HSBC has reiterated a mild overweight on US equities, arguing that resilient growth and AI leadership keep the case for staying invested intact. The bank frames the recent semiconductor selloff as rotation into software rather than a retreat from the AI trade, and projects AI capital spending to top $1 trillion by 2028.
HSBC has kept its mild overweight position on US equities, saying resilient economic growth, broadening earnings and continued AI leadership leave the case for staying invested largely intact. That call comes even as semiconductor stocks have come under pressure in recent weeks.
Chip weakness read as rotation, not retreat
HSBC analysts describe the move in chip and memory names as rotation rather than a genuine exit from the sector. They say the pullback reflects investors reassessing whether profit growth can keep pace with elevated expectations, not a decision to abandon the AI trade altogether.
The scale of the move has been notable. Samsung, SK Hynix, Intel and Micron have each fallen by roughly a third over the past month. Software stocks, an area HSBC calls sometimes unloved, have meanwhile come back into favor. The Morningstar Global Software-App Index has rebounded around 16% from its June low, and Salesforce, Workday and ServiceNow have all risen sharply over the past week.
Structural AI theme still intact, HSBC says
HSBC continues to view the structural AI theme as intact and maintains positions across Asia's AI ecosystem, spanning power, infrastructure and industrial automation. The bank notes that competition among AI models is intensifying and that pricing pressure is building as the era of subsidised AI access fades, pushing providers toward monetisation approaches such as Model-as-a-Service. In China, HSBC has observed renewed investor preference shifting toward biotech, internet platforms, hyperscalers and electric vehicle makers.
AI spending seen tripling by 2028
HSBC projects AI capital expenditure will rise from below $400 billion in 2025 to surpass $1 trillion by 2028, a trajectory it expects to open new revenue opportunities across the AI supply chain. The bank also expects Asia's data centre capacity to more than double by 2030 to around 40% of global capacity, a build-out it says should support demand for chips, semiconductor equipment, cooling systems, servers, commodities and on-site power generation.
Source: Investinglive
Trading involves risk.
Risk Disclaimer
