Goldman Sachs trader Matthieu Martal said in recent actions, hedge fund managers were forced to actively diversify their investments, and the extreme concentration of the US market is triggering large-scale theme rotation. He emphasized that the growth prospects of artificial intelligence (AI) are facing increasing challenges, and the risk of an AI capital expenditure bubble is obvious.
Matthieu warned that painful hedge fund positioning resets may last for a while before subsiding. "In addition, given its global impact on managers, there is a factor rotation and a crowd release dynamic that is expanding outside the US, making it worth hedging here."
(Source: ZeroHedge)
He continued, "Finally, undoubtedly, the profitability season of AI is crucial from now on. But good results are no longer enough. The pressure on free cash flow and implementation bottlenecks may disrupt the capital expenditure cycle, leading to significant downside risks."
(Source: ZeroHedge)
He also pointed out that investors are currently in a crowded market, and against the backdrop of cyclical slowdown, the rise of the Russell index seems to be overdone, but the shorts are squeezing, and this dynamic may continue for weeks.
"Although the United States has seen a record decline, ASML indicates that concentration problems also exist outside the United States, and hedge fund demand follows. Worse still, Europe is still a liquidity trap, and US momentum is being reset. Currently, Europe is at a critical level," he continued to explain.
However, this is more positioning-led rotation rather than momentum.
Goldman Sachs said it makes sense to hedge against unnecessary small-cap risks here.
Matthieu wrote that the growth prospects of AI are facing increasing challenges, and the risk of an AI capital expenditure bubble is obvious.
Capital expenditure cycles typically require 4-6 quarters of expansion, followed by 2-4 quarters of digestion.
He warned, "We are in the fifth cycle, which means that pressure may soon come. Moreover, since it is not uncommon for NVIDIA to recall more than 50% of capital expenditures during the rolling period, such as in 2006, 2010, and 2018, once its biggest 'general' is killed, the market may face a historic collapse of the technology bubble."
As building the power grid also takes time, it appears that the market lacks manpower to avoid data center bottlenecks.
How to determine the time of the technology bubble burst? Goldman Sachs trader Rich Privorotsky suggested waiting for technology sales corrections to turn negative, which will be the peak of the bubble.
Therefore, Goldman Sachs recommends profit-taking on AI capital expenditures in GSXEDATA and GSTMTAIP.
Finally, taking all of this into consideration, Goldman Sachs traders reviewed the latest sentiment/positioning snapshots and came to the conclusion. "We are increasingly bearish, with little stock risk premium, bad news is no longer good news, and systematic positioning is overshadowing the stock market."
"Commodity trading advisers (CTAs) are still close to maximum long positions, breaking through the NDX threshold, about 2% higher than the SPX threshold. Given that 70% of the EU's purchases this year come from the United States, there is a real risk of contagion. Hedge funds are bearish on Europe, but not yet. The volatility of information technology is the largest reduction since the COVID-19 pandemic. This may be why sentiment remains calm, with no signs of panic."