News

Key takeaways:

Yesterday’s session felt like an “unwind the AI trade” day. Major US indices fell, with the tech-heavy benchmarks leading the way and more defensive areas holding up relatively better.

What just happened in markets

The S&P 500, Nasdaq and Dow all closed lower, but the gap between them told the story. The Nasdaq, packed with AI and growth stocks, dropped the most in % terms, while the Dow, which has more traditional industries, fell less. That is what you would expect if investors are taking profits in crowded technology trades rather than pricing in a full-blown recession.

Under the surface, chipmakers and AI-linked names saw some of the sharpest moves. Nvidia swung from early gains to a clear loss by the close. Several other semiconductor and cloud-related stocks also fell by more than the broader market. In contrast, defensive sectors such as utilities and consumer staples were flat to slightly positive.

Macro data and yields added noise but not a clear disaster signal. A mixed jobs report and shifting expectations for central-bank rate cuts gave traders a reason to reassess how much good news was already in the price, especially for high-valuation growth stories.

Why this sell-off feels different (and how much it really changes)

This drop felt different because it hit on a “good news” day for AI. Strong earnings and upbeat guidance from key names were not enough to keep prices rising. That usually means positioning and expectations have run very hot. Many investors were already leaning the same way, so any wobble triggered profit taking.

It also highlighted how concentrated and expensive parts of the market have become. Based on our internal analysis on Bloomberg, the S&P 500 now trades around 24 times expected earnings, near its highest level in the past five years and well above its 10-year average near the high teens. The technology sector sits even richer, at about 32 times forward earnings, versus a 10-year average in the low 20s. In simple terms, investors are paying a much higher price than usual for each dollar of future profits, especially in tech and AI. Current P/E vs. historical averages

AI-exposed giants are at the heart of this. Large “hyperscalers” such as Microsoft, Alphabet, Amazon and Meta trade around 26 times expected two-year-ahead earnings on average, which is far below the near-70 times seen for the top tech names at the peak of the dot-com bubble but still far from cheap. Nvidia, the poster child of the AI build-out, also trades on a hefty forward price-to-earnings multiple of 27x, well above the broader market.

Valuation: what history is really telling us

History offers a useful reminder. The last time the S&P 500’s forward price-to-earnings (P/E) ratio sat around these levels was in mid-2020, when it peaked near 23.6 times as markets bounced back from the pandemic shock. Over the next five years, the index roughly doubled, not because valuations expanded forever, but because earnings also grew strongly.

Over the past year, the S&P 500’s performance has pulled away from its underlying earnings growth. That gap is just another way of describing multiple expansion, where prices rise faster than profits. In past cycles, stretches like this have often been followed by corrections or longer periods of flat returns as earnings catch up or valuations cool.

That is a good way to frame today. The market again trades at rich multiples, especially in AI and big tech. The long-term outcome will depend less on today’s exact P/E and more on whether earnings growth eventually “catches up” with the price investors are paying now. Valuation does not tell you what happens tomorrow, but it shapes how much room for error is left in the story.

Scenario analysis: base, bear and bull paths

From today’s starting point, it helps to think in scenarios rather than single “price targets”. All of them start from the same fact: markets, especially AI leaders, already price in a lot of good news.

Base case: growth continues, valuations cool In a reasonable base case, AI and big tech earnings keep growing at a solid pace as spending on data centres, chips and software stays high. Valuations do not stay at record levels, but they do not collapse either. Multiples drift down or move sideways while profits grow into them. Returns over the next five to ten years are positive but more modest, with pullbacks like the latest sell-off. This path rewards staying invested, but makes stock selection, entry price and time horizon more important.

Bear case: de-rating from a high starting point In the bear case, growth disappoints or interest rates stay higher for longer. AI projects take longer to pay off, customers become more cautious or margins feel pressure from competition and regulation. The market no longer wants to pay 20-plus times earnings for many winners, so rich multiples “de-rate” towards historical averages. Index returns can be weak even without an earnings collapse. The risk is not that AI disappears, but that investors paid too much, too early.

Bull case: earnings grow into the hype In the bull case, AI profits and productivity gains are stronger than expected, spreading across sectors. Earnings growth proves strong enough to “earn” today’s valuations, and the current wobble becomes just another shake-out in a longer structural uptrend.

What long-term investors should focus on now

If your horizon is 5 to 10 years, the key question is not why a stock moved 3% in an afternoon. It is whether the underlying business can keep growing earnings, defending its competitive “moat” and managing debt through different economic conditions. Prices will be noisy around that path, especially in hot themes.

You also do not need to bet the farm on a single scenario. Instead, you can ask a simple question: “If valuations stay high but drift down slowly, if they correct more sharply, or if earnings powerfully catch up, would my current portfolio still let me reach my goals?” That shifts the focus from predicting the next headline to checking whether your holdings can live with different futures.

Practical ways to limit risk and downside

The most powerful tools are simple and process-based rather than predictive. No one can forecast every drop, but you can decide how much damage a drop can do.

Start with position sizing. If a single stock falling 30% would derail your plan, the position is probably too large.

Then look at diversification. Mix sectors, regions and themes so that not everything depends on US big tech or one hot story.

Staggered buying, or drip-feeding, spreads entry points over time and reduces the regret of investing everything at a short-term peak.

Simple rebalancing rules help too, such as trimming a stock or sector once it climbs above a set share of your portfolio.

Finally, a small safety buffer in cash or short-duration bonds can cover near-term needs and stop you becoming a forced seller on a bad day.

The real lesson behind the volatility

For long-term investors, the most useful response is not to guess the next headline, but to use episodes like this as a health check on portfolio design and personal risk tolerance. If the moves felt painful, the solution is usually in position sizes, diversification, buffers and a clear view of scenarios, not in abandoning long-term themes altogether.

In the end, when AI meets gravity, the investors who cope best are not the ones who call every drop, but those whose portfolios are built to keep compounding through both the surges and the setbacks, whatever path valuations take from here.

  This material is marketing content and should not be regarded as investment advice. Trading financial instruments carries risks and historic performance is not a guarantee of future results. The instrument(s) referenced in this content may be issued by a partner, from whom Saxo receives promotional fees, payment or retrocessions. While Saxo may receive compensation from these partnerships, all content is created with the aim of providing clients with valuable information and options.