Intel's stock price swung wildly, and the "Pop Mart" of the AI world was born

Intel's recent stock price movements bring to mind a company that seems completely unrelated: Pop Mart.

One sells blind boxes, the other makes chips, yet they have staged a similar drama in the capital markets.

On November 2, Intel's stock plunged 8.8%.

Just one trading day earlier, on October 30, it had just crashed 28.58%, wiping out about $140 billion in market value in a single day.

This traditional hardware giant had just released its third-quarter earnings: revenue up 21% year-over-year, with data center and AI business revenue surging 78%. Yet weak fourth-quarter guidance still sent investors fleeing.

However, even after the late-October crash and the consecutive declines in early November, Intel's cumulative gain this year still exceeds 60%.

Looking back over the past few months, Intel's stock doubled in April, hit its yearly high in June, plunged in July, rebounded in September, and crashed again in October.

1

In April of this year, Intel could well be called the darling of the capital markets, surging 114.1% that month.

On April 24 alone, Intel's stock rose 23.6%, marking its largest single-day gain in nearly 39 years.

Wall Street analysts grew excited as well. That day, Evercore ISI upgraded Intel to "outperform" and raised its target price from $45 to $111; JPMorgan, while maintaining its "underweight" rating, also raised its target price from $35 to $45.

Intel shareholders on Reddit were even more jubilant. Some celebrated buying in at over $20, some were already imagining Intel catching up to TSMC's market value. One netizen even exclaimed: "This rally is just getting started."

One netizen, who was preparing to short Intel with $4,100 earned from delivering food, could only remark sourly: Intel merely didn't go bankrupt last night.

The following two months, Intel's stock kept climbing. By the end of June, it hit an intraday high of $142.35, nearly four times its $36.90 at the end of 2025.

The air inside and outside the community was filled with joy.

Unfortunately, in the capital markets, any celebration at any time is like opening the champagne at halftime. One moment the stock is held high, the next it can take a sharp turn downward.

In July, Intel's stock suddenly reversed, plunging 35.4% for the month, and August brought no relief.

Watching the earlier gains slip away bit by bit, Intel suddenly perked up again in September. As demand for AI server CPUs regained attention, the stock rebounded 34.3% for the month.

Investors, who had just been badly frightened, seemed to see hope for Intel's revival once again.

By October 19, JPMorgan, which had been bearish on Intel, finally upgraded its rating from "underweight" to "neutral."

As it turned out, 11 days later, Intel suffered the October 30 crash.

That day the stock closed at $65.76, down about 54% from its June high, and for the whole of October, Intel's stock fell 45.3% cumulatively.

That wasn't the end. On November 2, Intel's stock fell another 8.8%.

A few months ago, the market was still imagining Intel reclaiming semiconductor dominance; now the stock has fallen by more than half from its high. Fortunately, its starting point at the beginning of the year was low enough that, even after all this turmoil, as of November 2, Intel's cumulative gain this year still reached 62.52%, offering some comfort to investors who bought early in the year.

As for those who entered at the June high, they probably aren't feeling so good.

Intel's weak guidance also added fuel to an already tense semiconductor market. On October 30, AMD fell 10.9%, Arm about 10%, Nvidia 4.2%, and both Broadcom and Micron dropped more than 5%. The Philadelphia Semiconductor Index fell 7.2% that day, and almost the entire chip sector was hit.

CoreWeave and Nebius, which provide computing power leasing services for AI, saw their stocks fall about 37% and 35% respectively in October. Just a few months ago, the market was excited about how many GPUs Neoclouds could buy; now it is starting to worry about how these companies will repay the money they borrowed.

Musk even predicted that more than 200 Neoclouds will face massive death and consolidation, with only a few survivors remaining in the end.

On November 2, Anthropic updated its prospectus, disclosing that third-quarter revenue grew 47% quarter-over-quarter, but the growth rate was clearly lower than the 144% in the second quarter. According to the latest reports, some institutional investors want to keep its IPO valuation between $1.5 trillion and $1.6 trillion, lower than the roughly $2 trillion previously discussed.

According to foreign media reports, OpenAI has cut its new funding target from $30 billion to $20 billion, and lowered its proposed valuation from $1.4 trillion to $1.25 trillion, with about $11 billion in funding commitments secured so far.

Although these cannot yet be considered completed deals, the market's mood has indeed changed.

2

Intel missed the first wave of the AI boom dominated by GPUs, but as tech giants continued to expand data centers, it found itself facing a surge in demand for server CPUs.

GPUs excel at large-scale parallel computing, while CPUs handle general computing, task scheduling, and data processing. As AI inference and agent applications expand, server CPUs are becoming increasingly sought after.

On April 9, Intel announced an expansion of its multi-year partnership with Google, continuing to supply Xeon processors for Google Cloud's AI infrastructure and jointly developing custom infrastructure processors.

