Jensen Huang on his relationship with Trump: ‘calls me in the middle of the night’

Nvidia CEO Jensen Huang argued AI adoption will improve productivity, create jobs, and support U.S. competitiveness, while warning that energy shortages and unclear export controls could constrain growth. He backed a need for some government regulation and national-security focus, but said government ownership of AI companies is unnecessary because Americans already benefit through stock ownership, taxes, and jobs. The comments reinforce bullish long-term AI demand, with policy and power-supply constraints remaining the key risks.

Analysis

The core equity implication is not simply “AI is good for semis,” but that the next leg of AI capex is becoming more politically and physically constrained. If policymakers lean into national security screening while simultaneously treating power as the bottleneck, the market should increasingly reward vendors that monetize the constraint itself: interconnect, optics, thermal management, power conversion, and grid-adjacent infrastructure. That makes NVDA less of a pure compute story and more of a systems-orchestration story, while COHR becomes a levered beneficiary because efficiency improvements are one of the few ways to keep capex ROI intact as power availability tightens.

The second-order effect is that energy scarcity can become a gatekeeper on AI deployment rather than just a cost line. That favors companies selling “more compute per watt” and penalizes marginal data-center builds in constrained grids, especially where local politics can delay permits or rate hikes. Over 6-18 months, the market may start discounting a bifurcation between elite AI infra winners and everyone else in the ecosystem that needs cheap, reliable power but lacks it.

Contrarianly, the consensus may be underestimating how much regulation accelerates concentration. Clearer export-control and model-screening rules tend to entrench the largest, most compliant incumbents and raise the barrier to entry for smaller challengers; that is structurally positive for NVDA and for critical suppliers with de facto standard-setting power. The risk is that if the U.S. energy buildout fails to keep pace, sentiment can flip from “AI capex supercycle” to “AI bottleneck overspend” within one or two quarters, compressing multiples for the whole complex even if end-demand remains intact.

Key Decisions for Investors

• Add to NVDA on any 3-5% pullback; view as a 6-12 month leadership name as regulation raises moats, but size with a tighter stop if energy/power headlines turn negative and capex sentiment rolls over.
• Initiate a tactical long COHR vs short a basket of lower-quality AI infrastructure names over the next 1-3 months; COHR is a direct beneficiary of power-efficiency spend and should outpace the broader AI hardware complex if grid constraints intensify.
• Buy COHR call spreads 3-6 months out to express upside on power-efficiency adoption with defined risk; use a pullback entry rather than chasing strength after the event-driven move.
• Pair long NVDA / short a basket of high-beta data-center proxies that are more exposed to local permitting and power costs; thesis is that regulation and energy scarcity increase share concentration at the top of the stack.
• If AI policy headlines become more restrictive, take profits on crowded secondary beneficiaries and rotate capital into the picks-and-shovels names with pricing power and lower execution risk.