Nvidia, Broadcom seen weathering U.S. AI power bottlenecks as secondary chip suppliers face risk
Growing electricity constraints in U.S. data-center markets are emerging as a pressure point for the AI infrastructure buildout. Morgan Stanley says Nvidia and Broadcom remain relatively protected, but delays in deploying chip capacity could ripple through memory, optics and other supporting semiconductor segments.
Highlights
- Nvidia and Broadcom face no material risk to 2027 forecasts from U.S. data-center power constraints, according to Morgan Stanley.
- Morgan Stanley estimates a 34% net U.S. data-center power shortfall—equal to 32 gigawatts—through 2028, risking chip delivery delays or cancellations.
- Secondary suppliers of memory, optical, power-management and analog chips face greater inventory disruption risk as AI infrastructure expansion strains U.S. power systems.
Brokerage view on power constraints and chip exposure
As reported by Reuters, citing Morgan Stanley, Nvidia and Broadcom are not currently seen as facing material risk to their 2027 forecasts from the worsening U.S. data-center power crunch. The brokerage cites the companies' visibility into chip placement, geographic expansion and coordination across data centers, semiconductor suppliers and the power supply chain.Morgan Stanley estimated last month that U.S. data-center developers face a 34% net power shortfall through 2028, equal to 32 gigawatts, after accounting for measures such as behind-the-meter generation and fuel cells. The firm says any inability to deploy available chip capacity could lead customers to delay deliveries or cancel orders.
That leaves memory, optical, power-management and analog chip suppliers more exposed to inventory disruption than the leading AI chip names. The assessment points to uneven risk across the semiconductor supply chain as utilities and developers struggle to keep pace with AI-related electricity demand.
Broader implications for U.S. AI infrastructure
The AI boom is driving billions of dollars in data-center investment, but the computing intensity of generative AI is increasingly colliding with limits in the U.S. power system. That tension is becoming a central constraint on how quickly new AI infrastructure can be brought online.Morgan Stanley and Goldman Sachs have both warned about mounting barriers to data-center expansion in the U.S. Goldman Sachs expects limited near-term impact from political pushback, while Morgan Stanley points to a broader mix of labor, power and political challenges for the sector.
In our earlier article on Nvidia’s march toward a potential $6 trillion valuation, we examined how options markets were pricing in continued upside amid strong AI-driven demand and record highs in major tech indices. We also highlighted that while investor enthusiasm remains elevated, implied volatility and mixed probability distributions signal a more cautious near-term setup even as bullish sentiment persists.