UBS has reiterated its overweight stance on European pharmaceuticals, framing the sector as a more resilient home for capital than artificial intelligence at a point when the wider market is slowing.
The bank argued that the investment case now rests on more than defensive quality.
It pointed to improving earnings revisions, low relative valuations and lighter ownership after the sector became a source of funds earlier this year.
European pharma has risen more than 5% since mid-June, making it the best-performing sector over that period.
UBS described the shift as the "what if not AI" trade with better timing.
The bank said AI enablers remained attractive but were becoming more two-sided, as data-centre execution and physical bottlenecks could slow the pace of further upgrades.
Weight-loss drugs are the strongest near-term catalyst, according to the note.
The GLP-1 theme, which covers treatments such as Novo Nordisk's (NYSE:NVO) Wegovy, now tops UBS's ranking of investment themes in both the United States and Europe.
The bank flagged the start of Medicare coverage on 1 July as the next test of demand, with patients paying $50 a month and the US government covering the balance.
It rated Novo Nordisk (NYSE:NVO) as a tactical rather than structural opportunity, citing a 2027 US price reset and the group's heavy reliance on semaglutide.
UBS said investors should organise stock selection around the industry's looming patent cliff, which it expects to hit US drugmakers hardest in 2028 to 2030 and European groups in 2030 to 2033.
The bank favours what it calls patent-light compounders, naming Merck KGaA, whose life sciences arm supplies drug manufacturing inputs, and Galderma, the injectable aesthetics group.
Among larger companies, it prefers AstraZeneca PLC (LSE:AZN, NASDAQ:AZN) and Roche, both of which have pipeline readouts due before the end of the year.
UBS was more cautious on Novartis, citing a roughly $26 billion patent cliff from 2029, and on Sanofi, which it said lacked a clear catalyst.
On AI, the bank said the most investable use cases were mundane, covering sales targeting, trial recruitment and regulatory filings rather than drug discovery.