Swiss Stocks Turn Into a Haven as AI Fears Roil Global Markets

NextFin News - Switzerland's SMI index closed at 14,547.97 points on October 23, less than 1% below its 52-week high and within a fraction of the record set in August, as investors piled into the world's most defensive major equity benchmark while artificial-intelligence-linked growth stocks unraveled. The divergence is the clearest signal yet that the market has moved past asking whether the AI trade is crowded and is now deciding where to hide when it unwinds.

The Setup: A Growth Rout With a Defensive Bid

The contrast is stark enough to read as two different markets. In the week ended October 23, the S&P 500 lost roughly 2% to finish near 7,659, slipping back below the August high it had only just reclaimed, while the Nasdaq Composite surrendered about 4.5% and reversed its breakout above the summer peaks. The breadth of the reversal was concentrated rather than universal: eighteen of the S&P 500's twenty biggest weekly decliners had been among its hundred strongest performers between March 30 and October 6, according to market-data calculations. The hundred strongest stocks from that window lost roughly 5% on average for the week, while the hundred weakest gained roughly 4%.

Yet buyers had not abandoned the market outright. Roughly 72% of S&P 500 constituents still finished the week higher, and the equal-weighted version of the index eked out a gain. That is the fingerprint of a leadership rotation, not a broad risk-off collapse. Dell and Hewlett Packard Enterprise, both of which had more than tripled during the earlier advance on AI-infrastructure demand, each lost more than 20% in the week. Palantir, which had briefly reclaimed a record in November, confirmed a correction and ended the week down roughly 11%.

The catalyst stack behind the move is specific rather than vague. Expectations for AI capital spending have been pushed out and questioned: highly anticipated IPOs from AI startups including OpenAI and Anthropic have been delayed, forcing investors to reassess how much of the sector's valuation rests on future fundraising rather than current earnings. At the same time, the 10-year Treasury yield has climbed to its highest levels since 2002, raising the discount rate applied to long-duration growth cash flows — the exact mechanism through which a bond-market move becomes an equity-market repricing. Worries about AI safety and development have also entered the conversation, adding a regulatory overhang to the valuation one.

Against that backdrop, the Swiss benchmark barely blinked. The SMI's close of 14,547.97 sat inside a daily range of 14,408.96 to 14,552.72, with a 52-week span of 12,053.51 to 14,669.51. The index's all-time high of 14,672.80, recorded in August 2026, is essentially within reach. Year to date, the SMI is up about 8.4%, a gain that has accelerated precisely as the AI trade has wobbled.

The reason is mechanical before it is philosophical. The SMI is built like a fortress: its largest components are Roche Holding, Novartis and Nestle, followed by ABB, UBS and Zurich Insurance. Pharmaceuticals and food alone account for more than half the index, and equity strategists have long characterized the benchmark as leaning roughly 80% defensive against 20% cyclical. The three dominant names derive their revenues from patients, shoppers and insurers — not from data-center capital expenditure or graphics-chip orders. When the market reprices the future cash flows of artificial intelligence, it is repricing a slice of the global economy that Switzerland barely owns.

That is not an accident of this cycle. It is the legacy of what Switzerland lists. The country's public market is the equity arm of an economy built on drugs, food, insurance and precision engineering — industries with long product cycles, regulated pricing and demand that moves with demographics rather than with the Nasdaq. The SMI's twenty constituents represent about 85% of the Swiss equity market, so there is no broad small-cap universe to drag the index toward growth. Switzerland simply does not have a listed technology champion large enough to move the index, and that absence is the point.

Why the Haven Function Works — and Where It Can Break

The first-order mechanism is index construction, and it is unforgiving in both directions. An equity index is a portfolio, and portfolios have exposures. The SMI's heavy weighting in health care and consumer staples, and its thin technology sleeve — effectively limited to ABB's industrial automation and a handful of smaller names — makes the benchmark a low-beta position against a technology-led melt-up and a low-damage holder during a technology-led drawdown. There is no Swiss Nvidia, no Swiss Meta Platforms, no Swiss Palantir. When the megacap technology names drag the U.S. market lower, the Swiss index has almost no direct equity to mark down.

This structure also explains why the SMI can look boring and still compound. The defensive sectors carry their own return engine: pricing power in branded staples, patent-protected margins in pharmaceuticals, and fee income in insurance and asset management. Nestle's forward dividend yield sits near 3.7%, and Roche's near 2.5%, levels that matter when growth-stock dividends are measured in basis points. In a market suddenly concerned with cash flows you can hold rather than narratives you can only believe, that income stream is not a footnote — it is the trade.

The second-order channel is the one most investors miss: the Swiss franc. The same risk-off impulse that bids the SMI also bids the currency, and the two reinforce each other. Currency flows confirmed the pattern during the selloff — the dollar was far from the preferred safe haven, with the yen and the Swiss franc emerging as the main beneficiaries of the flight to safety. The franc's haven status is older and deeper than any single equity cycle: Switzerland's current-account surplus, low public debt and political neutrality have made the currency a store of value through wars, financial crises and inflation episodes.

A stronger franc would normally wound a Swiss exporter. But the SMI's giants earn the bulk of their revenue abroad, in dollars and euros, and their products are price-inelastic. A franc that strengthens on risk aversion compresses reported Swiss-franc earnings for Nestle and Roche, but it does not change the volume of KitKats sold or the number of cancer patients treated. Meanwhile, the discount rate applied to those stable cash flows falls as investors flee to safety. The equity effect dominates the currency effect. That is why the index can rise even as its home currency strengthens — a nuance that separates a currency hedge from a genuine earnings haven.

