AI regulation risk will outlast the midterm elections: Raymond James

The message is mixed for AI equities but negative for valuation certainty. If regulatory risk persists beyond the U.S. midterm elections, investors cannot treat the election as a clean policy catalyst. That raises the probability of prolonged uncertainty around data use, model liability, copyright, export controls, energy consumption, and permitting—factors that can delay projects or increase compliance costs.

Near-term implications:

• AI infrastructure remains supported: Data-center developers rushing to deploy capital could sustain demand for GPUs, networking equipment, power systems, cooling, construction, and electrical equipment. Potential beneficiaries include semiconductor, networking, data-center, and grid-infrastructure groups such as Nvidia, AMD, Broadcom, Arista Networks, Vertiv, Eaton, Equinix, and Digital Realty.
• Higher dispersion within technology: Companies selling essential infrastructure may be less exposed than AI application providers whose business models depend on unclear data rights, automated decision liability, or rapidly changing compliance rules.
• Valuation pressure remains a risk: Persistent policy uncertainty tends to weigh most heavily on long-duration, high-multiple stocks because investors apply a larger risk premium to future cash flows. Even strong earnings growth may not fully offset multiple compression.

The capex rush is not unambiguously bullish. It may pull forward orders and support supplier revenues, but it also increases the risk of overbuilding, financing stress, power constraints, and weaker returns on invested capital if regulation, electricity availability, or end-user AI demand limits utilization. Data-center operators and developers therefore face a more asymmetric setup than equipment vendors: near-term construction activity is positive, while longer-term asset economics remain uncertain.

The broader macro effect is also mixed. Accelerated data-center spending supports investment and demand for electricity, equipment, and construction labor, but it can intensify pressure on power markets and infrastructure costs. If regulation slows deployment, growth expectations for the technology-investment cycle could weaken; if regulation mainly raises compliance costs without stopping construction, the effect may be higher operating expenses and lower margins rather than an immediate collapse in demand.

For markets, this is more likely to produce sector rotation and headline volatility than a broad risk-off event. Traders should monitor actual legislative or agency actions, permitting decisions, export-control changes, utility interconnection timelines, hyperscaler capex guidance, data-center vacancy or utilization trends, and whether infrastructure suppliers report cancellations or only longer project timelines. The key distinction will be whether regulation merely increases the cost of AI expansion—or materially reduces the amount of capacity that can be deployed.