The National Day holiday has entered its final day. During the week when A-shares were closed, global technology stocks advanced collectively. From October 1 to 6, Brazil's BOVESPA index rose 10.46%, and Japan's Nikkei 225 gained 5.89% cumulatively, ranking among the top performers among major global stock indices. Markets in Taiwan, China and South Korea strengthened overall, driven by the memory chip and AI hardware sectors. On Wall Street, both the S&P 500 and the Nasdaq refreshed their closing records. (All content in this article is an objective listing of data and information and does not constitute any investment advice.)

Interestingly, this round of tech stock gains occurred precisely against the backdrop of U.S. Treasury yields surging all the way up. Since September, the yield on the 10-year U.S. Treasury note has climbed from around 4.8% to 5.31%, while the 30-year yield rose to 5.66%, both reaching highs not seen in more than two decades. On October 6, the two edged down slightly to around 5.28% and 5.65%, still at elevated levels.

Wall Street expects overall earnings for S&P 500 constituent companies to grow 30.6% year-on-year in the third quarter, with the technology sector projected to grow 66.5%. Goldman Sachs calculations also show that AI infrastructure-related companies may contribute more than half of the incremental earnings per share for the S&P 500 in the third quarter. In other words, high interest rates have compressed valuations, but AI-related earnings expectations have propped the index back up.

On October 6, the S&P 500 rose 0.58% and the Nasdaq rose 0.45%, both refreshing closing records. The combined market capitalization of the "Magnificent Seven" tech giants has approached $25 trillion, with a weighting of more than 40% in the S&P 500. Nvidia's share price has reclaimed its historical high, with a market value approaching $5.8 trillion, just one step away from $6 trillion. So far this year, Nvidia's stock has gained about 28% cumulatively, adding roughly $1.2 trillion in market value, making it the single largest contributor to the S&P 500's rise.

It is worth noting that on October 6, while U.S. indices hit new highs, the Russell 2000 small-cap index actually fell 0.6% that day. This shows that the new highs in U.S. indices are mainly driven by a handful of tech giants. Market-cap-weighted indices have a characteristic: the larger the company, the greater its impact on the index. A few giants continuing to rise means that even if a large number of small and mid-sized companies perform modestly, the S&P 500 and Nasdaq can still reach new highs.

Under high interest rates, the advantages of large technology companies have become even more prominent. In September, the Federal Reserve raised rates by 25 basis points, lifting the federal funds rate target range to 3.75%-4%. After financing conditions tightened, large tech companies with more cash, higher profit margins, and still-growing revenue are typically more resilient than smaller companies. Some funds therefore view them as a kind of "relative defense," rather than purely high-beta growth stocks.

So-called "relative defense" does not mean lower risk. Tech giants are investing more capital in AI. Beyond chips, they also need to build data centers, purchase servers, find land, and secure long-term power. The larger the investment, the higher the demands on future revenue, profits, and cash flow. Once the pace of returns falls short of expectations, high interest rates will likewise magnify the pressure.

The next stop for AI investment has already extended from computing power to electricity. On October 6, Google's parent company Alphabet splurged $4.3 billion on power, reaching a long-term power supply arrangement totaling 3,590 megawatts with Constellation Energy, including 890 megawatts of new nuclear capacity. Constellation plans to invest more than $4.3 billion for this. A week earlier, Amazon also signed a 20-year agreement with the company to support nuclear plant expansion.

These deals show that AI competition is no longer just chip competition. Data centers need large and stable power supplies. Companies such as Google and Amazon locking in nuclear power and grid resources in advance is essentially preparing for future computing expansion. For the capital markets, this not only shows that AI infrastructure investment is still continuing, but also that the funds needed to sustain this round of investment are growing.

SpaceX's new financing has put this issue squarely on the table. According to reports, SpaceX plans to raise about $40 billion, including $10 billion in bank loans and $30 billion in investment-grade debt, with the funds mainly used to purchase Nvidia chips, and the related transactions are expected to be completed in 2027. In June this year, SpaceX raised a net amount of about $85.7 billion through an IPO, then completed a $25 billion investment-grade bond issuance, and is now still preparing to continue borrowing to expand AI infrastructure.

Financing can accelerate expansion, but it also raises return requirements even higher. If the investment return cycle lengthens, companies will have to bear greater interest and cash flow pressure in a high interest rate environment. This is also where the biggest controversy in U.S. stocks is concentrated.

One side sees AI orders, chip demand, and corporate earnings still growing; the other side sees capital expenditures continuing to rise, debt financing increasing, and free cash flow potentially under pressure. The two are not in conflict; they simply focus on different time horizons: the former looks at current performance, while the latter looks at future returns.

In addition, the higher U.S. indices climb, the harder it is to ignore the issue of market breadth. As of October 6, about 44% of S&P 500 constituents were above their 200-day moving average, less than half. The equal-weighted S&P 500 index is still about 4.6% away from its historical high, and small-cap performance is also clearly lagging. In other words, U.S. indices have already hit new highs, but they are increasingly dependent on a small number of companies. Any company experiencing capital expenditure overruns, earnings below expectations, or declining free cash flow could have a greater impact on the entire market.

Market breadth is not only deteriorating. On October 6, 10 of the 11 S&P 500 sectors rose, and advancers on the New York Stock Exchange clearly outnumbered decliners. This shows that funds are beginning to spread beyond tech stocks, but the duration of this change is not yet long enough to confirm that the market has shifted to a broad-based rally.

Global asset performance during the holiday ultimately points to the same question. The current global market is not simply a case of rising risk appetite, nor is it a full-scale retreat of capital. High interest rates are still suppressing valuations, and funds can only seek earnings growth within a limited range. U.S. tech giants and Asia-Pacific semiconductor companies have relatively high AI weightings, so their performance has been relatively strong. Some European indices have different industry structures, so their trends have diverged.

For A-shares, the direction of tech stocks after the holiday has also attracted attention. New highs in U.S. stocks and strength in Japanese and Korean semiconductors reflect that upbeat expectations for the AI industry chain remain; the pullback in European markets, which are dominated by traditional industries, and the persistence of high U.S. Treasury yields also show that global funds have not entered a state of full optimism. After the holiday, what global markets face is still a tug-of-war between growth expectations and high funding costs.

What truly needs to be verified in the fourth quarter is not whether the indices can continue to hit new highs. The more important question is whether technology companies can turn continuously increasing capital expenditures into sufficiently fast revenue, profits, and cash flow, and ultimately have that reflected in their financial statements.