The AI trade's trick-or-treat earnings season

The AI trade is the investment bet that artificial intelligence will create major growth for technology companies. It includes spending on chips, cloud computing, data centers, and AI services. It matters because AI can expand businesses while also disrupting their most profitable products.

The latest hyperscaler reports showed both sides. Google’s AI Overviews increased search activity, but AI-mode results have lower margins than traditional search advertising. Meanwhile, Google Cloud benefited from demand for its TPU chips. Microsoft and Amazon reported strong cloud growth, but private investments in Axiom and Paradox also affected their profits.

That combination unsettled investors. The Nasdaq briefly entered correction territory, while Google lost more than $500 billion in market value after earnings. Companies continue spending heavily because demand is said to exceed supply. But if demand weakens, the resulting excess infrastructure could expose the AI trade as an overheated bet.

The market reaction was large enough to signal broad anxiety about the AI boom. The tech-heavy Nasdaq briefly entered correction territory, meaning it fell 10% from its prior peak. That decline reflected investor concern after several hyperscalers released earnings that highlighted risks alongside growth.

Across October, the technology sector lost $1.2 trillion in value. Google suffered an especially sharp reaction after earnings, losing more than $500 billion in market value. Its report raised concerns that AI Overviews could weaken the economics of its core search advertising business.

The sell-off did not mean investors rejected AI altogether. Some earnings details supported the bullish case, including strong cloud growth and demand for Google’s TPU chips. The market was instead weighing whether future AI revenue can justify enormous infrastructure spending and whether AI might damage existing profit engines.

AI Overviews can make Google’s search product more useful or engaging, encouraging people to conduct more searches. But higher activity does not automatically mean higher profits. The crucial issue is how each search is answered and monetized. Traditional search advertising has been Google’s core business, while AI-mode results have lower margins.

The article describes a trade-off: search queries are rising, yet AI Overviews may cannibalize advertising revenue. AI responses can provide an answer directly, changing the role of conventional search results and the economics around them. Google therefore faces more usage but less profit from some interactions.

This tension helps explain the sharp market reaction. Google lost more than $500 billion in market value after earnings. AI also creates an opportunity through Google Cloud and TPU demand, but the company must move quickly enough to replace or offset pressure on its advertising business.

Cloud-computing businesses are central because AI services require substantial computing infrastructure. Hyperscalers provide access to that infrastructure and can sell related services to businesses. Their cloud revenue therefore offers a direct signal of whether AI demand is becoming a commercial business rather than remaining only an investment theme.

Google Cloud is closing in on Microsoft’s Azure revenue, helped substantially by demand for Google’s TPU chips. Microsoft and Amazon also reported explosive cloud growth, although investors wanted even faster growth from Amazon. These results show how cloud platforms can benefit when customers need more AI capacity.

The cloud opportunity also creates risk. Technology companies are spending heavily to expand supply because they say demand exceeds it. If demand later falls, companies could be left with too much costly infrastructure. Investors are therefore watching both cloud growth and the quality of the demand behind it.

Investments in private AI companies can change reported earnings through accounting gains or losses. Those effects may appear in net income even though they do not come from selling cloud services, advertising, or other core products. Investors must therefore separate investment results from underlying business performance.

Amazon illustrates the distinction. Its operating profits soared, but net income fell because it did not repeat the pretax gain from its Paradox investment that had boosted earnings a year earlier. Microsoft had the opposite kind of effect: its Axiom investment slightly increased net income and lifted profit growth.

These changes can make a company’s earnings look stronger or weaker without showing an equivalent change in its main operations. That is why the article urges investors to check under the hood. The key question is whether growth comes from private AI investments or from the hyperscaler’s underlying business.

Hyperscalers are expanding AI infrastructure because they say demand is outstripping supply. That strategy can work if customers keep buying enough computing capacity and AI services to support the investment. It becomes dangerous when companies build far faster than demand develops.

The mechanism is straightforward. Companies spend heavily on data centers, chips, and related infrastructure to create more supply. If demand later falls, those assets may be underused while the spending has already reduced cash generation. The article warns that pouring trillions into supply before a demand shock could produce a popped-bubble situation.

The risk is not limited to ordinary market volatility. Cisco’s Jeetu Patel identified a major cyber event affecting core AI demand as a possible downside. Despite these concerns, AI capital spending continues. The next test is whether real demand remains strong enough to turn new capacity into lasting revenue.

Free cash flow is the cash a company generates after paying for its operations and capital investments. It shows how much money remains available for flexibility, debt reduction, acquisitions, or returns to investors. When infrastructure spending rises sharply, free cash flow can shrink or turn negative.

The article notes that some asset-light technology companies have become free-cash-flow negative while funding AI ambitions. Management argues this reflects demand exceeding supply, so investing now should create capacity that later produces revenue. The spending includes the infrastructure needed to support AI services.

Negative free cash flow is not automatically a failure. It can reflect deliberate growth investment. But it raises the stakes if demand weakens. Companies may then have committed enormous sums to infrastructure without enough revenue to support it. Investors are watching whether today’s cash outflow becomes tomorrow’s durable AI business.