AI Daily Brief: 02 April 2026

2 April 2026

A basic DevOps mistake at Anthropic exposed the entire source code of Claude Code to the world this week - 512,000 lines of TypeScript, forked over 41,500 times before anyone could blink. Meanwhile, the CMA is coming for Microsoft, Databricks is betting big on the UK, and OpenAI just closed the largest venture round in history. The money keeps flowing, the regulators keep circling, and the cracks in AI security keep showing.

Anthropic Accidentally Exposes Entire Claude Code Source - 512,000 Lines on npm

Anthropic shipped a sourcemap file inside its official npm package for Claude Code (v2.1.88), inadvertently exposing the full source code of its flagship AI coding tool. The .map file, a debugging artifact that should have been excluded via .npmignore, contained references to a zip archive on Anthropic's Cloudflare R2 storage holding approximately 1,900 TypeScript files and 512,000 lines of unobfuscated code.

Security researcher Chaofan Shou spotted the exposure on 31 March. Within hours, the code was mirrored to GitHub, where a clean-room Rust rewrite hit 50,000 stars in two hours - likely the fastest-growing repository in GitHub history. The repository has been forked over 41,500 times. Claude Code is built on Bun, which Anthropic acquired in late 2025. Bun generates sourcemaps by default, and the release team failed to exclude them from the published package.

The leak is not a hack or a sophisticated exploit. It is a basic DevOps mistake in the npm publishing pipeline - the kind of error that would fail a junior developer's code review. For UK businesses evaluating AI vendors, it is a stark reminder that even the most well-funded AI companies can stumble on fundamental software supply chain hygiene.

Our take: This is embarrassing for Anthropic, but the real story is what it reveals about the state of AI tooling security. If a company valued at tens of billions cannot get .npmignore right, what does that say about the supply chain security of the AI tools your business depends on? Every organisation using npm packages - AI or otherwise - should be auditing what gets published. The exposed code itself is fascinating (full system prompts, tool implementations, permission models), but the lesson for business leaders is simpler: trust, but verify. Audit your AI vendor's security practices the same way you would audit any other critical supplier.

CMA to Launch Investigation into Microsoft's Business Software Ecosystem

The UK's Competition and Markets Authority has announced it will launch a Strategic Market Status investigation into Microsoft in May 2026. The probe targets Microsoft's use of software licensing to restrict cloud competition, a concern that has grown more urgent as AI tools like Copilot become embedded in core productivity suites used by hundreds of thousands of UK businesses and public sector organisations.

CMA Chief Executive Sarah Cardell said the regulator is "not just responding to today's concerns but getting ahead of emerging issues too." Microsoft and Amazon have also committed to actions on cloud egress fees and interoperability following CMA engagement, though the regulator says further steps are needed.

For UK businesses that rely on Microsoft 365 and are evaluating AI add-ons, this investigation could reshape pricing, bundling, and the ability to mix-and-match AI tools from competing providers.

Our take: This is the most consequential UK AI regulatory action in months. If the CMA secures an SMS designation, it gains real enforcement power over how Microsoft bundles Copilot with Office. For UK businesses currently locked into the Microsoft ecosystem, this could open the door to genuine choice in AI productivity tools. Watch this space closely.

Databricks Commits $850m to UK with New London EMEA Headquarters

Data and AI firm Databricks has announced plans to invest more than $850 million in the UK over three years. The company will quadruple its London office footprint with a new 137,000 sq ft EMEA headquarters in Fitzrovia, growing its UK team from 500 to over 1,000 employees.

Databricks already counts more than 50% of the FTSE 100 as customers, including Unilever, Rolls Royce, Nationwide, and Octopus Energy. The firm also plans to train 100,000 people across the UK and Ireland in data and AI skills by 2028, with partnerships at the London School of Economics and University College Dublin.

Our take: Major AI companies are backing the UK with real money, not just press releases. Databricks choosing London as its EMEA hub signals confidence in the UK's AI talent pool and regulatory environment. The training commitment of 100,000 people is equally significant - the skills gap remains the biggest barrier to AI adoption for most UK businesses.

AI Set to Reshape £3.7bn of UK Retail Marketing Spend by 2030

New research from retail technology firm Voyado and Retail Economics estimates that £3.7 billion of UK retail marketing and e-commerce spend could be disrupted by artificial intelligence by 2030. The study highlights how AI is transforming customer acquisition, personalisation, and campaign optimisation across the sector.

For UK retailers already operating on thin margins, this represents both a competitive threat and an opportunity. Businesses that adopt AI-driven marketing early stand to capture disproportionate value, while those that delay risk falling behind.

Our take: £3.7bn is a serious number. UK retailers are already under pressure from cost-of-living impacts and shifting consumer behaviour. AI is not going to wait for them to catch up. The businesses that start experimenting now, even with modest budgets, will be the ones still standing in 2030.

