‘I am the Hoax Buster’: Trump escalates his attack on AI safety warnings

President Donald Trump dismissed concerns about artificial intelligence as a “HOAX” and questioned why AI industry leaders seek regulation. His comments, following remarks in Ireland criticizing efforts to slow AI development, signal a potentially less restrictive US policy stance toward AI. The lack of clarity on specific policy actions leaves regulatory implications uncertain for AI companies.

Analysis

The investable implication is not simply lower regulatory risk for AI leaders; it is a higher probability that compliance costs and deployment approvals become less binding relative to compute and power availability. That favors the AI infrastructure stack—NVDA, VRT, GEV, ETN and data-center REITs with secured power—over application vendors whose valuation already assumes rapid monetization. Over the next 1-3 months, rhetoric alone should not materially alter earnings estimates, but it can reinforce capex commitments if hyperscalers read federal policy as reducing the probability of a near-term national licensing regime.

The second-order risk is regulatory fragmentation rather than deregulation. If federal policy does not preempt state-level rules, large platforms can absorb a patchwork of privacy, safety and labor requirements while smaller model developers face disproportionate legal and distribution costs; this would incrementally strengthen MSFT, GOOGL, AMZN and META versus venture-backed competitors. Conversely, a permissive federal stance could intensify concerns around power prices, water usage, grid interconnection and workforce displacement, shifting political opposition from model regulation toward local data-center permitting over the next 6-18 months.

Consensus may overestimate the immediate benefit to AI software multiples. The binding constraint for incremental AI revenue remains inference economics, enterprise ROI and electricity availability—not policy messaging. The thesis is falsified if hyperscaler capex guidance decelerates, GPU lead times normalize sharply without corresponding inference demand, or state permitting actions delay large data-center projects despite a favorable federal narrative.

Key Decisions for Investors

• Maintain a 3-6 month overweight in AI physical infrastructure via long VRT and GEV versus a short basket of high-multiple, pre-profit AI software names; the relative trade benefits if policy optimism extends capex while software monetization remains uneven. Exit if Microsoft, Amazon, Alphabet and Meta collectively guide 2027 capex lower by more than 10%.
• Use a long MSFT / short smaller AI application-software basket as a 6-12 month regulatory-fragmentation hedge. Large incumbents have distribution, legal resources and cloud control; the key risk is a credible federal preemption framework that materially lowers compliance burdens for challengers.
• Do not add directional NVDA exposure solely on this development. Set an alert around the next hyperscaler earnings cycle: only add if aggregate AI capex guidance rises and management specifically cites faster deployment or reduced approval friction; absent that evidence, this is narrative support rather than an earnings catalyst.
• Monitor PJM, ERCOT and major state data-center permitting actions over the next quarter. Delays or power-cost escalation would favor a tactical reduction in data-center infrastructure exposure even if federal AI policy remains accommodative.