After the US stock market crash, has the century-long top arrived? Is the AI bubble about to burst?

Today, SK Hynix's earnings missed expectations, triggering another猛烈 sell-off in Asian tech stocks, leaving the market in turmoil.

Looking back at this correction:

• The Nasdaq has fallen about 8%–10% from its peak;
• The Philadelphia Semiconductor Index has dropped roughly 25% from its June high;
• South Korea's KOSPI index plunged 10.84% in a single day, marking a 35.81% decline from its all-time high.

But just as chip stocks were bleeding, Apple reclaimed the title of the world's most valuable company, and the Dow Jones Industrial Average hit a new high. It seems funds haven't entirely left the US stock market; they are simply rotating out of crowded AI hardware trades into assets with more stable cash flows and less valuation pressure.

This has split the market into two camps: one shouting "the AI bubble has finally burst," while the other insists "this is just a healthy correction; the AI narrative has never changed."

So, the real question is:

Is the US stock market facing a "once-in-a-century top," or is it still climbing a "wall of worry" on its bull run? Is the AI bubble about to fully burst, or is it merely undergoing a phase of valuation repricing?

⬇️ Below, let's watch the debate between the bulls and bears.

Bullish Side: The AI Narrative Remains Intact, This Isn't a Bubble Burst

Debater 1: Cathie Wood @CathieDWood｜ARK Invest Founder｜XHunt Rank: 170

Cathie Wood believes the market is still climbing a classic "wall of worry."

True market tops rarely occur when everyone is skeptical; they are born from extreme optimism – when investors universally believe "the sky is the limit," that's when risk can truly get out of control.

The current market is filled with controversy, fear, and doubt, which doesn't resemble a typical ultimate top.

Core View: This is likely a normal correction within a bull market. The long-term logic of disruptive innovation hasn't changed. A real top usually appears when consensus reaches extreme agreement.

Debater 2: RamenPanda @IamRamenPanda｜XHunt Rank: 3994

RamenPanda shifts his focus from AI hardware to the software and application layer.

The market worries that massive investments in data centers and computing infrastructure will lead to overcapacity. However, these infrastructures ultimately need to generate returns through software, platforms, and commercial applications.

The internet era also experienced a phase of infrastructure overbuilding. Cisco was the star in the early days, but the true long-term winners were Google, Microsoft, and Amazon that rose later.

The AI era may follow a similar path: hardware investment overheats first, and then value gradually shifts to software and application layers.

Core View: Infrastructure overcapacity might be a necessary stage in a technological revolution. The current phase is more about transitioning from hardware building to application value realization, not the end of the AI era.

Debater 3: Investment Talk Jun @TJ_Research｜XHunt Rank: 9918

Following SK Hynix's earnings release, its stock price initially opened about 8% lower but quickly rebounded.

Investment Talk Jun believes this contrasts with the "sell the news" pattern seen after Micron's earnings and could be a short-term bottoming signal.

If bad news is already fully priced in by the market, even an earnings miss might see selling pressure begin to wane. More importantly, there is no sign yet of a systemic weakening in AI demand guidance.

Core View: Short-term sentiment has been vented, but AI demand hasn't been disproven. The market now needs confirmation from data, not more panic.

Debater 4: Nico Investment Wisdom @tychozzz｜XHunt Rank: 13869

Nico argues the biggest lesson from this crash is: no matter how strong the fundamentals or low the P/E ratio, when an asset is at a high level, stock prices can still plummet if the market decides to compress valuations. The memory and storage sector is a prime example.

However, "valuation compression for stocks" doesn't equal "the AI industry thesis being disproven."

The current forward P/E of the Nasdaq is close to the level seen at the end of March this year. Instead of panic selling, the key is to control leverage, set profit-taking targets in advance, and hold indices and quality companies that can truly deliver on earnings.

Core View: Sentiment can quickly depress valuations, but ultimately, earnings determine stock prices. The AI narrative hasn't changed; it may only be a matter of time before quality assets return to their previous highs.

Debater 5: US Stock Sage @hanking66｜XHunt Rank: 44611

The US Stock Sage's view is more direct: don't sell stocks when fear is at its peak.

In his view, this decline already shows the emotional characteristics of a "final capitulation" – despair, suffocation, breaking points, where investors interpret any news in the most bearish way.

The fact that many stocks were heavily sold off before their earnings reports also suggests pessimistic expectations may have already been priced in.

Core View: Buy the dip. If you can endure the most extreme phase of sentiment, opportunities likely belong to those still able to step in.

Bearish Side: Cracks in AI's Creditworthiness Are Appearing, The Bubble Is Deflating

Debater 1: ZeroHedge @zerohedge｜Prominent Macro Analysis Media｜XHunt Rank: 200

ZeroHedge believes the market is re-entering a phase where it prices in the idea that "circular financing is actually terrible."

Multiple warning signs are already present:

• The Nasdaq 100 is nearing a 10% technical correction;
• Semiconductor and momentum factor trades have been hit hard;
• Cracks are appearing in the creditworthiness of AI-related companies;
• Deals involving companies like Nvidia and OpenAI are again raising circular financing concerns;
• Low-cost models and open-source tools from China are continuously pressuring profit expectations.