Here is the call that determines whether this story has legs: the AI repricing is cyclical; Switzerland's haven status is structural. The evidence for the cyclical read is straightforward. The selloff is a repricing of expectations around AI capital spending and the valuations attached to it — a sentiment and positioning shock, not a collapse in global output. Three features mark it as cyclical: it is concentrated in the prior leadership cohort, it is tied to a specific catalyst calendar including Nvidia earnings and delayed IPOs of AI startups, and it has not broadened into the majority of the S&P 500. Cyclical shocks mean-revert. When the AI narrative stabilizes, money rotates back into growth.

The evidence for the structural read is equally clear. The SMI's defensive composition is embedded in the index's construction and in Switzerland's economic structure. Equity strategists noted the dynamic when comparing the Swiss and Swedish benchmarks:

In a more troubled world, where growth slows due to some combination of today's worries, earnings could be less at risk as this index shines as a safe haven.

The practical implication cuts both ways. The haven bid will fade when the AI trade recovers — the cyclical leg. But the SMI will remain structurally underexposed to the next technology cycle — the structural leg. Investors treating Switzerland as permanent shelter will be right about the next correction and wrong about the next rally. That asymmetry is the whole story in one line: Switzerland wins the week you are scared and loses the year you are greedy.

The Counter-Thesis: Haven, or Expensive Trap?

The strongest argument against the haven narrative is that the SMI is not defensive — it is narrow, expensive, and one clinical-trial failure away from trouble. The concentration risk is real and measurable. Three stocks dominate the index, and the top ten names account for the overwhelming majority of its market capitalization. Nestle has been a laggard, down roughly 16% over the past year in some measures, and its weight drags the index even in a risk-off week. Roche and Novartis face patent cliffs, pricing pressure in the U.S. health-care system, and the binary risk of clinical readouts. A single failed phase III study in a blockbuster pipeline can move the entire index.

The valuation argument is equally uncomfortable. The Swiss and Swedish benchmarks have both traded near 17 to 18 times forward earnings, a premium of about 20% to the broader European index — a level the SMI has touched only three times since 2011, according to strategist data. When an index is priced for perfection, "defensive" can quickly become "expensive and fragile." If earnings disappoint, a defensive index does not receive the multiple expansion that usually cushions a growth index; it absorbs both lower earnings and a de-rating.

This counter-thesis is not easily dismissed. It attacks the core claim — that the SMI is a safe place to be — at its foundation, and it rests on valuation and concentration metrics rather than sentiment. The answer is that both statements are true, on different time horizons. Concentration and valuation are medium-term fundamental risks; the haven function is a short-term flow phenomenon. In a two-week risk-off episode, flows dominate fundamentals, and the SMI's low AI beta wins. Over two years, earnings growth and the multiple matter more, and the index's narrowness is a headwind.

The falsifying signal for the haven thesis is specific and observable: if the SMI falls more than the Nasdaq over a week in which the Nasdaq drops 3% or more on AI-driven selling, the defensive-beta argument is broken. Through October 23, the data said the opposite. The SMI's relative strength was not a rumor or a sentiment read — it was a price fact, recorded in the gap between a near-record Swiss close and a Nasdaq that had given back its breakout.

What to Watch: Beneficiaries, the Exposed, and the Signals

The beneficiaries are clear. Swiss health-care and consumer-staples names gain when the discount rate on stable cash flows falls and when growth-stock money looks for cover. Insurers such as Zurich Insurance and Swiss Life, and industrial names with contracted or regulated revenue such as ABB and Geberit, sit in the same defensive bucket. Lonza and Galderma add biotech and dermatology exposure with growth profiles that still carry less AI sensitivity than the U.S. technology complex. The exposed are the global technology leaders whose valuations embed years of AI-driven growth, and any investor whose benchmark is the S&P 500 or Nasdaq and who is now underweight defensives.

Three horizons frame the path ahead. In the short term, watch the Nasdaq's direction and the VIX. If the AI selloff deepens, the SMI's relative outperformance should extend, and a sharply stronger franc against the dollar would confirm the haven flow is broadening beyond equities. In the medium term, the quarterly results from Roche, Novartis and Nestle matter most. If the drugmakers guide lower on U.S. pricing, or if Nestle's volumes continue to disappoint, the valuation premium to the broader European index becomes the story instead of the defensive tilt. In the long term, the question is whether Switzerland lists more technology and growth companies. The haven status persists only as long as the index remains structurally underexposed to AI; a wave of Swiss technology IPOs would, over time, erode the very feature that makes the SMI attractive today.

The base case is that the AI trade stays volatile through year-end and the SMI holds near-record levels while U.S. growth stocks churn. The upside case is that a deeper AI repricing pushes the SMI to a fresh all-time high as global money seeks low-beta exposure. The downside case is that the AI narrative stabilizes, growth rotates back, and the SMI's valuation premium compresses toward the European benchmark.

Switzerland's stock market is not a bet on the future. It is a bet that the future, for once, can wait — and in a market obsessed with what artificial intelligence will do next year, that is exactly the trade nobody wants to own until they need it.

Explore more exclusive insights at nextfin.ai.