Semiconductor stocks face volatility amid political and leveraged pressures, but AI compute demand remains strong over the long term.
The semiconductor sector is experiencing significant volatility, with political challenges, company guidance, and leveraged trading positions contributing to a sharp selloff. Despite these pressures, the long-term demand for AI compute remains robust, according to an analysis by VanEck published on November 19, 2026.
On November 5, the VanEck Semiconductor ETF (SMH) dropped 11.6%, marking its worst single-day performance since March 2020. This decline followed the approval of restrictions on 40 out of 41 local data center ballot measures during the U.S. midterm elections. UBS estimates that these restrictions could jeopardize approximately 30% of the anticipated U.S. data center additions for 2028. However, most of the 2027 buildout is already underway.
Additional factors exacerbating the selloff included weaker-than-expected guidance from Qualcomm and Arm, Micron’s decision to pause a planned Taiwan plant, and the unwinding of leveraged positions. As of November 13, SMH was trading 27% below its June 2026 high but remained up 35% year-to-date.
4 GW by 2030, a fivefold increase. The report highlights that inference workloads will expand from 30% to 81% of total compute demand over this period. While political and power constraints are expected to slow the pace of new data center approvals, they are unlikely to diminish overall demand. Instead, companies may focus on upgrading existing facilities with chips that maximize compute per megawatt.
Source link







