US tech stocks see sharp falls as AI spending nears $1tn

Major US technology stocks suffered a broad sell-off for the second consecutive day. The Nasdaq, which tracks many large listed technology companies, fell more than 3.5% on Friday. Investors were unsettled because companies are spending heavily on AI while some results and forecasts disappointed.

Intel recorded the sharpest fall among the companies highlighted, with shares dropping as much as 30%. Its planned AI spending came alongside weaker computer demand, lower-than-expected revenue guidance, and an expectation that it would not generate more cash than it spent this year. Reddit fell 26% after reporting weaker daily user numbers and declining referral traffic.

Amazon shares fell as much as 6% overnight before recovering to about 1.6% lower. Roblox dropped nearly 20% after daily active users declined. Alphabet had already fallen more than 12% after warning that AI search summaries were affecting advertising revenue, and its stock continued to slide.

Stock-market value, also called market capitalisation, is the combined value of a company’s shares. It is calculated by multiplying the share price by the number of shares available. It reflects what investors collectively think the company is worth at that moment, rather than the cash sitting in its bank account.

If investors suddenly expect weaker growth, lower profits, or greater costs, many may sell. The share price then falls. Because the new price applies to every share, even a modest percentage decline can erase an enormous amount of quoted value for a very large company. The value disappears from market pricing, not necessarily from the company’s bank balance.

Alphabet’s stock-market value plunged by more than $500bn in one day after it linked AI search summaries to weaker core search advertising revenue. Intel was on track to lose more than $130bn in value as its shares continued falling. These figures show how quickly expectations can change.

Major technology companies are committing extraordinary sums to artificial intelligence. Meta, Alphabet, Oracle, Amazon, and Microsoft together plan to spend $800.6bn in 2026. Alphabet separately expects to invest up to $205bn this year, almost entirely on AI-related projects. Amazon has committed to spending $220bn this year.

This money pays for the physical and technical foundation of AI. Companies are building data centres and buying specialised chips, servers, and network equipment. They also spend on software development and the systems needed to train AI models and deliver answers to users. These projects require large upfront investments before their financial returns are certain.

The spending is already affecting company finances. Amazon said it spent more cash than it made, with spending rising six-fold over three months. Intel also increased equipment investment despite weaker computer demand. Investors are therefore watching whether this huge build-out eventually produces enough revenue and profit.

Building AI systems requires enormous spending before companies know how much money those systems will make. Firms must finance data centres, specialised equipment, servers, networking, and software. These costs can weigh on cash flow and profits, especially when companies are still expanding capacity rather than selling enough AI services to cover it.

The central concern is a mismatch between investment and returns. A company may spend billions training models and operating infrastructure, but customers or advertisers may not generate enough additional revenue. If earnings miss expectations, investors may decide that future profits will be smaller than previously thought. Share prices can then fall, even if the business continues growing.

The article describes this as a possible AI bubble. Amazon spent more cash than it made, while Intel expected no positive cash surplus this year. Analysts said major firms were showing growing pains. Investors have become less patient and are punishing disappointing earnings, though many technology shares remain above last year’s levels.

Google earns important advertising revenue from its traditional search results. Advertisers pay for visibility when people use search and interact with result pages. AI-generated summaries change that experience by presenting a direct answer as the top response, rather than primarily directing users toward a list of links.

If the summary satisfies a user’s question, that person may have less reason to scroll through traditional results, click links, or interact with advertisements displayed around them. That can reduce opportunities for advertising revenue. The exact financial effect depends on how users behave and how Google places ads alongside AI answers, but the commercial risk is clear.

Alphabet told investors that its AI summaries were affecting core search advertising revenue. Its market value then plunged by more than $500bn, and the stock fell over 12% on Thursday. Alphabet still expects to spend up to $205bn this year, almost entirely on AI, making the revenue question especially important.

AI systems perform demanding computing tasks. They process large amounts of information, train software to recognise patterns, and generate responses when users ask questions. That work requires far more computing capacity than ordinary office applications. Data centres bring together the electricity, cooling, security, and hardware needed to operate these systems at scale.

Specialised chips carry out the calculations efficiently. Servers hold and process data, while networking equipment moves information between machines and connects the centres to users. If any part is too small or slow, training can take longer and responses can become less reliable or more expensive to deliver. Companies therefore invest in the whole infrastructure chain, not just AI software.

The article identifies AI data centres, network equipment, and servers as major uses of Amazon’s planned spending. It also says Intel increased equipment investment while computer demand weakened. These costs help explain why AI expansion can reduce cash flow before it produces matching revenue.

Investors buy shares partly for the profits they expect a company to generate in the future. Strong earnings and convincing forecasts can make those shares seem more valuable, encouraging investors to pay higher prices. Weak earnings, rising costs, or lower guidance can have the opposite effect, because expected future returns appear less attractive.

This process is driven by changing expectations. Amazon’s earnings fell short of investor expectations, and the company said it was spending more cash than it made. Intel forecast lower-than-expected fourth-quarter revenue and said it would not bring in more cash than it spent this year. Investors responded by selling shares, pushing prices down.

The same pressure affected Alphabet, whose AI summaries were linked to weaker core search advertising revenue. Analysts are now asking whether huge AI investments will produce enough profit. Investors have become ruthless toward disappointing third-quarter earnings, although the Nasdaq and several companies remain above their levels from a year earlier.