Seven things the tech sell-off tells us — and three things it doesn't
- Robin Powell
- 46 minutes ago
- 9 min read
Tech and AI stocks have suffered their sharpest falls in months, with more than $800 billion wiped from software stocks alone. Commentators are scrambling to explain what it all means. Here are seven things the tech sell-off tells evidence-based investors, and three things it doesn't.
You may have been watching the numbers this week with a knot in your stomach. Nvidia down. Palantir down 11.6% in a single session. AMD losing 17.3%, its worst day since 2017. The London Stock Exchange Group shed 12.8% on Tuesday. Relx dropped 14.4%. WPP fell 11.8%. And it wasn't just tech. Bitcoin slid below $67,000. The VIX — a measure of expected market volatility — hit its highest level since November.
If you own a global index fund, you felt this. Maybe more than you expected to. The seven largest stocks in the S&P 500 now account for around a third of the index. When they fall, your "diversified" portfolio falls with them.
But here's what set this week apart. The sell-off wasn't triggered by AI failing. It was triggered by AI working. When Anthropic released new tools that automate legal, marketing, and finance tasks, investors didn't panic about the technology. They panicked about what it means for the companies they assumed would profit from it. Software and services stocks have lost more than $800 billion in market value since late January, according to Reuters.
We've seen this before. GPS navigation was a real revolution. It killed the paper map, gutted the standalone sat-nav market, and enabled businesses that didn't exist a decade earlier. But investors who backed the obvious winners got burned. TomTom's share price collapsed by more than 90% from its 2007 peak. The real value went to companies that used the technology — Uber, Deliveroo, Google Maps — not the ones that built it.
So what does this sell-off tell us? And what doesn't it?
What the tech sell-off teaches us
1. Technology revolutions have always been better for users than for investors
This is the most important lesson, and it applies far beyond AI. Technologies that change the world enrich the people who use them, not the people who invest in them. The pattern has repeated for two centuries.
British railways changed everything about how people and goods moved around the country. By the 1840s, investors were pouring money into railway stocks, convinced that a technology this powerful had to generate extraordinary returns. They were half right. The technology was powerful. But railway mania left most investors ruined, even as Britain got a world-class transport network.
The canals those railways replaced? Their investors lost everything too.
The internet followed the same script. Cisco, the company that built much of the internet's physical infrastructure, fell roughly 90% from its 2000 peak and took over 20 years to recover in nominal terms. Amazon dropped 94% before eventually becoming one of the most valuable companies in history. Nobody in 2001 could have told you which was the survivor and which was the cautionary tale.
The economist William Nordhaus estimated that innovators capture only around 2.2% of the total value their innovations create. The rest flows to consumers and to companies that find ways to apply the technology — often years later, in ways nobody predicted.
AI fits this pattern uncomfortably well. The technology is almost certainly real. The question is whether today's investors are buying the equivalent of railway shares in 1845 or Cisco in 2000.
2. The market is trying to pick AI winners in real time — and getting it wrong
This sell-off isn't a blanket rejection of AI. It's investors frantically sorting companies into two buckets — those AI will help and those it will destroy — with incomplete information and a lot of fear.
Look at what happened this week. On Thursday, while the Nasdaq cratered, Broadcom jumped nearly 4% and Nvidia edged up 0.5%. Investors decided these were infrastructure winners. The same day, AMD fell 17.3%, Palantir dropped 11.6%, and SanDisk lost 16%. The JPMorgan index tracking US software stocks is down 18% since the start of the year.
The market has landed on a neat story: hardware and chips good, software and services bad. Mike Zigmont at Visdom Investment Group put it bluntly to the Financial Times: "This concept that AI will cannibalise the most tech-savvy firms first is an interesting new narrative in the AI story arc."
But that's all it is. A narrative.
Today's infrastructure play could become tomorrow's commodity as competition drives prices down. And plenty of the software companies being sold off will adapt and survive. Marija Veitmane, head of equity research at State Street, told the FT the market's reaction was "overly pessimistic" and that the real story is "incremental improvement, rather than total revolution."
