Meta Decided to Sell AI Computing Power to Third Parties — Asian Chipmakers' Stocks Crashed

Meta announced plans to sell access to its AI computing power to third-party companies. The news immediately crashed semiconductor stocks in Asia: investors feared the AI infrastructure market would become oversaturated. Analyst Brian Belsky calls this 'crowded trades' — too many investors held identical bets on the AI boom.

AI-processed from Bloomberg Tech; edited by Hamidun News
Meta Decided to Sell AI Computing Power to Third Parties — Asian Chipmakers' Stocks Crashed
Source: Bloomberg Tech. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceBloomberg Tech ↗
Editor-in-Chief
Key metric #1AI chips are in short supply, GPU delivery queues stretch for months
Key metric #2Tech giants increase capital investments in their own infrastructure quarter after quarter
Key metric #3Analysts forecast this demand cycle to continue for at least several more years

July 3, 2026: Asian semiconductor manufacturers' stocks plummeted following a Wall Street selloff triggered by Meta Platforms' plan to establish a business selling AI computing capacity to third-party clients, sparking a new wave of concerns about market oversupply.

What Meta Plans to Do

Meta Platforms has announced its intention to build a separate commercial division: the company will sell external organizations access to its AI infrastructure. Until now, Meta had constructed massive data centers exclusively for its own products and AI labs—primarily for training the Llama model series. Now the company wants to transform these capabilities into a paid service and generate revenue rather than simply bearing costs.

The analogy is obvious: Amazon Web Services once grew from exactly this logic—the monetization of server infrastructure that Amazon had built for its own needs. Later, Microsoft, Google, and Oracle repeated this step. In 2026, Meta intends to capture a portion of the cloud AI computing market, possessing one of the world's largest AI infrastructures. The difference is that Amazon entered an empty field in the early 2000s, whereas Meta is entering a space where strong competitors with years of competitive advantage are already entrenched.

Why the Market Perceived This as a Threat

Valuations of NVIDIA, Broadcom, TSMC, Samsung, and dozens of other semiconductor chain companies over the past two years were built on specific assumptions.

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

When one of the largest buyers of AI equipment announces that it will now sell computing capacity itself, these assumptions begin to wobble. Companies that would previously have purchased servers and GPUs can instead rent infrastructure from Meta—and aggregate chip demand will fall short of forecasts. Moreover, the emergence of a new major supplier threatens a price war among hyperscalers, which pressures margins across the industry.

What Analysts Are Saying

Brian Belski, CEO and founder of Humilis Investments, described the situation in a Bloomberg comment as a "crowded trade" crisis. According to him, too many investors simultaneously held similar bets on AI growth in the technology sector, and any unexpected development triggers a wave of synchronized profit-taking. This explains why Asian semiconductor stocks, which had reached record levels on AI boom expectations, reacted to the news particularly sharply.

What This Means

The July 3 selloff exposed a structural contradiction in the AI rally: the market is simultaneously betting on infrastructure scarcity and on its mass proliferation—two scenarios that are poorly compatible. Meta has not yet revealed the pricing policy for the new division or named initial corporate clients, so the true scale of the threat to the industry remains unclear. However, the announcement alone was sufficient reason for the market to begin revaluing positions in the semiconductor sector worldwide.

*Meta is recognized as an extremist organization and banned in the Russian Federation.

Story

Key Questions Answered

What Meta Plans to Do?

Meta Platforms has announced its intention to build a separate commercial division: the company will sell external organizations access to its AI infrastructure. Until now, Meta had constructed massive data centers exclusively for its own products and AI labs—primarily for training the Llama model series. Now the company wants to transform these capabilities into a paid service and generate revenue rather than simply bearing costs. The analogy is obvious: Amazon Web Services once grew from exactly this logic—the monetization of server infrastructure that Amazon had built fo

Why the Market Perceived This as a Threat?

Valuations of NVIDIA, Broadcom, TSMC, Samsung, and dozens of other semiconductor chain companies over the past two years were built on specific assumptions. AI chips are in short supply, GPU delivery queues stretch for months Tech giants increase capital investments in their own infrastructure quarter after quarter Analysts forecast this demand cycle to continue for at least several more years When one of the largest buyers of AI equipment announces that it will now sell computing capacity itself, these assumptions begin to wobble. Companies that would previously have purch

What Analysts Are Saying?

Brian Belski, CEO and founder of Humilis Investments, described the situation in a Bloomberg comment as a "crowded trade" crisis. According to him, too many investors simultaneously held similar bets on AI growth in the technology sector, and any unexpected development triggers a wave of synchronized profit-taking. This explains why Asian semiconductor stocks, which had reached record levels on AI boom expectations, reacted to the news particularly sharply.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…