Key Points

• Retail investors made a record net purchase of $700 million in neocloud stocks.
• As financing concerns intensify, the yield of CoreWeave bonds soars
• Despite retail investors continuing to buy, institutional positions have shifted towards a defensive stance.

In the latest statistics, retail investors invested a record $700 million in AI cloud infrastructure stocks, despite debt concerns affecting some of these companies, and institutional investors are also becoming increasingly cautious.

This divergence highlights that the rift in artificial intelligence trading is widening: individual investors are buying stocks of companies that build large-scale computing infrastructure, while the credit market demands higher compensation for lending to these highly indebted operators.

According to the Citadel Securities October report, as of the week of October 11th, retail investors made a net purchase of $9.1 billion in U.S. stocks on their platform, which is the highest weekly total in 15 weeks.

Within this total amount, semiconductor stocks attracted $2.6 billion, infrastructure stocks in the broad sense of AI garnered $2.4 billion, and neocloud companies recorded a net purchase of $700 million, which is the largest single-week cumulative purchase of such stocks in the Citadel database.

As institutional sentiment towards AI stocks becomes more cautious, retail investors are increasing their buying efforts.

Retail investors' buying momentum has accelerated for three consecutive weeks, rising from $4.8 billion to $6.4 billion, and ultimately reaching $9.1 billion. However, Citadel indicates that overall trading activity is still below the historical average, suggesting that the increase in buying momentum reflects a stronger belief in buying, rather than abnormally large trading volumes.

The concentrated buying of neocloud stocks is particularly noteworthy. These companies include CoreWeave, Nebius, and IREN, which provide the computational infrastructure necessary for training and running artificial intelligence models, and usually require a significant investment in NVIDIA GPU and data centers.

At the same time, as bond yields rise, inflation persists, and concerns about infrastructure financing related to AI increase, institutional investors have become more defensive.

Business Insider quotes Vanda Research's Viraj Patel as stating that there has been a 'buyer strike' in institutional position indicators, which includes exposures from hedge funds and options activities.

However, this contrast does not mean that individual investors are bullish while institutions are bearish in a completely divided manner. Different institutions adopt different stances, and moreover, the data from Citadel only covers the trading activities visible on their platforms.

The bond market of CoreWeave has sent out different signals.

CoreWeave provides one of the clearest examples of the risks behind the infrastructure boom of AI.

The company's stock price fell by more than 22% that week due to investors questioning whether its rapid expansion could sustain the rising financing costs.

The yield of one bond maturing in 2031 exceeds 16%, and the protection costs against potential defaults have also risen significantly.

Even earlier, the credit default swap of CoreWeave had exceeded 1000 basis points, indicating that investors demanded higher compensation for credit risk.

It should be noted that an increase in bond yields on the secondary market does not automatically lead to higher interest expenses for existing fixed-rate debt. However, for companies that must continuously raise funds for expansion, this may mean a higher refinancing risk.

CoreWeave currently has approximately $104 billion in contracted revenue orders on hold, but future commitments do not equate to immediately available cash. The company still needs to raise funds for its infrastructure before it can collect most of that revenue.

Can retail buying support the infrastructure trading of AI?

The importance of this difference lies in the fact that neocloud relies on two different markets to support its expansion.

Stock investors focus on the expected growth, while lenders assess the debt repayment ability, collateral, and the reliability of future cash flows.

Strong retail demand can support stock prices, but it cannot eliminate the cost of refinancing, nor can it guarantee that continuously expanding data centers will generate sufficient returns.

Coinpaper Previous analysis of AI's financing has shown that these companies are becoming increasingly dependent on borrowed capital to fund the procurement and infrastructure development for GPU.

The latest buying momentum from retail investors indicates that, despite some selling, the market's enthusiasm for AI remains high. Whether this demand can support a sustained rebound will increasingly depend on corporate profitability, the ability to obtain low-cost financing, and whether cloud service providers can convert their substantial infrastructure investments into cash flow.