She might be right. She might be wrong. That's the point. Nobody can reliably sort winners from losers mid-revolution. If they could, Cisco investors would have sold in 1999 and Amazon investors would have held through a 94% drawdown. Almost nobody did either.
3. Valuations were priced for perfection — and perfection didn't arrive
Many AI-linked stocks needed everything to go right to justify their prices. The moment anything wobbled, those valuations cracked.
Take Alphabet. The company beat earnings expectations, grew revenue 18%, and posted a 48% jump in cloud income. A strong quarter, on paper. But when it announced capex plans of $175–185 billion for 2026 — roughly 50% higher than Wall Street's $120 billion estimate — shares fell 5%. Investors weren't punishing bad results. They were repricing the cost of chasing good ones.
Qualcomm told a different version of the same story. Solid first-quarter numbers, but a weak forecast dragged shares down 10% as memory shortages — driven partly by AI's appetite for data centre components — rippled through the supply chain.
That's the trap with momentum-driven markets. When prices already reflect the most optimistic outcome, there's no room for "good but not perfect". Any friction — rising costs, supply constraints, competitive pressure — turns expensive into fragile.
Academic research has long flagged this. Companies that aggressively expand their asset base tend to deliver lower subsequent returns — a pattern captured by the investment factor identified by Fama and French. When firms pour billions into expansion, the market overestimates the returns that spending will generate.
4. Narrative-driven investing is dangerous — especially when the narrative is true
The most dangerous investment stories aren't the obviously wrong ones. They're the ones that are partly right. AI is reshaping industries. That doesn't make AI stocks good value.
There's a reason investors keep falling for this. AI headlines are everywhere, which triggers availability bias — we overweight information that's easy to recall. CEOs talk up AI's potential at every earnings call, feeding authority bias. Recent outperformance gets extrapolated into future returns through recency bias. And the fear of missing out keeps money flowing in long after valuations have stopped making sense.
As Deutsche Bank strategist Jim Reid observed this week, there's been "a clear shift in markets from AI euphoria towards more differentiation between companies, and growing concern about its disruption to existing business models."
Most investors didn't buy AI stocks because they'd run a discounted cash flow model and decided the price was attractive. They bought because everybody else was talking about it. That's how narratives work — they bypass analysis entirely.
Research on bubble psychology tells the same story every time. Investors collectively convince themselves that "this time is different," not because the evidence supports it, but because the story feels compelling. Sam Altman, the CEO of OpenAI, has himself described the AI market as a bubble. When the people building the technology are warning you, that should give any investor pause.
5. Your "diversified" index fund may be less diversified than you think
Market-cap-weighted index funds have become heavily concentrated in a handful of tech stocks. That's not a reason to abandon indexing — but it is a reason to understand what you own.
The Magnificent Seven now account for roughly 34% of the S&P 500, up from around 12% a decade ago. If you own a global index tracker, the picture isn't much better. US stocks make up about 72% of the MSCI World Index, according to iShares data as of December 2025 — well above the long-term average of 57%, as State Street has noted. A sell-off in a handful of Californian tech companies can send shockwaves through a fund that's supposed to give you exposure to 23 countries.
Does this mean indexing is broken? Absolutely not. Active fund managers hold the same stocks in roughly the same proportions — they have to, because straying too far from the benchmark is career risk. And deliberately tilting away from mega-caps carries its own dangers: you might dodge the crash, but you'll also miss the rally that precedes it.
The real lesson is simpler. If your entire equity allocation sits in a single market-cap-weighted global tracker, you're more exposed to US tech than you probably realise. Better to know that now than discover it during the next sell-off.
6. Sell-offs are when behavioural discipline matters most
The urge to sell after prices have already fallen is the single most expensive instinct in investing.
