The rebound in AI stocks is weak: the lack of three major supports is difficult to replicate the July market

According to Woofun AI, JPMorgan Chase (JPM.US)'s pessimistic expectations for the strength of the AI stock rebound in October are in stark contrast to July. Although the stock price has now recovered, it is difficult for the market to recover quickly and strongly. Strategist Nikolaos Panigirtzoglou emphasized that the current decline in investors' willingness to buy on dips has weakened, compounded by reduced corporate repurchases and insufficient short positions, so the lack of basic momentum for this round of decline has been transformed into a strong rebound.

There is an essential difference between this round of adjustments and July in terms of capital and in-depth structure of positions. According to data compiled by Woofun AI, although the stock prices of South Korean memory chip giants SK Hynix (SKHY.US) and Samsung Electronics both fell by nearly one-third in the past month, related stocks have yet to show obvious oversold characteristics, and the decline itself is insufficient to form a basis for a rebound. Market concerns focus on a possible slowdown in the AI capital spending boom and the possibility that lower-priced Chinese rivals could erode growth and profit margins. The recent sell-off spread to the entire industry chain. The Philadelphia Semiconductor Index fell about 10% in the past five trading days, and Intel (INTC.US) declined roughly the same period. Unlike July, ETF investors during the October adjustment did not buy on dips through leveraged and unleveraged ETFs as before; similar purchases rarely occurred. The capital that previously helped the market take on the pressure of the sell-off has yet to be clearly returned. Potential support was further weakened by differences in short positions. In July, the largest Korean stock ETF listed in the US — iShares MSCI Korea ETF (EWY.US) — had a marked increase in short positions, but the fund's short positions did not change much this month.

Meanwhile, short positions on products such as VanEck Semiconductor ETF (SMH.US) and Roundhill Memory Chip ETF (DRAM.US) are also at a low level. As buys formed by successful bears' settlements can sometimes drive a rebound, the current low level means that this portion of potential buying is relatively limited.

The more critical variable is the re-accumulation of leveraged risk. The asset management scale of leveraged storage chip ETFs that Panigirtzoglou is concerned about compared to the total market value of underlying stocks rose again in August and September after falling to a low point at the end of July.

This means that some of the leverage risks released after the previous round of adjustments have now returned to the market, which may amplify fluctuations. Anthropic and OpenAI are still struggling to achieve profits, as well as the high interest rate environment, causing investors to re-examine the growth prospects of AI infrastructure expansion, leading to a lack of bottom-up capital.

Changes in corporate behavior are another key difference. After recording strong profits in the second quarter, SK Hynix and Samsung Electronics announced a total share repurchase plan of $40 billion. These plans were implemented faster than expected and were completed ahead of schedule in mid-October. For the market, buyback means that a company uses its own funds to buy shares to provide relatively continuous demand. When investors reduce their positions and market sentiment weakens, this part of the buying can handle part of the selling order and ease the downward pressure on stock prices.

However, the early completion of the buyback meant that demand dropped earlier than expected. Selling pressure, which was previously absorbed by corporate capital, now needs to be taken over by other investors. Even if the company's business prospects do not deteriorate further, changes in trading power may make it more difficult to stabilize stock prices. At the end of July, the market did not face the same sharp decline in repurchase funds, so although South Korea's memory chip stocks were the first to bear pressure in both rounds of adjustments, there was a clear difference in the financial conditions supporting subsequent repairs. This also explains why J.P. Morgan Chase emphasized that stock prices have fallen, which does not mean that positions and capital flows have been adjusted enough to support a strong rebound.

Under multiple constraints, the AI stock repair path has changed. Judging from buying sources, ETF investors have not clearly bottomed out as in July; short positions are low, which means that demand for reimbursement formed by closing positions with profit is limited; the end of corporate repurchases has also reduced the strength to continue buying.

Meanwhile, the relative size of leveraged memory chip ETFs rose again in August and September, causing the market to face the risk of amplified fluctuations in leverage while its ability to withstand weakening. The combination of these factors made the rebound more dependent on new capital entering the market. Falling stock prices can reduce buying costs, but they cannot automatically create the will to buy. Faced with issues such as a possible slowdown in AI capital expenditure, increased competition, and uncertainties about profit cashing out at high interest rates, investors may need more evidence before they are willing to re-increase their positions. Therefore, J.P. Morgan Chase's judgment mainly points to its ability to recover after this round of adjustments: the purchases of several stocks that drove a rapid recovery in the market in July have either weakened or not yet appeared, and leverage risks have accumulated. This does not rule out a short-term rebound, but it means that it is difficult for the market to expect an equally rapid and equally strong recovery based on previous experience alone. Following this analytical framework, it is worth observing: whether South Korea and semiconductor-related ETFs can re-attract capital, whether leveraged positions are digested, and whether companies launch new repurchase plans. Until these supports improve, this round of AI stock adjustments may take longer to re-establish a sustained buying base.