AI Daily Brief: 29 May 2026

29 May 2026

A day defined by extraordinary valuations and expanding AI autonomy. Anthropic has overtaken OpenAI with a near-trillion-dollar funding round, Google launched a 24/7 AI agent at I/O 2026, and Adobe's fresh UK data confirms AI is now outperforming traditional search in retail conversions. On the regulatory front, the UK compliance picture is tightening across criminal law, cyber security, and copyright.

Anthropic raises $65bn at $965bn valuation, surpassing OpenAI as world's most valuable AI company

Anthropic, the AI safety company behind the Claude family of models, announced on Thursday it has raised $65 billion in a Series H funding round, valuing the company at $965 billion post-money. The round makes Anthropic the most valuable private AI company in the world, overtaking OpenAI for the first time.

For UK businesses, the development underscores how quickly the competitive landscape is shifting. Anthropic's Claude models are widely deployed across enterprise use cases in financial services, legal, and professional services - sectors where UK firms have been among the earliest adopters. A $965 billion valuation implies deep investor confidence that AI foundation model companies will capture enormous economic value from the enterprise market over the coming decade.

The fundraise signals a continuing arms race in compute infrastructure. Anthropic has stated it will use the capital to expand computing capacity, meaning more powerful Claude models are expected in the months ahead. With both Anthropic and OpenAI now operating at near-trillion-dollar valuations, the cost of accessing frontier AI is unlikely to fall significantly in the near term.

Our take: Anthropic overtaking OpenAI in valuation is a genuine watershed moment - and a signal that the foundation model market is not consolidating around a single winner. UK businesses betting everything on one AI provider should reconsider. The competitive pressure between Anthropic, OpenAI, and Google is, for now, keeping innovation rapid and pricing competitive. Evaluate your options across all three before locking in long-term contracts.

Google I/O 2026: Gemini Spark launches as a 24/7 AI agent that works while your phone is off

At Google I/O 2026, the company unveiled Gemini Spark, an always-on AI agent that runs continuously in the background using virtual machines on Google Cloud. Unlike a standard assistant that responds to prompts, Spark proactively drafts emails, monitors inboxes, tracks deadlines, watches for hidden credit card fees, and runs multi-step tasks - even when a user's phone or laptop is switched off.

Powered by Gemini 3.5 Flash - Google's new flagship model also announced at I/O - Spark connects to Google Workspace applications including Docs, Gmail, Sheets, and Slides, as well as third-party tools such as Canva and Instacart. Google also announced plans to extend Spark to access local files through the Gemini app on macOS.

Alongside Spark, Google launched Gemini 3.5 Flash as the new default model for the Gemini app and AI Mode in Search, describing it as significantly faster and better at agentic tasks than its predecessor. A new Gemini Omni model family was also announced, capable of generating video from text, photos, video, and audio inputs. The breadth of the I/O announcements signals Google's intent to embed AI agents deeply into professional workflows.

Our take: Gemini Spark is the clearest example yet of AI shifting from a tool you use to a system that acts on your behalf. The 24/7 background operation model - autonomous agents working without direct human prompting - is where enterprise AI is heading. UK businesses need to be thinking now about where they want agents to act independently and where they require a human in the loop. The governance question is no longer theoretical.

AI now converts UK shoppers better than traditional search, Adobe data shows

Shoppers clicking through to UK retail sites from AI sources are now more likely to complete a purchase than those arriving via traditional search engines, according to figures from Adobe Digital Insights. The UK AI shopping conversion rate was up 182% year-on-year in May 2026, and up 543% since January 2025. Adobe has been tracking AI referrals to retail sites since August 2024 and found AI-driven traffic grew 393% in the first quarter of this year compared to the same period in 2025.

The data reflects a structural shift in how consumers discover and evaluate products. Rather than typing keywords into a search engine and scrolling through multiple results pages, shoppers are asking AI a question and receiving a direct recommendation - often with a link to purchase immediately. Some 37% of UK consumers have already used an AI assistant for online shopping, most commonly for research (43%) and product recommendations (40%).

The conversion advantage appears linked to trust: 65% of those using AI said they trusted it to provide accurate results, while 70% said AI assistants were now their primary source of product research. Adobe's Vivek Pandya noted that AI traffic converts better than traditional channels partly because consumers using AI feel more assured about their purchases and return fewer items.