On June 16, Intel also announced that its advanced 18A-P process had entered the risk production stage. Two days later, Trump claimed that Apple would partner with Intel to design and manufacture chips (neither side had confirmed this at the time).

These gains are not just a story: in the third quarter of this year, Intel's Data Center and AI business revenue reached $7.3 billion, up 78% year over year, marking its strongest data center revenue growth in 15 years.

However, a closer look at this earnings report reveals some noteworthy details.

In the third quarter, the average selling price of Intel's server products rose 54% year over year, while shipments grew 10%. The revenue surge largely came from a higher mix of high-end products and price increases on some products.

Intel's server CPUs are already in short supply, but precisely when it most needs to sell more chips, Intel is constrained by wafer manufacturing capacity and advanced packaging substrate supply, and production cannot keep up.

At the same time, strong demand for high-end memory from AI data centers has pushed up memory prices and also raised the manufacturing cost of ordinary PCs, further hurting the consumer market—yet the PC business remains Intel's revenue mainstay. In the third quarter, Client Computing Group revenue was $8.1 billion, down 5% year over year, accounting for nearly half of the company's total revenue.

So on one side, the most sought-after chips are not available in sufficient quantities to sell; on the other, the traditional business that accounts for half the company cannot sell.

Analyst Jay Goldberg, in an interview with the media, offered rather sharp criticism of Intel's capacity planning. In his view, Intel produced too many consumer-grade chips precisely in the weakest PC market in 15 years, yet in the strongest growth cycle of the semiconductor industry, it could not produce enough AI server chips.

He attributed the problem directly to poor planning and believed that the sharp selloff reflected investors' loss of confidence in management's credibility.

In addition, in October a serious explosion occurred at Intel's factory in Israel, interrupting production for about two weeks and expected to reduce revenue by about $600 million combined in the fourth quarter and the first quarter of 2027.

These pressures ultimately converged in the fourth-quarter guidance released on October 29.

Although third-quarter revenue and profit performance were decent, Intel's fourth-quarter revenue guidance was only $15 billion to $16 billion, below analysts' previous expectation of $17.07 billion. Adjusted earnings per share are expected to be $0.25, also far below the market expectation of $0.43.

Investors had still been looking forward to Intel's revival, but when they opened the earnings report, next quarter's expectations were a direct shock.

What's more troublesome is that Intel still needs to keep spending.

In the third quarter, Intel's foundry division revenue was about $6.7 billion, but about $6.3 billion of that came from internal business; revenue truly from external customers was only $411 million, and was mainly driven by a non-recurring contract. The foundry division as a whole still lost $1.853 billion.

To expand capacity and advance advanced processes, Intel expects capital expenditure this year to reach $20 billion, and to increase substantially in 2027. The chief financial officer even said that positive free cash flow may still not be achieved next year.

3

Speaking of the fickleness of the capital markets, Pop Mart's shareholders should also have deep personal experience.

In March this year, Pop Mart delivered a rather impressive report card: 2025 revenue up 184.7% year over year, with net profit also growing severalfold. As a result, the stock price plunged 22.5% that day.

But as we have discussed before, Pop Mart runs a business highly dependent on consumer sentiment.

Today consumers are willing to line up and pay premiums for Labubu; tomorrow they may fall in love with another trendy toy. How long an IP can stay hot, and when the next hit will appear, no one can say for sure.

The capital market's valuation of it naturally keeps swinging between frenzy and doubt.

Put simply, believe and it rises; doubt and it falls.

Intel's Data Center and AI business revenue grew 78% year over year, and that was not earned from investors' imagination.

Yet these results seem unable to satisfy the market's appetite—or perhaps they do not deserve the price the market has given them.

So investors began to scrutinize. Guidance that misses expectations, a slight decline in profit margins, or a capital expenditure item with no return in sight could all suddenly make them turn hostile.

When high-speed growth has already become a market consensus, growth itself is no longer enough to support ever-rising valuations.

Investors also need to believe that such growth can continue, that the money spent today can be earned back in the future, and preferably earned back in ever-greater amounts.

Previously, the market was willing to believe that surging revenue at chip companies was proof of strong AI demand, and that tech giants' increasing capital expenditure meant AI still had broad room for growth. Now the business is still the same business, but Wall Street has suddenly put on a different pair of glasses—speed alone is not enough; acceleration is required.

Except that Pop Mart still has to worry that consumers may suddenly stop liking Labubu one day, while the AI industry may encounter a capital market selloff first, even while demand remains strong and revenue continues to grow.

From selling blind boxes to making chips, the capital market's game is surprisingly consistent: when it believes in the future, no amount of growth is too much; when doubt appears, no amount of growth is enough.

Shareholders had better keep some fast-acting heart pills on hand and be ready to "collapse to the floor" at any moment.

This article comes from the WeChat public account "Letter List" (ID: wujicaijing), author: Yuan Xinyue, published by 36Kr with authorization.