Meta’s Computing Power "Upsets the Table": Is the Golden Age of AI Chips About to Hit the Brakes?

Waking up to big news: Meta’s Mark Zuckerberg announced that Meta will officially start selling its own AI computing cloud services to external customers.

As soon as the news broke, the chip sector in the US stock market was hit by a wave of heavy selling:

• Philadelphia Semiconductor Index: plummeted 6.27% in a single day
• Micron: plunged 10.57%
• SanDisk: plunged 10.62%
• Intel: plunged 9.03%
• Corning: plunged 13%

A computing power storm unleashed by Meta is dragging the entire AI hardware supply chain into a heated debate about a "demands ceiling."

1. Meta’s Pivot: From Spending Big to Selling Shovels

To understand why the market reacted so violently, we first need to look at what Meta has been doing over the past few years.

Over the last two years, Meta has been one of the most aggressive "buyers" in the AI arms race. It poured $125 billion to $145 billion in annual capital expenditure, frantically purchasing GPUs, network equipment, optical modules, power, and cooling infrastructure – all to catch up with OpenAI and Anthropic in the race for large language models.

The problem? The money was spent, but the models didn't deliver. Meta's own large models consistently lagged behind its main competitors in performance, leaving much of its deployed computing power idle or operating at low efficiency.

To put it figuratively: This was an AI-era marathon. Meta was desperately chasing the leaders on the track, only to find itself falling further behind. So, it sat down by the roadside and started selling water to the runners passing by.

The moment it sat down, it became a "computing power seller."

Specific figures better illustrate the scale of its resources:

• By the end of 2025, Meta's AI computing power was equivalent to approximately 2.5 million H100 GPUs, with a total power capacity of about 2GW.
• In 2026, Meta's capital expenditure guidance is $135 billion, corresponding to an additional 2-3GW of computing capacity.
• A simple projection suggests that by the end of 2026, Meta's total computing capacity could approach 5GW.

What does this mean? It means Meta's AI computing capabilities may already surpass those of any tech company except the largest hyperscale cloud providers. Crucially, Meta itself cannot fully utilize this capacity.

Therefore, the logic of renting out idle computing power makes sense—letting assets sit idle is wasteful. Monetizing them can at least improve the balance sheet.

2. Why Did Chip Stocks Fall? Two Market Narratives

Regarding Meta selling computing power, two different interpretations currently exist in the market.

Interpretation 1: Bearish – The "Computing Power Glut" Signal Fired

The bearish logic chain is straightforward:

Meta has been one of the largest buyers of AI chips over the past two years. Now, it is starting to rent out its surplus capacity – signaling that it no longer needs to buy as much for itself.

If even a player of Meta’s caliber realizes it "bought too much," what about other companies? Is the large model arms race shifting from "full sprint" to "hitting the brakes"?

The further implication is that once a consensus on "computing power glut" forms, expectations for demand growth in the entire AI hardware market will be significantly downgraded. Key players in the supply chain like Samsung, TSMC, Micron, and Nvidia could face slowing order growth or even order cuts. The two-year bull run narrative for the hardware supply chain might be starting to fade.

Interpretation 2: Bullish – "Selling Shovels" to Continue the Marathon

The bullish counterargument is equally strong:

Hundreds of billions of dollars Meta spent on GPUs in the previous two years are sunk costs. Monetizing these idle assets now is not abandoning the race but a return to business rationality.

If renting out computing power can generate revenue, Meta's confidence in subsequently purchasing GPUs, network equipment, and optical modules could actually increase. A self-sustaining cycle emerges: spend money on equipment → recover costs by renting idle capacity → use recovered funds to buy more equipment. This is a positive cycle, not a zero-sum game.

3. Who Wins? The Market’s Answer is Clear

These two viewpoints are currently clashing, and volatility in chip stocks is unlikely to subside in the near term. However, one thing is certain: the capital market is rewarding Meta’s decision to "sell water" with real money.

In the trading session following the announcement, Meta’s own stock price did not crash alongside the chip stocks. Instead, it surged nearly 9%. This clearly shows investor sentiment: regardless of how chip stocks fare, this move is a positive for Meta.

Why is this?

The core reason isn't "how much money renting out computing power will make." Even if Meta rents out all its surplus capacity, the short-term incremental net profit is hard to estimate accurately. Perhaps the initial figure is only $2 billion to $3 billion – not a critical number for a company with annual revenues exceeding $100 billion.

What the market truly cares about is the shift in attitude.

In recent years, Wall Street's biggest anxiety about Meta wasn't its GPU purchases, but its "unlimited spending." Annual capital expenditures of $125 billion to $145 billion looked like a bottomless pit, growing investor skepticism about return on investment (ROI).

Against this backdrop, Zuckerberg’s willingness to monetize surplus computing power – even if the amount is small – sends a signal: management is starting to focus on capital efficiency. The "crazy spending" phase may be peaking.

This is the signal Wall Street has been waiting for. So, even as chip stocks plummeted, Meta’s stock surged against the trend. Capital is not rewarding the computing power rental business itself, but the strategic shift from "unlimited arms race" back to "business rationality."

4. The Real "Black Swan" May Not Have Arrived Yet

Although chip stocks took a heavy hit, declaring the end of the AI hardware bull market might be premature.

Currently, aside from Meta’s move itself, the biggest uncertainty is whether other tech giants will follow suit.

Meta is the first major tech company to publicly state, "We bought too much computing power and need to rent it out." What about Microsoft and Amazon? If these cloud giants, also heavily investing in AI infrastructure, buckle under capital market pressure and announce "rational investment" or reduced capital expenditure plans – that would truly shake the foundation of the entire AI hardware supply chain.

So far, Microsoft and Amazon have not released similar signals. But if "capex cuts" become a collective action among tech giants in the coming weeks or months, the value of AI chip hardware will need to be reassessed.

In other words, the market is shifting from one question to another: The issue isn't Meta itself, but whether Meta is the first domino to fall.

5. Final Thoughts: Volatility is Certain, Direction is Not

For investors focused on chip stocks, the current situation can be summed up in one sentence: Uncertainty is extremely high, and short-term volatility will intensify.

How significant will Meta's rental of computing power be for chip demand? Is it an isolated event or the start of an industry trend? Will Microsoft and Amazon follow suit? These questions remain unanswered. Until answers emerge, the chip sector is likely to experience continued price swings.

For investors looking to trade chip stock volatility, the BIT (formerly Matrixport) platform offers access to trade real chip stocks listed on NASDAQ – covering core targets like Micron, Intel, AMD, and Nvidia. You are buying actual shares registered within the DTC system, not tokenized certificates, allowing flexible trading and transfer via ACATS.

Combined with BIT's 24/7 near-instant USDT deposit channel, even if you are in the Asian time zone, you can have your funds ready before the US stock market opens, ensuring you don't miss any important market windows.

The story of AI hardware is not over yet, but the script is being rewritten.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. The market interpretations and data herein are based on publicly available information and may contain lags or biases. The US stock market and chip sector carry high volatility risk; past performance does not guarantee future results. Specific terms of related trading services are subject to the official statements of the BIT platform. Users in different regions must confirm compliance on their own. Investing involves risks. Please consult a professional advisor before making decisions and bear the corresponding risks and consequences.