Our take: The 182% year-on-year conversion improvement is the number every UK retailer and e-commerce business needs to see. If your digital presence is not optimised for AI discovery - meaning clear, structured product data that AI tools can surface accurately - you are already losing customers to competitors who are. This is not a future consideration: the conversion gap is open and widening now.

GCHQ chief reveals AI-powered national cyber defence blueprint at Bletchley Park

Anne Keast-Butler, Director of GCHQ, used the agency's inaugural annual lecture to announce that the UK's signals intelligence agency has developed a blueprint for a new national cyber defence capability built on agentic AI. Speaking at Bletchley Park - the wartime home of GCHQ's predecessor - she described the capability as hardwiring cutting-edge agentic AI into machine-speed cyber defence, drawing on decades of machine learning expertise.

Keast-Butler warned that AI is an unstoppable force carrying both opportunity and significant risk, and called on the technology industry and national security community to work together at the speed of the frontier. She pointed to the rapid pace of model releases, increasingly sophisticated agents, and greater system autonomy as transforming the landscape at a rate that makes traditional cyber security approaches insufficient.

The announcement positions the UK government as treating autonomous AI systems as both a defensive tool and a new source of vulnerability. Keast-Butler urged action from boardrooms to living rooms, framing cyber security as an intergenerational responsibility in the AI era. For UK businesses managing sensitive data or operating critical infrastructure, the message is clear: if state-level defence is being rebuilt around AI-speed threats, commercial organisations need to assess whether their own defences are keeping pace.

Our take: GCHQ using agentic AI for machine-speed cyber defence reflects a hard reality: human-speed security operations are no longer adequate when AI can probe for and exploit vulnerabilities faster than any analyst can respond. UK businesses should treat this as a direct prompt to review whether their security tools, vendors, and partners are built for the same pace of threat - not just whether they tick compliance boxes.

UK Crime and Policing Act 2026 creates new criminal AI offences with corporate liability

The Crime and Policing Act 2026, which received Royal Assent on 29 April, has introduced new criminal offences directly targeting AI tools. Under the Act, it is now a criminal offence to make, adapt, or supply a model optimised to produce child sexual abuse material, or to make or supply a tool used as a generator of deepfake intimate imagery - what the legislation terms purported intimate image generators. Both offences apply to corporate bodies as well as individuals, sharply raising the compliance stakes for AI vendors and platforms operating in the UK.

The Act also amends the Online Safety Act 2023 to give ministers powers to regulate AI-generated illegal content and AI services used to commit priority offences. Deepfake intimate imagery offences have been designated as priority offences, meaning regulated platforms must not only remove such content but take proactive steps to prevent it appearing in the first place.

For businesses building or deploying AI products, the corporate liability dimension is the critical compliance shift. Any organisation whose AI product or service could plausibly function as a generator of illegal content now faces potential criminal exposure - not just regulatory sanction. Legal teams will need to assess whether content moderation controls and product safeguards meet the new standard, particularly for generative AI tools deployed in consumer-facing applications.

Our take: Most UK businesses using mainstream AI tools from major providers will not face direct exposure from this Act. However, any organisation building their own generative AI products, running fine-tuned models, or operating platforms where users interact with AI-generated content needs to review their safeguards now. The corporate liability extension makes this a board-level risk question, not a technical one to be delegated to engineering.

Quick Hits

• SK Hynix and Micron both surpassed $1 trillion in market capitalisation within 24 hours of each other, driven by soaring demand for high-bandwidth memory chips used in AI systems - SK Hynix's entire 2026 HBM production is already sold out.
• EU negotiators agreed a provisional Digital Omnibus package delaying high-risk AI system compliance deadlines by up to 16 months, giving organisations more time to prepare documentation, testing, and third-party assessments.
• Google launched Gemini Omni at I/O 2026, a new model family capable of generating video from any combination of text, photos, video, and audio - with Google stating the model will eventually be able to create anything from any input.
• Chancellor Rachel Reeves has told ministers to prioritise British suppliers across four key industries including AI, instructing procurement decisions to weigh domestic origin alongside cost in a push to support home-grown capability.

Frequently Asked Questions

How often is the AI Daily Brief published?

Every morning at 7:30am UK time, covering the previous 24 hours of AI news from over 30 sources.

How are stories selected?

UK-relevant stories are prioritised first, then by business impact and practical implications for UK organisations adopting AI.

Why should business leaders follow AI news?

AI is moving faster than any technology in history. Staying informed is essential for making smart decisions about AI investment, adoption, and governance.