Nvidia and Broadcom are relatively insulated from a worsening US data-center power crunch, but delays in AI deployments could hit makers of memory, optical and other secondary chip components, Morgan Stanley said in a research note published Monday. The brokerage estimated last month that US data-center developers face a 34 percent net power shortfall through 2028, equivalent to 32 gigawatts, even after accounting for mitigation such as behind-the-meter generation and fuel cells. The figure was independently corroborated in separate coverage of the note.
The context is the collision between the AI investment wave and the US power grid. Generative AI’s computing and electricity needs are running up against supply constraints just as hyperscalers and developers have committed hundreds of billions to new capacity. Goldman Sachs has also flagged mounting constraints on the buildout, though it expects limited near-term impact from political pushback, while Morgan Stanley cites labor, power and political challenges together.
Morgan Stanley’s differentiation is specific: it does not see the bottlenecks putting Nvidia’s or Broadcom’s 2027 forecasts at risk, citing their visibility into chip placement, geographic expansion and coordination across data centers, semiconductor suppliers and the power supply chain. The risk instead sits with the companies that ship when racks get energized. If chip capacity cannot be deployed, customers could push out deliveries or cancel orders, with memory, optics, power-management and analog components most exposed to inventory disruption. That framing helps explain Monday’s tape: Nvidia rose 2.12 percent to a record $238.90 and Broadcom gained 2.08 percent, while memory and secondary names lagged.