Wells Fargo: Positioning-driven rotation in US stocks is nearing its end; focus shifts to stock selection

Why does Wells Fargo think the positioning adjustment is nearing its end?

The bank’s sentiment indicator triggered its first contrarian buy signal since April 2025 last week. As of 23 October, the reading was negative 0.9, a slight improvement from negative 1.0 the previous week. Upside skew for semiconductor ETF call options also fell to the 39th percentile over the past 12 months, its lowest since the July low.
Historically, after this buy signal appeared, the S&P 500 rose over the following three months 89% of the time, with an average gain of 6.5%; however, returns were usually weaker when profit growth slowed and liquidity was negative. Since the latest signal, the index has continued to trade in a range, and the bank’s liquidity indicator stands at negative 0.1.

Over the past month, the relative performance of hyperscalers and software stocks versus the S&P 500 became positively correlated with that of semiconductor stocks, unlike in July. Wells Fargo believes the market previously traded AI and technology stocks without sufficient differentiation. It therefore sees opportunities in hyperscalers and infrastructure software, while noting that the market wants to see capital spending discipline.
The bank remains positive on healthcare because it is one of the few sectors where earnings per share growth is expected to accelerate. However, sentiment indicators for pharmaceuticals, biotechnology and life sciences have triggered contrarian sell signals, and software is also approaching that level, making individual stock fundamentals more important.
Within software, the bank’s analysts prefer areas benefiting from AI, such as infrastructure and cybersecurity, to application software. By style, they continue to favor high-quality stocks over high-beta stocks.

Wells Fargo believes that if different parties control the two chambers of Congress after the US midterm elections on 3 November, regulatory pressure on AI could be limited, benefiting AI stocks. The report says that as of 26 October, Kalshi put the probability of Democrats controlling both chambers at 64%; the bank’s scenario analysis considers that outcome potentially unfavorable for AI because of regulatory expectations.
Gubernatorial elections also matter for data center construction: Texas and Georgia account for 27% and 7%, respectively, of planned US data center power capacity. The bank says Democratic candidates in both states are more opposed to data centers than their Republican rivals, and it is paying particular attention to the Texas result.

Wells Fargo credit strategists expect the seven major cloud computing providers to issue 400 billion dollars in bonds in 2027. Separately, the bank estimates that emerging cloud providers face a 80 billion dollar financing gap between operating cash flow and capital expenditure plus dividends.
After accounting for factors including potential IPOs, the bank expects US corporate financing needs over the next year to reach as much as 1.6 trillion dollars, or about 7% of M2, close to the previous cycle’s peak in 2021. The estimate assumes corporate bond issuance remains at the same share of M2 and includes the IPO pipeline and expected financing by cloud providers.

Wells Fargo’s model estimates that commodity trading advisor strategies (CTAs) sold 4 billion dollars of US stocks in the week through 26 October, mainly from the Nasdaq 100 and S&P 500. The model forecasts continued selling pressure on both indices over the next five trading days, while the Dow attracts buying.
In the week through 23 October, ETFs tracking S&P 500 companies saw net inflows of 8.7 billion dollars, down 16.2 billion dollars from the previous week; Nasdaq-related ETFs had net inflows of 6.1 billion dollars, down 11.1 billion dollars week on week; and Russell 2000-related ETFs saw net outflows of 1.2 billion dollars.