Trump Calls AI Danger Warnings a 'Hoax' as Nuclear Deals for AI Power Keep Stacking Up

Since NuScale, Oklo and other reactor developers saw their stocks swing on policy headlines earlier today with only one company actually filing to build, two more storylines have moved: a fresh paper deal on the nuclear side, and a political brawl over whether AI itself is dangerous.

Another Letter of Intent, Not a Reactor

NANO Nuclear Energy (NASDAQ: NNE) announced Friday, September 25, that it signed a Letter of Intent with IP3 Corporation and a company called Cybernetic Intelligence to jointly pursue nuclear-powered AI infrastructure projects, according to a GlobeNewswire release carried by Markets Insider. The LOI names NANO Nuclear as a preferred provider of microreactor technology, fuel supply and reactor operations for projects the three parties hope to develop.

Nothing here is built or funded yet. The release itself describes the arrangement as "expected to include identifying potential end-users and host sites, assessing applicable regulatory and policy frameworks" and evaluating financing. IP3 brings project origination and government engagement; Cybernetic Intelligence brings AI infrastructure and "compute sovereignty" connections. This is a framework to evaluate projects, not a signed reactor order.

Italy Ends Its Chernobyl-Era Ban

A more concrete legislative step came out of Rome. Italy's Senate gave final approval Wednesday to legislation reversing the country's post-Chernobyl nuclear ban, according to the Associated Press as carried by Breitbart. The vote passed 81-51 with seven abstentions, following earlier passage in the Chamber of Deputies.

The law creates a regulatory framework, giving the government 12 months to draft licensing, safety and waste-management rules for next-generation reactors, mainly small modular reactors. It does not authorize a single reactor to be built. Environment Minister Gilberto Pichetto Fratin called it "a historic step forward," saying nuclear "isn't a return to the past" but a technology that, paired with renewables, can strengthen energy security. Italians voted down nuclear power twice, after Chernobyl in 1987 and again after Fukushima in 2011. A June 2026 survey by Only Numbers found roughly 55% of Italians now back next-generation nuclear, per Breitbart, with support strongest among younger voters with no memory of Chernobyl.

Trump vs. the AI Safety Warnings

President Trump has rejected industry warnings about AI risk as a "hoax" and a "conspiracy," according to Fox News. The characterization drew a direct rebuttal from Anthropic cofounder Jack Clark on Bret Baier's program. Clark argued that safety standards are what preserve, not threaten, America's AI lead over China. "The worst time to react to an exponential is when it's too late," he said, comparing proposed safeguards to existing safety standards for children's toys. Clark also denied that Anthropic's push is a financial play tied to a future IPO, calling that theory "an absurd 4D chess move."

The fight was fueled by a Wall Street Journal report, cited by Daily Wire, that an Anthropic researcher named Jacob Coxon quit the company over what he described as "out-of-control" AI fears. Daily Wire's opinion coverage argued that Anthropic, which is reportedly planning a blockbuster IPO that could value the company in the trillions, has little obvious business incentive to scare the public away from its own product, and suggested the episode reads as calculated messaging aimed at inviting government regulation that would lock in Anthropic's position. That is Daily Wire's interpretation of Anthropic's motives, not an established fact, and Anthropic's Jack Clark directly disputed any financial motive on the record.

Researchers at Anthropic and elsewhere have argued publicly that AI systems are scaling in capability faster than tools to understand or control them, and that waiting for a serious incident before regulating carries real risk. This position is widely held inside the AI research community.

Not everyone in the industry wants Washington to slow down. Boom Supersonic CEO Blake Scholl told Fox News that companies, not regulators, should bear responsibility for what they ship. "If someone's concerned about making a product that could do some harm, don't ship the product," he said, though he drew a line at systems that act autonomously: "If a model maker creates a model that goes off and does bad on its own, that is the model maker's responsibility." Former House Speaker Newt Gingrich told Fox's "Kudlow" that policymakers face two separate problems: AI escaping human control and the more immediate strain data centers put on local water and electricity costs. He argued communities need contractual protections against "some trillion-dollar corporation."

What Happens Next

Senate negotiators are considering legislation that would impose a "duty of care" standard on advanced AI developers, according to Fox News, though no bill text, vote date or sponsor list has been finalized. Trump has not said whether he will raise AI safety standards in talks with Chinese President Xi Jinping, something Clark explicitly urged him to do. Whether Congress moves on binding rules before or after Anthropic's IPO, if one proceeds, remains an open question none of the sources answer.