might buy right at the top.

3. Do nothing more

An underrated but often wisest option: keep your world ETF and add no bet. You capture AI's growth without over-concentration or stress.

The uncomfortable word: bubble

Impossible to discuss this honestly without raising it. AI's momentum is real, but two signals call for caution:

• Many companies in the sector are not yet profitable: their value rests on future promises.
• Valuations echo the dot-com bubble of 2000: prices already factor in very ambitious growth scenarios.

The lesson of 2000: even companies that ended up dominating the world (Amazon, etc.) saw their share price fall 80-90% before rebounding years later. Being "right" about AI doesn't protect you from buying at the wrong price.

The sensible approach: core-satellite

If AI excites you and you want more exposure, do it with method rather than with this month's paycheck:

An AI ETF does not replace a diversified core portfolio: its sector concentration means higher volatility. It complements it, as a small satellite sleeve you accept seeing fall sharply.

What not to do

• Bet everything on one AI stock out of fear of missing out.
• Stack an AI ETF on top of a world ETF without realising you're doubling the same exposure.
• Invest money you'll need short-term in such a volatile sector.
• Confuse "great technology" with "great investment": the price you pay matters as much as the quality.

Frequently asked questions

AI ETF or individual stocks — which is better?

For a beginner, an AI ETF is clearly less risky than a single stock: it spreads the risk across dozens of companies. An individual stock can outperform… or collapse. If you're not ready to analyse balance sheets and stomach 50% drops, the ETF is the more sensible route.

Does DCA work on an AI ETF?

Yes. Investing a small amount regularly rather than a lump sum smooths your buying price and reduces the risk of buying the top — especially relevant in such a turbulent sector. It's the same logic as DCA applied to standard ETFs.