Meta plans to sell AI computing capacity to outsiders, and Asian chipmaker shares tumble

On July 3, 2026, shares of Asian semiconductor makers fell sharply, following a selloff on Wall Street. Meta Platforms' plan to build a business selling AI computing capacity to outside customers set off a fresh wave of worries about market oversupply.

What Meta is planning

Meta Platforms has announced that it intends to build a separate commercial business: the company will sell outside organizations access to its AI infrastructure. Until now, Meta built huge data centers solely for its own products and AI labs, primarily to train its Llama family of models. Now it wants to turn that capacity into a paid service and earn revenue, rather than only bear the costs.

The analogy is obvious: Amazon Web Services grew out of exactly this logic, openly monetizing the server infrastructure Amazon had built for its own needs. Microsoft, Google and Oracle later followed. In 2026, Meta intends to claim a share of the cloud AI computing market, with one of the largest AI infrastructures in the world. The difference is that Amazon entered an empty field in the early 2000s, whereas Meta is entering a market where strong competitors with years of head start are already entrenched.

Why did the market see this as a threat?

Valuations of NVIDIA, Broadcom, TSMC, Samsung and dozens of other companies in the semiconductor chain have rested on specific assumptions for the past two years.

• AI chips are in short supply, and GPU delivery queues stretch for months
• Tech giants raise capital spending on their own infrastructure quarter after quarter
• Analysts forecast this demand cycle to last at least several more years

When one of the largest buyers of AI hardware announces that it will now sell computing capacity itself, those assumptions begin to wobble. Companies that would previously have bought servers and GPUs could rent infrastructure from Meta instead, leaving total chip demand below forecasts. A major new supplier also risks a price war among hyperscalers, which squeezes margins across the industry.

What analysts say

Brian Belski, CEO and founder of Humilis Investments, described the situation to Bloomberg as a "crowded trade" crisis. In his words, too many investors held similar bets on AI growth in the tech sector at the same time, and any unexpected development triggers a wave of synchronized profit-taking. That explains why Asian semiconductor stocks, which had hit record highs on expectations of an AI boom, reacted to the news especially sharply.

What it means

The July 3 selloff exposed a structural contradiction in the AI rally: the market is betting at once on scarce infrastructure and on its mass availability, two scenarios that fit poorly together. Meta has not yet disclosed pricing for the new business or named its first corporate customers, so the real scale of the threat to the industry remains unclear. But the announcement alone was enough for the market to start repricing positions in the semiconductor sector worldwide.

*Meta has been designated an extremist organization and is banned in Russia.