Editor's Note: On October 22, a notable capital rotation occurred in the U.S. stock market. As Meta's latest earnings report revealed massive funding needs for AI infrastructure buildout, investors began reassessing capital expenditures, financing pressures, and future returns across the AI supply chain. The selloff spread from large tech companies to semiconductor equipment, AI cloud computing, and data center infrastructure firms, with the Nasdaq index posting its largest single-day decline in four months.

But this was not a broad-based equity market exodus. In contrast to the concentrated selloff in tech stocks, the Dow Jones Industrial Average closed nearly flat, U.S. Treasuries attracted strong buying, and some traditional blue-chip stocks and companies with solid earnings rose against the trend. More notably, despite oil prices rising to a five-month high—which typically pushes up inflation expectations and bond yields—the 10-year Treasury yield fell sharply, breaking below the 5% threshold.

This means the market's short-term trading logic is shifting: investors are no longer focused solely on how fast AI demand can grow, but are beginning to examine who bears the massive capital investment, how it will be financed, and when it will translate into cash flow.

The full article follows:

AI Stocks Hit by Concentrated Selloff, Capital Flows to U.S. Treasuries

Capital Retreats from AI Supply Chain, Semiconductors and Cloud Computing Hit Hardest

On October 22, the three major U.S. stock indices diverged markedly. The Nasdaq Composite posted its largest single-day decline in four months, while the Dow Jones Industrial Average closed nearly flat. The Philadelphia Semiconductor Index fell about 6.5%, becoming one of the most heavily sold sectors of the day.

This divergence reflects that the market adjustment was not driven by synchronous declines across all sectors, but was mainly concentrated in tech companies that had previously benefited from the AI investment boom.

Among them, Meta became the representative of the shift in market sentiment. Despite the company posting record quarterly revenue, its stock price still fell about 7%, with market capitalization evaporating by more than $120 billion in a single day. Investors' focus has shifted from revenue growth to expanding AI capital expenditures and financing needs.

The semiconductor sector also came under pressure. Nvidia's stock fell about 4.4%, and since chip stocks peaked earlier this month, its market capitalization has shrunk by more than $600 billion cumulatively. Semiconductor equipment maker Lam Research fell about 11.8%, with post-earnings selling spreading further across the entire chip equipment sector.

The impact on AI cloud computing companies has been even more pronounced. CoreWeave fell about 10%, and Nebius dropped roughly 8%. Even after Nebius announced a new $3.5 billion credit facility, it failed to reverse market concerns about financing pressure on AI cloud service providers.

As AI-related debt risks draw more attention, investors shorting new AI cloud computing companies (Neoclouds) have already reaped substantial gains. According to data cited by The Economic Times, related short positions have accumulated profits of approximately $6 billion.

Bank of America analyst Tal Liani, while maintaining a buy rating on CoreWeave, believes the company may still need to raise about $20 billion to fund support for existing customer commitments.

This indicates that investor concerns about the AI industry have extended from valuations to balance sheets.

For AI infrastructure companies that rely heavily on financing for expansion, even if customer demand remains strong, their business models could still come under pressure if financing costs rise and capital recovery cycles lengthen.

In the past, growth in AI demand typically meant higher revenue expectations and valuation premiums. But now, the market is beginning to focus on another question: how much capital do companies need to invest upfront to meet these demands, and how long will it take to recoup the costs?

U.S. Treasuries become the main safe-haven destination, 10-year yield falls below 5%

As tech stocks faced concentrated selling, U.S. Treasuries became an important destination for capital flows.

On October 22, the U.S. 10-year Treasury yield fell sharply in a single day, breaking below the 5% threshold and marking the largest one-day decline in more than two years. Since bond prices and yields move inversely, this means U.S. Treasury prices rose significantly.

Joe Saluzzi of Themis Trading described the day's market action as a typical rotation: investors sold AI stocks and moved funds into bonds and major Dow components.

It is worth noting that the rally in U.S. Treasuries occurred against the backdrop of rising oil prices.

In previous months, rising international oil prices were often accompanied by heightened inflation concerns, which in turn pushed U.S. Treasury yields higher. However, the market moves on October 22 broke this pattern.

That day, Brent crude rose to $113.17 per barrel, hitting a five-month high, yet the 10-year U.S. Treasury yield still fell sharply.

This indicates that, at least on the day, investors' concerns about tech stock risks outweighed the inflationary pressure brought by rising oil prices. As capital reduced its exposure to the AI sector, U.S. Treasuries once again became an important defensive allocation tool.

From an asset allocation perspective, this shift is somewhat representative.

