The US stock market experienced an unprecedented rally last night.

The Dow Jones Industrial Average surged 594 points, closing at a new all-time high.

McDonald's, Coca-Cola, Johnson & Johnson—these "old economy" stocks all danced together.

On the other hand?

The Philadelphia Semiconductor Index fell 11% in two days.

SanDisk fell over 14% in one day, dropping 27% from its peak, directly entering a bear market.

Teradyne fell 13%, KLA fell 11%, Applied Materials fell over 7%.

SK Hynix lost $160 billion in market cap in a single day.

Under the same sky, some are celebrating while others are being buried.

Let's first talk about what happened.

The trigger was a piece of news—Meta is going to sell computing power externally.

Meta's AI capital expenditure cap this year is $145 billion. It hoarded massive amounts of GPUs, but its own AI large models underperformed, leaving computing power unused.

So Zuckerberg came up with an idea—to rent out the idle computing power.

The decision itself is not a problem. The problem is that it punctured the story the market has been telling for two years.

What was the core logic behind the AI hardware sector's surge over the past two years?

Scarcity of computing power.

GPUs were in short supply, tech giants scrambled to buy, NVIDIA became the world's most valuable company, and memory manufacturers like Micron and SanDisk rode the wave.

But now Meta says: I have unused computing power, who wants it?

The narrative of computing power scarcity collapsed overnight.

How could the market not panic?

Then panic began to spread.

On July 1, the Philadelphia Semiconductor Index fell 6.27%.

On July 2, it fell another 5.44%.

Combined over two days, 11% was wiped out.

The storage sector suffered the most—Goldman Sachs' basket of memory stocks fell over 18% in two days, the sharpest two-day decline in 12 years.

The decline quickly spread to Asia. South Korea's KOSPI plunged nearly 8%, SK Hynix fell over 14%, and Japan's Kioxia plummeted 12%.

Global AI hardware has collapsed across the board.

Goldman Sachs' tracked "AI beneficiary vs. damaged stocks" pair trade fell 16% in two days—the worst performance on record.

For the past two years, the market only recognized one logic—whoever buys more GPUs is the winner.

And now?

The market is beginning to ask: You bought so many GPUs, did you actually make money?

The AI industry is moving from competing on capital expenditure to competing on capital efficiency.

Meta selling computing power, Anthropic developing its own chips, everyone starting to calculate ROI—

This is not the end of AI; this is AI transitioning from adolescence to adulthood.

During adolescence, you can rely on parents' money to spend freely.

What does this wave of AI panic mean for BTC and crypto assets?

Short-term, pressure.

Tech stock risk appetite cools, liquidity tightening expectations rise, and the crypto market cannot remain immune.

But long-term—

AI computing power moving from "scarcity" to "abundance" could be good for the crypto world.

Cheaper computing power = cheaper ZK proofs = cheaper on-chain computation = cheaper Layer2 operating costs.

Don't get the direction reversed.

"When the whole world is panicking about excess computing power, smart people are calculating—who can survive in the era of dirt-cheap computing power."

Meta selling computing power is not the death knell of AI.

It is AI's coming-of-age ceremony. #GateCard上线积分体系 #非农爆冷打压加息预期 #沃什宣告终结前瞻指引 $BTC $METAG $ETH