AI News Roundup – Trump administration readies executive order to block state AI regulations, Google unveils newest AI model, UAE and Saudi Arabia gain access to advanced Nvidia AI chips, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• • The Trump administration is preparing an executive order to block state-level efforts to regulate AI, according to POLITICO. The order would direct the Department of Justice to challenge state AI laws in court, reviving a Republican effort to block AI laws that they claim to harm AI innovation. Several states, most notably California, have introduced AI regulations in recent months in the absence of federal action on the issue. Federal lawyers would challenge the state laws as unconstitutional regulations of interstate commerce, a power reserved to Congress, or otherwise preempted by existing federal regulations. The order also directs the Secretary of Commerce to review state-level AI laws and restrict federal broadband internet funds to states with “onerous” laws. This plan mirrors a past proposal to condition federal broadband funding to states who complied with an AI regulation moratorium that was stripped out of the GOP’s One Big Beautiful Bill package this past summer. The finalized executive order could be released as soon as this week. 
  • The New York Times reports on the release of Google’s newest AI model, Gemini 3. According to the company, the new model improves on its predecessor in many abilities, including coding, “agentic” operations, and combining images and text in model outputs. According to the Times, the impending release of Gemini 3 caused apprehension at rival companies OpenAI and Anthropic, who worried about the possibility of Gemini outperforming their latest GPT or Claude models, respectively. Google’s Gemini app has over 650 million monthly users, an increase from 350 million in March of this year. Google will make its AI Pro subscription free for U.S. college students. The company will also target business customers, including users of cloud competitors Microsoft and Amazon, by focusing on Google Cloud’s AI features.
  • The Trump administration has approved the export of some advanced AI chips to the United Arab Emirates and Saudi Arabia, according to Bloomberg. The UAE’s G42 and the Saudi Humain will each receive chips equivalent in power to 35,000 GB300 processors, Nvidia’s most advanced models. The companies must comply with “rigorous security and reporting requirements” for the deal to go through, according to the U.S. Department of Commerce, citing concerns over the two countries’ close ties to China. The move comes soon after several deals with Humain were reached by U.S. tech companies, including Elon Musk’s xAI, Amazon Web Services, AMD, and Cisco, further cementing ties between Saudi Arabia and the U.S. as the former seeks to build up its tech industry to lessen its economic dependence on petroleum products. 
  • NBC News reports on the explosion of AI-generated evidence appearing in American courtrooms. In a housing dispute before the Alameda County Superior Court in California, an AI deepfake of a witness speaking was submitted into evidence. Due to the use of AI, Judge Victoria Kolakowski dismissed the case. The rise of AI-generated content has worried judges, especially as realistic AI content could jeopardize the legitimacy of court proceedings. Even before the AI boom, courts have often encountered the “Liar’s Dividend,” where parties cast doubt on authentic evidence as AI-generated to support their case, but the recent Alameda County case was just the opposite. Several advocates concerned about AI in courtrooms have proposed changes to the Federal Rules of Evidence to change the rules surrounding deepfakes, but the U.S. Judicial Conference rejected those proposals in May 2025. One law professor told NBC News that judges should raise their guard against AI, “instead of trust but verify, we should be saying: Don’t trust and verify.” 
  • Microsoft and Nvidia have announced a multibillion dollar partnership with Anthropic, according to the Wall Street Journal. According to the companies, Microsoft and Nvidia will invest up to $15 billion in Anthropic, maker of the Claude AI chatbot, while Anthropic in return will purchase up to $30 billion worth of Microsoft Azure cloud computing capacity running on Nvidia AI hardware. The new deal adds new connections in the ever-growing web of financial ties between major players in the AI industry, as reported by this AI Roundup last month. Microsoft also announced that Anthropic’s AI models will be available in Azure, alongside models from rival OpenAI. The complicated “circular” financing structure between AI companies has raised concern about the possibility of an AI market bubble, which has caused market instability in recent weeks, though there appears to be no indication of a slowdown or pullback in financial and technical cooperation between major tech companies in this space.