Jefferies analysts argue that the introduction of regulatory guardrails for artificial intelligence could temper the dominance of major technology firms even as such measures may prove advantageous for the broader equity market. In a recent report, the firm suggests that growing public concern about AI makes stricter oversight more likely, and that outcome would change the risk-reward profile across sectors.
The report notes that the recent decline in software and financial stocks largely reflects an assumption that AI will proceed without meaningful regulatory constraints. But as public sentiment increasingly favors safeguards over unfettered innovation, policymakers may be pushed toward rules, disclosure mandates and enforcement actions - developments that could be relatively more punitive for large technology names while easing competitive pressure elsewhere.
Jefferies highlights polling that shows Americans express greater concern about AI than populations in other major economies. According to the analysts, a majority of U.S. respondents prefer regulatory safeguards to rapid, unrestricted development. That preference, the report says, raises the probability of tighter regulation ahead of the U.S. midterm elections.
State legislatures are already taking an active role. The analysts report that more than 1,200 AI-related bills were introduced across U.S. states in 2025, and 874 of those measures remain active or are under consideration. Roughly 40% of the introduced bills focus explicitly on private-sector uses of AI, a pattern Jefferies interprets as evidence of broad and sustained state-level activism.
On infrastructure specifically, Jefferies points out that the five states with the largest data-center footprints have all enacted or proposed some form of restrictions on new builds. The firm concludes that these state-level constraints are likely to materially influence near-term deployment of AI systems and the pace of infrastructure expansion.
At the federal level, the analysts see limited momentum for a comprehensive preemption of state measures. While the report recalls that the Trump administration issued an executive order on AI policy in December 2025, it also notes that a federal AI moratorium is unpopular with many state lawmakers, reducing the likelihood of nationwide moratorium measures advancing.
Jefferies additionally points to congressional action removing a moratorium provision: the Senate voted 99 to 1 to strip moratorium language from the House’s legislative package known as the "One Big Beautiful Bill." The firm frames these developments as indicative of a fragmented policymaking environment in which state-level initiatives will play a substantial role.
Key takeaways
- Potential AI regulations could weigh on large technology companies but may shift gains to other market segments.
- State legislatures are highly active on AI policy, with more than 1,200 bills introduced in 2025 and 874 still active or under consideration.
- Data-center restrictions in the five largest-state footprints may materially affect AI infrastructure buildouts.
Risks and uncertainties
- Public concern might not translate into uniform federal action - uneven policy responses could create localized regulatory risk for infrastructure and cloud providers.
- State-level restrictions on data-center construction could slow or complicate AI deployment timelines, particularly for firms dependent on large-capacity sites.
- Removal of moratorium language at the federal level suggests legislative outcomes remain uncertain and subject to shifting votes and priorities.