
Nvidia, Alphabet and Dutch chipmakers endured a weak October. Demand for AI continues to grow across businesses, but investors are increasingly concerned about the billions needed to fund that growth. The differences between chipmakers, cloud companies and software providers are becoming harder to ignore.
The US technology index Nasdaq Composite closed on Friday, October 30, down 2.86 percent. The S&P 500 fell 1.82 percent. Reuters linked the sell-off to disappointing company updates and concerns over how long it will take to recoup massive AI investments.
The selling hit several links in the AI supply chain. Nvidia and AMD make computing chips, Dutch equipment makers help produce them, and cloud companies build the data centers that run them. They earn money at different stages of the same investment boom—and face different risks along the way.
For people using AI, the figures reveal just how much money is behind a chatbot, coding assistant or business AI agent. For investors, the question also comes down to how much they are willing to pay for future profits.
AI stocks slide, but the damage varies widely
October’s sell-off hit both household-name tech companies and specialist suppliers. AI computing providers CoreWeave and Nebius suffered some of the steepest losses. Dutch chip equipment makers ASML, ASM International and Besi also ended the month sharply lower.
Microsoft and German software company SAP, by contrast, gained ground. The market is clearly not treating every company with an AI story the same way.
| Stock | Price change in October 2026 |
| Microsoft | +1.76% |
| SAP, Xetra | +13.78% |
| Nvidia | −10.39% |
| Alphabet | −18.67% |
| AMD | −17.48% |
| ASML, Amsterdam | −11.52% |
| ASM International | −17.82% |
| Besi | −19.71% |
| CoreWeave | −37.02% |
| Nebius | −34.65% |
The figures compare regular closing prices on September 30 and October 30. US stocks are measured in dollars, while the European listings shown are in euros. Dividends and currency conversions are excluded, so returns for Dutch investors holding US stocks may differ as the dollar moves.
Monthly figures tell a different story from a single trading day or a move after the market closes. A stock can rise several percent after strong results and still finish the month lower. AI Wereld previously reported how Microsoft and Alphabet moved in opposite directions after their quarterly results.
The reasons for a stock’s decline also vary. Disappointing revenue, a weaker profit outlook, expensive financing or a shift in market sentiment can all play a part. A falling share price alone does not reveal what is going wrong at a company.
Microsoft, Google and Amazon are growing—but so is the bill
For now, the major cloud providers are still posting strong growth. Microsoft reported quarterly revenue of $92.3 billion for the three months through September, up 19 percent year over year. Azure, the cloud platform businesses use to run AI applications and other services, grew 48 percent.
Microsoft generated $58 billion in operating cash flow and spent $39.4 billion on property and equipment, including buildings and hardware. According to Microsoft’s financial disclosures, that left about $18.6 billion in free cash flow.
Free cash flow shows how much money remains after a company’s operations and relevant investments. It helps indicate whether a business can fund its expansion from its own resources. It is calculated differently from net income.
The picture is different at Google parent Alphabet. Google Cloud revenue surged 103 percent in the third quarter to $30.7 billion. Yet Alphabet’s free cash flow for the quarter was negative.
Alphabet generated $42.3 billion in operating cash flow and invested $51.4 billion in property and equipment. That left free cash flow at minus $9.1 billion, according to Alphabet’s quarterly results.
A company can be profitable while investing more than it brings in through operating cash flow. The cost of a data center or server is often spread over several years in profit calculations, even though the cash may be paid out much earlier.
Amazon is also showing strong cloud growth. Amazon Web Services, usually known as AWS, generated $46 billion in quarterly revenue, up 39 percent. AWS posted operating income of $17.8 billion.
Over the past 12 months, free cash flow across the Amazon group stood at minus $42.4 billion. That is a full-year figure, while the cash flow figures cited for Microsoft and Alphabet cover just one quarter. Amazon points to its expanding AI infrastructure as a major reason for higher investment.
Negative free cash flow does not, by itself, prove that an investment has failed. New capacity can generate revenue later. The key questions are how many customers will pay for it, how fully the equipment will be used, and whether the returns will outweigh construction, electricity, maintenance and financing costs.
Cloud revenue also extends well beyond AI. Azure, Google Cloud and AWS host databases, websites and other business applications. Presenting all cloud growth as AI revenue would overstate AI’s share.
ASML, ASM and Besi profit from different stages of the chip supply chain
For the Netherlands, the AI investment boom has a clear industrial dimension. ASML, ASM International and Besi supply equipment for different stages of chip production. Their customers can use these machines to make advanced AI processors, but also other types of semiconductors.
