Over the past few days, I kept seeing large moves in companies that seemed, at first, to be part of the same story. Alphabet fell sharply, then Intel collapsed after its results. Nvidia and Amazon declined too. Yet energy and some European healthcare shares rose. I wanted to put the moves next to one another and understand whether the market had discovered a common problem for AI and technology, or whether it was reacting to quite different problems at individual companies.

The answer is a bit of both. Investors are questioning the scale of AI spending and how long it will take to earn an adequate return. But Intel’s PC outlook and manufacturing costs are not the same as Alphabet’s cloud expansion or Nvidia’s position as a supplier. A broad index can magnify the shared concern even while the individual explanations remain different.

The week in prices

The table uses closing share or index price changes against the previous local trading day. Unless a row says Thursday, the move is for Friday 30 October 2026. This gives us a snapshot of the reaction, rather than a measure of each company’s long-term performance. Percentages are rounded; they do not include dividends or the currency conversion a euro investor would experience. Sector key: blue for technology, grey for indexes, green for healthcare, amber for energy and consumer, and violet for media.

United States

Europe

Asia confirms that this was not confined to Wall Street. On Friday, South Korea’s KOSPI and Taiwan’s benchmark were down about 4.5% and 2.5% respectively in a Reuters early-session snapshot. Samsung Electronics and SK Hynix also fell. Those Asian figures were reported before the final close, so I have kept them out of the closing-price tables rather than mixing different measurement points. [1–3]

One theme, several different businesses

Amazon is a useful place to start because its quarter was hardly a simple story of weak demand. AWS sales rose 39% year on year to $46.0 billion, and Amazon reported $28.1 billion of operating income. At the same time, its trailing-twelve-month free cash flow moved from a $14.8 billion inflow a year earlier to a $42.4 billion outflow. The company attributed much of the change to a sharp increase in purchases of property and equipment, primarily for AI, and said it expected roughly $220 billion of cash capital expenditure in 2026. The market could believe in AWS growth and still ask when that enormous investment will yield cash for shareholders. [4]

Free cash flow is not a perfect measure of a project’s value, particularly during a build-out. A data centre built today may generate revenue for years. But the cash leaves before all of those future returns arrive, and the assumptions about utilisation, pricing and useful life have to prove themselves. If a company keeps building because its customers demand capacity, investors must still decide whether the return on each new investment is attractive.

Alphabet illustrates the same broad question with a different starting point. It reported revenue growth of 25% and Google Cloud revenue growth of 103% in the third quarter. Its operating margin was 31.8%. Those figures make it hard to describe the business as failing. The sharp Thursday share-price decline instead reflected investors weighing cash spending, future returns and the expectations that had already been priced in. It would be misleading to use Friday alone as the date of Alphabet’s main reaction. [5, 6]

Intel’s fall requires a more company-specific explanation. Revenue grew in the reported quarter, but its fourth-quarter forecast disappointed and management did not expect positive cash flow this year. Weak PC demand, manufacturing execution and the cost of expanding production all matter to this case. It was not simply a vote on whether AI will succeed. The 28.6% fall is therefore a useful warning against putting every semiconductor company in a single box. [7, 8]

Nvidia, AMD and European chip suppliers nevertheless fell with the sector. Their shares are exposed to expectations about how much AI infrastructure customers will buy and when. If investors lower their estimate of future spending, or demand more evidence that buyers can earn returns, suppliers can reprice even without a new company-specific earnings warning that day. That is a market transmission channel, not proof that every affected supplier faces the same operating problem. [1, 2]

What is shared across AI and what is company specific

The shared issue is the distance between spending and payoff. Cloud providers and technology platforms buy equipment and construct capacity now. Chip and equipment suppliers may record revenue earlier in that cycle. End customers must eventually use AI products enough, and pay enough, to support the bill. If capacity expands faster than profitable use, the value of some projects will disappoint. Financing costs can also matter more as the amounts involved grow.

Competition complicates the calculation. A provider may have to invest heavily merely to maintain its position, while customers gain the benefit through cheaper or better services. In that case, AI can be economically valuable without each investor in the supply chain earning an exceptional return. Conversely, strong and lasting demand could justify today’s spending if utilisation and margins improve over time. The weekly share moves cannot settle which outcome will occur.