For some time, AI-related companies have been favored by the market, with capital willing to pay high valuations for their future growth. Now, as capital expenditures continue to expand and profit realization is delayed, some investors have begun to reassess the risks and rewards of holding high-growth stocks and are turning to Treasuries with lower credit risk and relatively certain returns.

Meanwhile, the cost of credit default swaps (CDS) on Meta's debt hit a record, showing that market concerns about AI financing risks have spread from the equity market to the credit market.

Even tech giants with massive operating scale and financing capacity are beginning to face investor scrutiny over the sustainability of capital expenditures.

However, falling U.S. Treasury yields only indicate stronger bond buying and cannot be used to precisely calculate how much capital flowed directly from tech stocks into Treasuries. Market price performance reflects changes in asset allocation preferences, not complete capital flow statistics.

Traditional blue chips buck the trend, safe-haven assets show divergence

Capital has not fully exited the stock market; traditional blue chips and companies with solid earnings buck the trend

In sharp contrast to the steep decline in the Nasdaq index, the Dow Jones Industrial Average closed nearly flat.

This divergence shows that investors are not indiscriminately selling stocks but reallocating risk exposure across different sectors.

On the day, some traditional industries, consumer companies, and companies with stronger earnings performance rose against the trend, reflecting increased market attention to earnings certainty and cash flow quality.

Among them, Blackstone's stock rose about 8.4%. The company reported earnings per share of $1.32, above the market expectation of $1.26. Its chief financial officer said that signed transactions are expected to generate fee revenue exceeding the relevant revenue actually achieved in the previous quarter.

Spotify's stock rose about 11.1%. The company added 6 million paid subscribers, bringing the total to 306 million, exceeding prior expectations, and it expects to add another 10 million users in the next quarter.

Defense company Lockheed Martin also performed strongly, with its stock rising about 7%. The company raised its 2026 sales forecast to $80 billion to $82 billion, with a record order backlog of $292 billion, including a $52.5 billion Patriot missile interceptor contract.

These companies span different sectors, but their stock performance that day shared one common feature: investors remain willing to pay a premium for relatively clear earnings growth, business demand, and order visibility.

Compared with AI infrastructure companies that rely on long-term capital investment and future growth expectations, companies with higher earnings visibility have shown stronger resilience when market risk appetite declines.

However, this rotation does not mean that all traditional industries will benefit, nor should it be simply interpreted as capital systematically withdrawing from the tech sector.

More accurately, investors are re-distinguishing the growth quality of different companies. For companies already able to generate stable cash flow through their core businesses, the market remains willing to offer support; for companies that need continuous financing and rely on revenue over many future years to recoup investment, investors are beginning to demand a clearer path to profitability.

From this perspective, the market action on October 22 looks more like a concentrated repricing of the AI high-capex model rather than a comprehensive reversal of the entire stock market's growth logic.

Gold and Bitcoin fail to absorb safe-haven demand, capital defense diverges

Although safe-haven sentiment clearly intensified in the market that day, not all traditionally perceived safe-haven assets were sought after.

Gold prices fell about 2.7%, and Bitcoin dropped about 3.8%, in sharp contrast to the major rally in U.S. Treasuries. This divergence indicates that investors did not adopt a uniform buying strategy across all safe-haven assets. U.S. Treasuries were favored for their credit attributes and relatively certain yields, while gold and Bitcoin failed to show similar defensive characteristics that day.

At the same time, global markets continue to face multiple external risks.

International oil prices surged sharply due to tensions in the Middle East. The original text cited a New York Times report saying that the United States and Israel may launch a new round of military strikes against Iran after the U.S. midterm elections on November 3.

Although oil exports from the Gulf region have recovered above pre-war levels, global oil inventories remain low, meaning any new supply disruption could amplify market volatility.

European markets are also under pressure. The yield spread between French and German government bonds has breached 150 basis points, reaching a 15-year high. Ongoing social protests and an upcoming credit rating review have intensified investor concerns about France's fiscal and political outlook.

Therefore, the popularity of U.S. Treasuries does not mean that global risks are receding. On the contrary, uncertainty over AI investment returns, energy supply shocks, and European fiscal risks are jointly weighing on market sentiment. What investors are choosing is not safe-haven assets in a broad sense, but rather assets more likely to provide relatively certain returns and liquidity.

AI Investment Bills Come Due, Wall Street Begins Repricing

AI Funding Chain Becomes Focus: Market Begins Recalculating Investment Return Cycles

The fundamental driver of this capital rotation remains investors' reassessment of AI infrastructure construction costs and investment return cycles.

