Wells Fargo: US equity sentiment triggers a buy signal as the focus of AI investment shifts to capital discipline

What does the sentiment buy signal mean?

As of 16 October, Wells Fargo’s sentiment indicator fell to -1.0 from -0.5 the previous week, triggering its first buy signal since June 2025. The contrarian indicator measures sentiment using data including share prices, options, futures positioning and ETF flows.
In backtests dating to 1998, the S&P 500 rose over the following three months with a probability of 89% after 46 buy signals, with an average gain of 6.5%.
But Wells Fargo notes that returns following the signal are usually weaker when earnings growth is slowing and liquidity is negative; the current liquidity indicator is -0.1. The bank therefore expects US stocks to remain volatile through year-end rather than rise in a straight line.

Wells Fargo believes the near-term support for AI infrastructure lies in 2027: earnings are expected to be strong, while supply and demand will remain very tight. Rising GPU rental prices also indicate constrained computing capacity supply.
The longer-term concern is that the industry could slow after more supply comes online in 2028. Wells Fargo expects semiconductor stocks to trade within a range, with progressively lower rebound highs.
Based on expected 2027 sales and the historical median forward price-to-earnings ratio of 19.3 times, current semiconductor share prices imply a through-cycle net margin of 37%, below the current 46% but still above the pre-AI boom peak of 32%.

Wells Fargo estimates that emerging cloud providers accounted for 19% of AI capital spending over the past 12 months, and that their share will fall to 15% in 2027. They remain an important source of semiconductor demand, but not the whole of it.
Under the bank’s scenarios, a 20% cut in capital spending by emerging cloud providers could reduce semiconductor earnings per share by 3%; a 50% cut could reduce them by 8%. The calculation does not yet account for the potential impact of falling chip prices.

Over the past few years, AI capital spending that exceeded expectations helped drive semiconductor stocks higher; Wells Fargo believes the bullish case going forward will rest on capital discipline. More spending could lead to more debt issuance, crowding out financing for US Treasuries and emerging cloud providers.
The computing capacity price curve shows high near-term prices and lower prices further out. Wells Fargo worries that more supply could intensify this pattern, making investors focus more on the duration of the AI buildout cycle rather than only on tight supply and demand in 2027.
The bank continues to favor large cloud providers and infrastructure software, noting that they recently fell alongside semiconductor stocks and could lead the next bull market.

Wells Fargo’s model estimates that commodity trading advisers (CTAs) net sold 11 billion US dollars of US equities over the past week, and expects CTA flows in US stocks to be tilted to the downside over the next five trading days.
ETF flows paint a different picture: inflows into S&P 500-related equities were 25.3 billion US dollars last week, up 10.8 billion US dollars from the previous week; Nasdaq-related inflows were 17.3 billion US dollars, while Russell 2000-related outflows were 1.2 billion US dollars.