OpenAI Raises $122bn in Record Funding Round at $852bn Valuation

OpenAI has closed the largest venture funding round ever recorded, raising $122 billion at a valuation of $852 billion. SoftBank co-led the round with Andreessen Horowitz, D.E. Shaw Ventures, MGX, TPG, and T. Rowe Price, with participation from Amazon, Nvidia, and Microsoft. Around $3 billion came from individual investors.

SoftBank separately secured a $40 billion bridge loan partly to fund its OpenAI commitment. OpenAI said the capital will support its continued investment in compute and infrastructure, with an IPO expected later this year.

Our take: The numbers are staggering, but they also highlight the immense capital requirements of frontier AI. OpenAI is valued higher than most countries' GDP, yet it is still burning cash at an extraordinary rate. For UK businesses, the practical takeaway is that the tools built with this funding will reach you - the question is whether you are ready to use them effectively.

Oracle Cuts Up to 30,000 Jobs to Fund AI Data Centre Buildout

Oracle has begun laying off thousands of employees in what analysts estimate could affect up to 30,000 people, or 18% of its 162,000-person workforce. TD Cowen estimates the cuts will free up $8-10 billion in cash flow to fund the company's aggressive AI data centre expansion.

Oracle's stock is down 25% this year as investors question whether the company can compete with larger cloud rivals like Amazon while taking on significant debt for AI infrastructure. The company raised $50 billion in debt and equity in January and executives said during the last earnings call that no further debt would be raised in 2026.

Our take: This is the human cost of the AI gold rush laid bare. Oracle is cutting nearly one in five employees to fund infrastructure that may or may not pay off. It is a pattern we are seeing across big tech: jobs displaced not by AI doing the work, but by the sheer capital required to build AI capacity. UK businesses should take note - AI transformation is expensive, and the costs are not always where you expect them.

OpenAI Brings Conversational Ads to ChatGPT via Smartly Partnership

Six weeks after launching static advertisements inside ChatGPT for free-tier users, OpenAI is escalating its ad strategy. The company has partnered with Helsinki-based Smartly to build conversational, interactive ad units that respond to users in turn, rather than sitting passively beside a response.

The initial ad pilot has already crossed $100 million in annualised revenue with over 600 advertisers, reaching fewer than a fifth of eligible users. Criteo is serving as OpenAI's first formal ad-tech partner, connecting 17,000 advertisers to ChatGPT's inventory. Anthropic, maker of Claude, has positioned itself as the ad-free alternative.

Our take: This is a pivotal moment for the AI assistant market. OpenAI is betting users will tolerate ads that talk back to them inside a tool they rely on for work and research. Anthropic is making the opposite bet. For businesses using ChatGPT as a work tool, the question is whether you want your AI assistant to also be a salesperson. The split between ad-funded and subscription-only AI is now real.

Baidu Robotaxis Freeze Mid-Ride in Wuhan, Trapping Over 100 Passengers

More than 100 Baidu Apollo Go robotaxis stopped mid-traffic in the Chinese city of Wuhan due to a system-wide failure, trapping passengers inside vehicles and causing highway collisions. Videos shared on social media showed vehicles halted in moving traffic with passengers unable to exit safely.

Wuhan police confirmed the incident was caused by a "system failure" and said services have since resumed. The incident has reignited safety concerns about autonomous vehicle deployment at scale, particularly in densely populated urban areas.

Our take: This is exactly the kind of incident that can set an entire industry back. When one system failure can immobilise an entire fleet simultaneously, the risk profile is fundamentally different from individual car breakdowns. UK policymakers drafting autonomous vehicle regulations should study this closely. The technology works until it does not, and when it fails, it fails everywhere at once.

Q1 2026 Venture Funding Hits $300bn All-Time Record, AI Takes 80%

Crunchbase data shows investors poured $300 billion into 6,000 startups globally in Q1 2026, up more than 150% quarter-on-quarter and year-on-year. That single quarter equals nearly 70% of all venture capital deployed across the whole of 2025.

AI dominated the numbers, accounting for $242 billion or 80% of total funding. Four of the five largest venture rounds ever were closed this quarter: OpenAI ($122bn), Anthropic ($30bn), xAI ($20bn), and Waymo ($16bn). These four deals alone represented 65% of all global venture investment in the period.

Our take: These figures are almost beyond comprehension. The concentration of capital in a handful of frontier AI companies is unprecedented in the history of technology investment. For UK businesses, the practical impact is clear: the tools, infrastructure, and capabilities being built with this money will reshape every sector. The question is not whether to engage with AI, but how quickly you can build the internal capability to use it.

Quick Hits

• Intel plans to invest another $15m in SambaNova, the AI chip startup chaired by Intel CEO Lip-Bu Tan, raising governance questions about the relationship.
• India's Wipro has named Nagendra Bandaru as CEO of its newly created AI segment, while Americas 2 unit head Suzanne Dann has resigned.
• EY research finds 96% of enterprises use AI for cybersecurity, but 85% say budgets are underfunded relative to AI-powered threats they face.

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.