Meta's latest disclosed filings show that its AI-related spending commitments amount to approximately $825 billion, higher than the nearly $700 billion level in July. The company's third-quarter free cash flow also turned negative for the first time in more than a decade.

At the same time, Meta's introduction of a new non-voting share class has drawn greater investor attention to how future AI expansion will be funded and what financing costs existing shareholders may bear.

Although Meta's revenue continues to grow, the market is clearly no longer satisfied with revenue growth alone.

Investors are beginning to ask when massive AI investments will generate sufficient cash flow, and whether expanding financing needs will erode shareholder returns.

This concern is not limited to Meta.

NVIDIA CEO Jensen Huang said demand from new AI cloud computing companies is extremely strong, but the capital needed to meet that demand remains scarce. IBM CEO Arvind Krishna noted that some large enterprise clients are already worried that certain AI spending could spiral completely out of control without generating real returns on investment. Miller Tabak strategist Matthew Maley also believes the payback period for AI investments could be much longer than the market previously expected.

These views collectively point to the core contradiction facing the AI industry: demand growth and capital returns do not necessarily move in sync.

For companies that rely on external financing to purchase chips, build data centers, and lease computing power, the business model typically requires them to bear massive capital expenditures upfront and then gradually recover costs through customer revenue over the following years.

Once the investment payback period lengthens or the financing environment tightens, companies may face larger funding gaps even as order volumes continue to grow. This also explains why new AI cloud computing companies such as CoreWeave and Nebius have seen declines significantly larger than those of some large tech companies. Large tech companies can at least rely on existing businesses to provide some cash flow support, while AI infrastructure companies that are highly dependent on debt and external financing are more vulnerable to financing costs, debt repayment, and customer performance risks.

The market is not suddenly rejecting AI demand, but rather reassessing the relationship between demand growth and capital investment. In the past, strong AI demand typically meant that chip, data center, and cloud computing companies could expect higher valuations. But now, investors are beginning to distinguish between companies that can support expansion through operating cash flow and those that need to continuously borrow or issue shares to sustain growth.

The former may still have opportunities to gain market support, while the latter are more vulnerable to financing costs and uncertainty over investment returns.

Wall Street's Next Step: Can the Capital Rotation Continue?

In the short term, changes in Wall Street capital flows still depend on several key factors.

First is the financing capacity of AI cloud computing companies.

CoreWeave will report earnings in November, and the market's focus is not only on revenue and order growth, but also on whether the company can raise sufficient funds for its existing customer commitments, and whether capital expenditures can match future cash flows.

If future earnings reports continue to show strong AI demand growth but the scale of capital companies need to invest continues to expand, the market may become further cautious about the financing capacity and valuation levels of related companies.

Second is whether U.S. Treasuries can continue to attract safe-haven buying.

On October 22, U.S. Treasury yields fell sharply against the backdrop of rising oil prices, reflecting a notable increase in market demand for safe assets. But whether this trend can continue still depends on inflation expectations, energy prices, Federal Reserve policy, and the overall economic growth outlook.

If tech stocks continue to come under pressure while U.S. Treasury yields keep declining, it would further indicate that investors are reducing risk exposure and increasing the allocation of bonds in their portfolios.

Finally, there is the question of whether the rotation within the stock market is sustainable. That day, the relative resilience of the Dow Jones Industrial Average, as well as gains in companies such as Blackstone, Spotify, and Lockheed Martin, showed that the market is still willing to support companies with relatively clear earnings prospects.

If this trend continues, market capital may further shift from the high-capex, high-growth-expectation AI infrastructure sector toward areas with stable cash flow, higher earnings certainty, or defensive characteristics.

Overall, the market adjustment on October 22 was not simply investors fleeing U.S. stocks across the board, but rather appeared to be a concentrated repricing of AI investment costs by Wall Street.

The clearest market signals that day were the concentrated selloff in tech stocks, rising U.S. Treasury prices, and the counter-trend strength of some traditional industries and companies with solid performance.

Wall Street has not abandoned growth assets; rather, it has begun raising its requirements for financing capacity, free cash flow, and certainty of investment returns.

For the AI industry chain, the key to valuation in the next phase may no longer be simply how much more computing power demand can grow, but rather how much of that demand can actually be converted into profit, and how high a cost of capital companies must bear to achieve growth.

For the broader market, the deeper change revealed by this adjustment is that as AI investment bills gradually come due, capital is shifting from optimistic expectations for future growth toward a reassessment of actual returns and asset safety.

[Original link]

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia