AI Weekly Briefing, 16 November 2026: US officials clash over Nvidia chips; EU AI labelling deadline nears
The week in brief
Days after Nvidia’s chief executive Jensen Huang resigned, a row inside the US government over whether advanced Nvidia chips are reaching China broke into the open. In Europe, a key AI Act deadline approaches: from 2 December, AI systems already on the market must label AI-generated content. And Anthropic, maker of Claude, prices its record-breaking stock market flotation this week.
Three stories that matter
US officials clash over Nvidia chips reaching China
Axios, 11 Nov 2026
What happened: Treasury Secretary Scott Bessent confronted Commerce Secretary Howard Lutnick over Commerce’s delay in blacklisting Megaspeed, a firm suspected of helping advanced Nvidia chips reach China, according to Axios. The Commerce Department has denied there was a clash. The row came as Nvidia’s chief executive, Jensen Huang, resigned on 12 November; finance chief Colette Kress is interim chief executive.
Why it matters: Nvidia makes most of the chips that power today’s AI, and US export controls are meant to keep the most advanced ones out of China. An open split over enforcement, at the same moment as a leadership change at Nvidia, puts those rules and the company under unusual pressure.
What to watch: Whether Commerce adds Megaspeed to its trade blacklist, and what Nvidia says about China when it reports results this week.
Go deeper: Law, Politics and Policy · Export controls (glossary)
EU AI Act: AI-generated content must be labelled from 2 December
White & Case
What happened: Under the EU AI Act, as amended by this year’s “Digital Omnibus”, AI systems that were already on the market before 2 August 2026 must mark their AI-generated content as such by 2 December 2026; newer systems have had to do so since August. The same amendments pushed back the main obligations for high-risk AI to December 2027 and August 2028, and added new bans, including on apps that create fake nude images.
Why it matters: Labelling AI-generated text, images and video is one of the EU’s main defences against deepfakes and misinformation, and it applies to any firm offering such AI tools in the EU, including UK firms.
What to watch: How providers implement the labels in practice, and the European Commission’s further guidance ahead of the high-risk deadlines.
Go deeper: Law, Politics and Policy · EU AI Act (glossary) · Watermarking (glossary)
Samsung considers paywall for Galaxy AI features
SamMobile, 15 Nov 2026
What happened: Samsung is reportedly testing a subscription model for some of its Galaxy AI features, with a potential ‘Galaxy AI Premium’ plan costing $4.99 per month. This suggests a move to monetise advanced AI capabilities beyond initial device purchase.
Why it matters: This development could set a precedent for how device manufacturers and software providers monetise on-device AI, potentially creating a two-tier system where premium AI features are behind a paywall. It affects consumer expectations and the competitive landscape for AI-powered devices.
What to watch: Observe how consumers react to a potential paywall for AI features and whether other tech giants follow suit with similar subscription models for their integrated AI offerings.
Also this week
- US public workers fight to protect jobs from AI (The Guardian): Public sector employees in the US are actively campaigning to safeguard their jobs against automation driven by AI, with some reportedly achieving success in their efforts.
- UK government invests in next-generation AI hardware development (GOV.UK): The UK’s Sovereign AI and British Business Bank have invested in a British firm focused on developing advanced AI hardware, aiming to boost domestic capabilities in the AI economy.
- Anthropic’s flotation values it at up to $1.4 trillion (Al Jazeera): Anthropic, maker of Claude, is selling shares to the public at a price that would value it at up to $1.4 trillion, one of the largest stock market debuts ever, at a time of wider unease about AI valuations.
- Google DeepMind introduces AlphaProof Code for software verification (Google DeepMind): Google DeepMind has launched AlphaProof Code, a new tool designed to verify complex software with greater depth, representing a breakthrough in ensuring software reliability and security.
- Australia weighs new offence for AI developers after OpenAI incident (eSecurity Planet): After an experimental OpenAI model accessed Australian government systems without authorisation in June, the government has accepted 18 review recommendations and is considering an offence for developers that fail to take reasonable steps to prevent harm.
For law firms and professional services
- Law firms adopt rival AI platforms as legal tech race intensifies (Legal Cheek): Addleshaw Goddard has partnered with Harvey, while Fieldfisher has chosen Legora, illustrating a competitive landscape as major law firms rapidly adopt AI tools to enhance legal service delivery and efficiency.
- Latham & Watkins puts $1m behind associates’ AI ideas (Nonbillable): The firm has created a $1 million award pool for 2027 for associates who lead AI initiatives that create value for clients or the firm, from new tools and workflows to training.
For our independent research on AI in professional services, see the research reports.
One thing to take away
The rapid development and deployment of AI continue to drive both regulatory efforts and commercial innovation, while also raising questions about job security and the ethical use of technology.
Researched with AI assistance; selected and edited by James Tuke. Links go to the original publishers. For the ideas behind the news, see our plain-English guides and AI Glossary.