Company-specific evidence still has to be read on its own terms: Intel’s PC and foundry challenges, Amazon’s AWS growth and spending, Alphabet’s mix of advertising and cloud, and Nvidia’s exposure to data-centre demand are not interchangeable. Nor did every sector decline. In Europe, Sanofi and Universal Music rose on their own results. US energy shares gained. That unevenness is why I find the week more informative than a generic “market panic” headline. [1, 2]

The link with the earlier papers

In Macro to Portfolio No. 06, “Markets Look Resilient. So Why Are the Risks Getting Harder to See?”, I discussed the possibility that a handful of very large companies could make a widely diversified equity fund sensitive to one set of AI expectations. This week offers a concrete example. Owning hundreds of stocks spreads company risk, but it does not guarantee that the largest holdings will respond to different economic assumptions. The earlier paper also distinguished a trigger from an amplifier: an earnings disappointment can be the trigger, while crowded exposures or borrowing can intensify a move. I do not have evidence that leverage caused this week’s sell-off, so I would keep that as a risk to monitor rather than a diagnosis. [9]

The currency point from No. 04 matters too. A euro-denominated trading line for a global ETF does not remove the US equity exposure underneath it. The dollar’s movement can change the euro return received by an investor, separately from the local share-price changes shown in the table. I have deliberately avoided turning the Friday percentage moves into supposed euro-portfolio returns without specifying an actual fund, its holdings and the relevant exchange rates. [10]

From market moves to the four portfolios

Capital Preservation

I would check whether the equity sleeve can experience an AI-led drawdown without interfering with the portfolio’s short-term purpose. Liquidity and high-quality bonds still have distinct jobs. A fall in growth shares does not, on its own, tell us to abandon equities, and a bond ETF is not identical to cash. The earlier euro yield-curve paper explains why maturity and duration matter in that defensive allocation. [11]

Early-Career Growth

Regular contributions and a long horizon can make a decline easier to absorb. But a global ETF already owns large US technology companies. A separate Nasdaq or AI fund would be an intentional concentration, not a new source of broad diversification. I would measure the combined exposure before deciding whether a thematic position belongs beside the core.

Balanced Entrepreneur

The venture reserve for a potential business need within three years should not be exposed to a technology sell-off. For the long-term portfolio, I would look at the entrepreneur’s existing business exposure as well as the securities account. A business dependent on technology spending may already share some of the same economic risk as AI-heavy equities.

Multi-Generational Family Stewardship

The family can invest over a longer horizon, but should map liquid assets against spending and capital calls. If private investments and public shares both depend on sustained data-centre spending, separate labels will not make them independent risks. As No. 05 and No. 06 discussed, private valuations can update more slowly than quoted shares; the apparent calm of an illiquid holding does not establish that it has escaped the same economic shock. [9, 12]

My current view

I would not rebuild any of the four portfolios around one week’s prices. I would use the week to check what we already own, especially overlapping technology exposure across global funds, thematic funds and individual shares. If an active AI position is proposed, its case should explain how the company earns a return on the money it spends, what must go right and what valuation already assumes.

The distinction that stays with me is simple: growth in AI usage, growth in a company’s revenue and a good return for its shareholders are three separate steps. This week’s falls suggest that investors are paying closer attention to the links between them. The gains elsewhere in the market remind me to examine each company’s evidence before calling the whole episode systemic.

What I will monitor next

In the next results I will compare capital expenditure with incremental revenue, operating margins and free cash flow; read management’s expected timetable for returns; and watch whether cloud capacity is being used profitably. For suppliers, I will look at orders and customer concentration rather than infer demand solely from a platform’s spending announcement. Across markets, I will watch whether pressure remains focused on AI-linked names or spreads to credit, other sectors and funding conditions. That would change the assessment of how broad the shock has become.

Conclusion

The late-October sell-off exposed a shared question about the economics of AI investment, but its sharpest moves had different immediate causes. Intel’s outlook was not Amazon’s cash-flow story; Alphabet’s Thursday reaction was not Nvidia’s Friday move; and Europe’s technology decline did not prevent gains in healthcare and other companies. For the portfolios, the practical response is to understand concentrations, preserve money needed soon and require a clear return case for every extra technology bet.

[1] Reuters, “Wall Street closes sharply lower as AI cash burn fears spark tech selloff,” 30 October 2026. Source

[2] Reuters, “European shares stumble at end of best month since June on AI worries,” 30 October 2026. Source

[3] Reuters, “Tech rout hits S. Korea, Taiwan,” 30 October 2026, early-session snapshot. Source

[4] Amazon, third-quarter 2026 results, 29 October 2026. Source

[5] Alphabet, third-quarter 2026 results, 28 October 2026. Source

[6] Reuters, “Stocks fall on AI cash burn worries,” 30 October 2026. Source

[7] Intel, third-quarter 2026 results, 29 October 2026. Source

[8] Reuters, “Intel on course for worst day ever, triggers global chip rout,” 30 October 2026. Source

[9] Macro to Portfolio No. 06, “Markets Look Resilient,” 17 October 2026. Source

[10] Macro to Portfolio No. 04, “The ETF Is Priced in Euros,” 9 September 2026. Source

[11] Macro to Portfolio, “Cash Pays Again, but Long Bonds Still Hurt,” 3 September 2026. Source

[12] Macro to Portfolio No. 05, “Private Credit Is Everywhere,” 17 October 2026. Source

Educational project based on public information available on 1 November 2026. This article is not personalised investment, legal or tax advice, nor a recommendation to buy or sell a security. The four client profiles are fictional; no real client money is involved.