Tech Stocks Plunge as $725B AI Spending Fails to Deliver Returns

NEW YORK, JOURNALARTA.COM  – AI stocks sold off sharply on Wall Street after investors began pressing a harder question: when will the billions poured into chips, data centers, and power generation turn into real profit? The Nasdaq fell 1.5% and the S&P 500 dropped 0.5% as the market reassessed the scale of 2026 spending, which is now estimated at more than $725 billion.

The pullback hit big technology names and exposed how dependent the recent rally has been on faith in future AI earnings. IBM was among the hardest-hit stocks, with shares plunging 25.2% in a single day after the company said some clients shifted IT budgets toward AI infrastructure and delayed software deals.

Why AI stocks lost momentum

The first trigger is simple. Investors no longer want headlines about ambition. They want proof that the money will come back. Four of the largest U.S. AI players are expected to spend more than $725 billion in 2026 on computing hardware, new facilities, and electricity, a level of capex that dwarfs earlier spending cycles.

That scale has started to unsettle the market. For months, traders treated heavy investment as a sign of strength. Now the same spending is being read as a risk: if revenue growth does not keep pace, margins can shrink fast and valuations can reset even faster.

The second pressure point came from corporate customers. IBM said its business slowed because clients diverted budgets toward AI infrastructure rather than traditional software and services. That shift hurt deal flow and dragged the stock lower in a move that analysts said marked the company’s worst one-day drop since 1968. ServiceNow and Workday also came under pressure as investors worried the budget rotation could spread across enterprise software.

China’s AI push added more pressure

Sentiment weakened further after China’s new open-source model Kimi K3 entered the market. Investors took that as another warning that U.S. AI valuations may be too rich, especially if rivals can offer capable models at lower cost and with faster adoption.

Chipmakers felt the hit. Nvidia slipped 1.3%, while SK Hynix fell 14% and Sandisk dropped 13%. The Semiconductor Index is now about 20% below its June peak, a sign that the correction has moved beyond a single stock or one weak earnings print.

Hedge funds have also been trimming AI exposure over the past five to six weeks, according to market tracking cited in trading desks. That matters because it suggests the unwind may not be a one-day panic. It looks more like positioning is being rebuilt around a new rule: show the returns, or pay the price.

What this means for investors and the market

The immediate impact is not limited to traders holding tech names. If AI spending keeps rising without a clearer path to revenue, the pressure can spread into hardware suppliers, cloud providers, software vendors, and even utilities tied to data-center buildouts. That makes the correction important far beyond the Nasdaq tape.

Data-center projects also face a less glamorous obstacle: power and water. New facilities need large infrastructure commitments, and local communities in several regions have pushed back against fresh builds. Financing costs are rising too, which means even companies with strong balance sheets will have to choose their projects more carefully.

For now, analysts are still describing the move as a healthy correction rather than the start of a full AI crash. The next test comes with earnings from hyperscalers such as Microsoft, Amazon, and Alphabet. Investors will be looking for two things at once: sustained capex growth and a believable route to profit.

The market has not abandoned AI. It has just raised the bar. And that bar now sits much higher than a year ago, when spending alone was enough to keep the stock bid alive.

Disclaimer: This article is for informational purposes only and does not constitute financial advice. Investing carries risk, and readers should do their own research before making any decisions.