Year after year, Dalbar's Quantitative Analysis of Investor Behavior reports show the same thing: the average investor earns less than the funds they invest in, largely because they sell after falls and buy after rallies. The precise size of this gap is debated, with some researchers putting it lower than Dalbar does, but the direction is not.
There's a neurobiological reason it feels so hard to resist. When your portfolio drops sharply, your amygdala triggers a threat response. Cortisol floods your system. Your brain, wired to escape physical danger, screams at you to do something. In evolutionary terms, that response kept your ancestors alive. In financial terms, it reliably makes things worse.
The investors who'll suffer most from this sell-off aren't the ones who bought AI stocks at the top. They're the ones who panic-sell at the bottom.
7. The case for boring, low-cost, globally diversified investing got stronger
You don't need to predict which technologies will succeed, which companies will win, or when bubbles will burst. A globally diversified, low-cost portfolio removes the need to get any of those calls right.
Broad global index funds give you exposure to thousands of companies across dozens of countries. Geographic diversification — not just the US — means no single market's fortunes dictate your retirement. Regular rebalancing enforces the discipline of trimming winners and topping up laggards. And low costs protect the one variable you can control.
Harry Markowitz called diversification the only free lunch in finance. Weeks like this one show why.
What the tech sell-off doesn't tell us
It doesn't tell us the bubble has burst
A sharp sell-off feels like a verdict. It isn't. It's a data point — and a single data point tells you almost nothing about what comes next.
Consider the LTCM crisis in 1998. The Nasdaq dropped sharply that autumn, and plenty of commentators declared the party over. It wasn't. The index went on to reach extraordinary new highs before the dotcom bubble finally burst two years later.
So is this moment 1998 or 2000? Nobody knows. And anyone who claims otherwise is selling false certainty — which, in markets, is one of the most expensive products you can buy.
It doesn't tell us AI is overhyped
It's tempting to read falling software stocks as proof that AI was all smoke. But the sell-off isn't about the technology failing. If anything, it reflects the opposite fear — that AI works so well it's driving down prices, compressing margins, and undermining the very business models investors were betting on.
A technology can be world-changing and still a terrible investment at the wrong price. Railways reshaped Britain; most railway investors lost their shirts. The technology and the trade are not the same thing, and confusing the two leads to poor conclusions in both directions.
It doesn't tell us it's time to buy the dip
A lower price is not the same as a good price. "Cheaper than last week" tells you where something was, not what it's worth. Shops that mark down overpriced goods aren't offering bargains — they're correcting a mistake.
The questions that matter haven't changed. What are the realistic future returns? How fierce is the competition? How much exposure does an investor already carry to this part of the market? None of those answers come from watching a price fall. They require the kind of slow, unglamorous analysis that rarely fits in a headline.
The quiet advantage
Technology revolutions are real. They reshape industries, create new ones, and render old certainties obsolete. But the history of investing in those revolutions is far messier than the history of the revolutions themselves.
The investors who fare best over the long run aren't the ones who correctly called the top, bought the dip, or predicted which companies would survive. They're the ones who built portfolios that didn't require them to get any of those calls right. Diversified, low-cost, and patient — not because it's exciting, but because it works when the exciting stuff stops working.
That's the quiet advantage no sell-off can take away.
Resources
Fama, E. F., & French, K. R. (2015). A five-factor model of expected stock returns. Journal of Financial Economics, 116(1), 1-22.
Nordhaus, W. D. (2004). Schumpeterian profits in the American economy: Theory and measurement. NBER Working Paper No. 10433.
You understand the evidence. Now find an adviser who believes it too
Our Find an adviser directory lists professionals committed to low-cost, globally diversified portfolios and straight answers. Connect with someone who'll respect your knowledge instead of trying to sell you expensive active funds. Search your area now.
Tired of YouTube investment channels pushing expensive products?
Subscribe to our channel for content that actually serves your interests. Evidence-based insights, clear explanations, zero sales pitch. New videos weekly on the investing approach that really works.