AI spending hasn’t decreased—it’s just becoming more concentrated. The market is entering a “winner-takes-all” era.

Recently, an interesting phenomenon has appeared in the US stock market.

Even though they’re both AI concept stocks, they’ve moved in completely opposite directions.

On one side, SK hynix ADR surged 27.29% in a single day, as money went on a frenzy.

On the other side, IBM plunged more than 20% at one point before the bell due to results coming in below expectations, setting a record for the largest single-day drop in history.

Many people’s first reaction is:

Is AI not working anymore?

On the contrary.

What’s truly changing isn’t AI—it’s the market starting to redefine which companies are the real ones that can make money from AI.

⸻

AI investment enters the second half

Over the past two years, as long as it was tied to AI, the stock price almost always went up.

GPU, servers, software, cloud computing, databases…

The entire industrial chain has benefited from valuation boosts brought by AI.

Because the market believes:

AI will benefit every company.

But now, this story is starting to change.

More and more businesses are finding that their IT budgets aren’t increasing much.

They’re just slowly shifting the money originally spent on software procurement toward AI infrastructure.

In simple terms:

Previously:

A budget of 100

60 for software

40 for hardware

Now:

70 for AI GPU

20 for software

10 for other

The budget hasn’t changed.

It’s just that the money is flowing to a new place.

⸻

Why is Hynix up, while IBM is down?

The reason is actually very straightforward.

Training AI large models is inseparable from three things:

* GPU
* High-bandwidth storage (HBM)
* Compute servers

And SK hynix is one of the biggest current beneficiaries of HBM.

As demand for AI chips keeps increasing, HBM has become the scarcest resource across the entire industrial chain.

When supply can’t meet demand, profit growth naturally follows.

IBM’s situation is different.

Even though it has AI products and keeps rolling out new enterprise solutions,

the capital market cares more about:

Do these AI products actually bring new revenue?

If AI only replaces existing business and doesn’t create incremental value, then valuations will naturally be reassessed.

That’s why you get today’s market—ice and fire at the same time.

⸻

The market is starting to stop believing in “big and all-encompassing”

Previously, the investment logic was:

AI will make the entire tech sector earn together.

Now, the market believes more that:

AI will only make the most core companies earn more money.

As for other companies, they may simply end up bearing higher costs.

So lately, everyone has been noticing:

Capital is becoming more and more concentrated.

Fewer and fewer stocks are hitting new highs.

More and more stocks are starting to fall behind.

This is actually the classic:

Winner Takes Most.

⸻

What does this mean for Crypto?

Many people think this is just a matter of US stocks.

In fact, the crypto market is going through the same kind of change.

Back in the 2021 bull market,

as long as a project’s name included AI, Layer2, GameFi,

money would come in.

What about now?

The market has started asking three questions:

Are there real users?

Is there sustained revenue?

Is there cash flow or Buyback?

The narrative still matters.

But it’s no longer enough.

What really stays are the data that shows products, revenue, and user growth.

In this respect, both US stocks and Crypto are becoming increasingly consistent.

⸻

AI won’t end—it will only get more expensive

Many people think:

The AI bubble is about to burst.

Actually, from the perspective of corporate capital expenditures, that’s not the case.

AI investment is still increasing.

It’s just that capital is becoming more concentrated.

More and more money is flowing to the most core infrastructure—chips, compute power, data centers, and so on.

And companies that can’t prove that AI can create profits are starting to be repriced by the market.

In the next few years, what we may see isn’t AI cooling down.

It’s that within the AI industry, differentiation will become more and more severe.

The winners keep innovating and keep hitting new highs.

The losers are gradually forgotten by the market.

⸻

Written at the end

The biggest change in the market has never been that the hype disappears—it’s that capital has started to get smarter.

Before, you could go up just by telling a story.

Now, the market is starting to look at profits.

Before, standing on the right bandwagon was enough.

Now, you must prove you can truly make money.

Whether it’s AI or Crypto, the next phase isn’t about who just rides the hottest trend—it’s about who can continuously create value.

The AI story isn’t over yet. It’s just that the market is starting to eliminate supporting roles and only hand the chips to the real main characters.
#PreIPOs第二期OpenAI认购 #预测世界杯阿根廷VS英格兰 #USDT充值理财双重奏 $BTC ‌$ETH ‌$GT ‌

LAYOUT REFERENCE (source): total_lines=194, non_empty_lines=99, blank_lines=95