ASML makes lithography machines that chipmakers use to print patterns onto silicon wafers. The company reported revenue of €11.66 billion and net profit of €4.96 billion in the third quarter, according to ASML’s quarterly results.
One detail is important when looking at that profit. Chief Financial Officer Roger Dassen said about €500 million came from a revaluation of ASML’s stake in Mistral AI. That change in value should be distinguished from the profit ASML earns from machines, maintenance and upgrades.
ASM International supplies, among other things, machines that deposit ultra-thin layers of material. These are essential for building complex chip structures with precision. According to ASM’s quarterly report, the company generated €1.16 billion in revenue in the third quarter, up 45 percent from a year earlier.
At constant exchange rates, the company expects revenue to exceed €5.7 billion in 2027. AI Wereld previously explained how ASM is bringing forward its revenue target. But it remains a forecast: the machines still need to be delivered, and the associated revenue has yet to materialize.
Besi focuses on chip assembly and interconnection technologies, enabling manufacturers to combine different chip components with precision. These techniques are becoming more important as designs incorporate multiple components placed side by side or stacked on top of one another.
According to Besi’s results, quarterly revenue rose 113.6 percent to €283.4 million. Orders totaled €342.6 million. Some deliveries are tied to production capacity that customers expect to have available in the first half of 2027. An order received, then, does not automatically count as revenue for this quarter.
Further down the supply chain, the business models diverge too. Nvidia sells a combination of computing chips, systems, networking and software. For the second quarter of fiscal 2027, which ended in July 2026, the company reported revenue of $96.2 billion. Of that, $89 billion came from Data Center. The segment is closely tied to AI infrastructure, but it is not a standardized measure of AI revenue alone.
Broadcom does report a separate figure for AI semiconductors. In the third quarter of its fiscal year, revenue from those chips reached $16.7 billion, against total revenue of $29.6 billion. AI semiconductors therefore generated more than half of the company’s revenue, according to Broadcom’s financial release.
The comparison shows why a single ranking of AI companies can quickly mislead. An equipment maker sells machines to a chip factory, a chip designer sells computing platforms, and a cloud provider rents out capacity. Their revenue arrives at different points and each business requires a different level of capital.
Debt and valuation determine how much growth a stock needs
Specialized AI cloud providers show just how much financing can weigh on a business. CoreWeave generated $2.58 billion in revenue in the second quarter, up 112 percent from a year earlier. But it also posted a net loss of $626 million.
Net interest expense came to $640 million, according to CoreWeave’s quarterly results. Those financing costs show why revenue growth alone is not enough to assess the business model.
Customer concentration adds another risk. The company’s three largest customers together accounted for 72 percent of quarterly revenue, based on CoreWeave’s quarterly filing. A major customer changing its plans could therefore have a significant impact.
Amsterdam-headquartered Nebius reported second-quarter 2026 revenue of $582 million. The company generated $2.25 billion in operating cash flow, against $5.66 billion in purchases of property, plant, equipment and intangible assets. Nebius’s financial results also show that customer prepayments are helping support cash flow.
Prepayments can help fund construction, but the company still has to deliver services in return. To understand its financial position, investors also need to consider the construction schedule, contract terms and future capital requirements.
The company’s performance is only part of the story. Its market valuation matters, too. A business can increase sales and profits while its stock falls if investors had already priced in even stronger growth.
The price-to-earnings ratio helps illustrate this gap. It compares a company’s share price with its earnings per share. Based on reported earnings over the past 12 months, Nvidia closed at around 26 times earnings on October 30. Palantir stood at around 157 times.
Those figures are not buy recommendations. Growth prospects, risks and earnings quality vary. But they do show why the label “AI stock” says little about how much future growth may already be priced in.
The same applies to the Dutch chip sector. ASML traded on the Amsterdam exchange at around 40 times its earnings over the past 12 months. When assessing those earnings, investors must account for both machine sales and the previously mentioned revaluation of its stake in Mistral.
A simplified example shows how a changing valuation can affect a stock. If a hypothetical company’s earnings per share rise by 20 percent, but its price-to-earnings ratio falls from 40 to 30, the implied share price still drops by 10 percent. Profit growth and stock gains do not automatically move in lockstep.
Upcoming quarterly results will provide more clues. AMD reports on November 3, followed by Nvidia on November 17. At both companies, revenue, margins and guidance could reveal how much of the announced AI expansion is now showing up in their results.