If AI giants are investing in each other, buying from each other, and contributing to each other's revenue, it's hard for the market to determine how much comes from genuine end-user demand versus capital circulating within the industry chain.

Core View: The credit expansion within the AI industry is hitting a wall. The phase of bubble deflation may arrive sooner than most investors expect.

Debater 2: Phyrex @PhyrexNi｜XHunt Rank: 784

Phyrex is more focused on the leverage risk underlying US stocks.

Currently, US investors have borrowed approximately $1.53 trillion to buy stocks, with broker net credit balances falling to around negative $1.061 trillion. Margin debt increased by about $86 billion in a single month and has risen for three consecutive months.

This means US stocks are not only overvalued but also increasingly dependent on borrowed funds. If new capital inflows slow down, the mechanical buying that drove past rallies could turn into mechanical selling.

This risk can also transmit to Crypto. During cross-market deleveraging, BTC is more likely to be sold first as a high-beta risk asset rather than immediately rising independently as "digital gold."

However, short positions in US stocks are also at high levels. The short interest ratio for the Russell 3000 is near 6%, and for the S&P 500 around 3.5%, so the possibility of a short squeeze rebound exists.

Core View: The real risk isn't necessarily the AI technology being disproven, but the combination of high valuations, high leverage, and slowing capital inflows. Once deleveraging starts, both US stocks and BTC could face mechanical selling.

Debater 3: Peter Schiff @PeterSchiff｜Economist, SchiffGold Founder｜XHunt Rank: 1486

Peter Schiff argues that AI itself is a real technology, but AI-related stocks have formed a serious bubble.

Low-cost models from China, open-source technology, and intensifying industry competition will continue to pressure the profit margins and moats of US AI companies.

If massive capital expenditures cannot be converted into matching cash flows, more valuation "air" still needs to be squeezed out.

Core View: AI technology is real, but the AI stock bubble is equally real. The bubble may have already burst, and what follows will be a continuous period of deflation.

Debater 4: CryptoPainter @CryptoPainter｜XHunt Rank: 2207

CryptoPainter views the South Korean stock market crash as a stress test for the AI infrastructure narrative.

He doesn't deny the demand for AI computing power and data centers but warns against the circular investment, excessive financing, and high leverage built around this narrative.

Therefore, this correction may not replicate the full-scale industry collapse of 2000 but could instead mirror the violent deleveraging seen in the Korean market: the technological direction is still valid, but asset prices suffer severe pullbacks due to fragile valuations and financing structures.

For Crypto, if BTC breaks down again, one of the most probable external triggers remains further declines in the US stock market.

Core View: AI technology is real, but that doesn't mean related asset prices are reasonable. Compared to the industry going to zero, the more realistic risk is the unwinding of financing and leverage, which can then transmit to BTC via US stocks.

Debater 5: Hellen @peng_hellen Mother｜XHunt Rank: 40426

Hellen compares today's AI craze to the British Railway Mania and the Internet Bubble.

Railways and the internet truly changed the world, but capital frenzy still led to overbuilding of infrastructure. Some projects saw losses worsen the more they invested and operated.

This mirrors the dilemma faced by some large model companies: more users mean higher computing and inference costs, but revenue may not cover the investment.

The internet didn't disappear after the bubble burst, but the Nasdaq still fell for 31 consecutive months, declining about 78% from its peak.

Core View: A great technology and a huge bubble can exist simultaneously. AI might change the world, but the prices of related assets can still be severely disconnected from their true value and demand.

Summary of Both Sides' Views

The Bulls Argue

The current crash is mainly caused by the collapse of crowded trades and sentiment-driven valuation compression. There is no data yet to prove that AI demand has systemically collapsed.

New competitors and low-cost models may not necessarily destroy US tech giants; they could instead lower the barrier to AI adoption and expand the application market.

The AI narrative is not over, and the market may continue to advance up a "wall of worry."

The Bears Argue

The moats of AI companies are not as deep as the market imagines.

High valuations, massive capital expenditures, leverage, and circular financing have all underpinned the past rally. If revenue and cash flow fail to match the spending, the entire valuation system will be repriced.

Technological revolutions won't disappear, but investment bubbles can certainly burst prematurely.

Finally

What US tech stocks are experiencing is not a wholesale rejection of AI's long-term logic. Instead, the market is beginning to recalculate the math:

If returns on capital expenditure are insufficient, circular financing recedes, and low-cost models from China keep pressuring prices, how much of a moat is really left for US tech leaders?

The Nasdaq's correction, the halving of memory chip stocks, and the crash of South Korean chip stocks are essentially a concentrated release of this concern.

But declaring a "once-in-a-century top has arrived" might be premature.

The true watershed moment isn't in today's sentiment-driven valuation compression, but in the coming quarters and beyond: whether AI revenue and free cash flow can match the enormous capital expenditures, and whether industry chain demand can spread from circular procurement among giants to genuinely paying end-users.