#1Anthropic's Secret "Mythos" Model Leaked — and It's a Cybersecurity Nightmare
A misconfigured content management system exposed ~3,000 unpublished Anthropic assets, including a draft blog post revealing a new frontier model codenamed "Mythos" (internally "Capybara"). Anthropic confirmed it's real, calling it a "step change" over Claude Opus 4.6 with dramatic gains in coding, reasoning, and — notably — cyberattack capability. The company is quietly briefing government officials that Mythos makes large-scale intrusions "significantly more likely."
Source: [Fortune](https://fortune.com/2026/03/26/anthropic-says-testing-mythos-powerful-new-ai-model-after-data-leak-reveals-its-existence-step-change-in-capabilities/) | [Axios](https://www.axios.com/2026/03/29/claude-mythos-anthropic-cyberattack-ai-agents) | [Euronews](https://www.euronews.com/next/2026/03/30/what-is-anthropics-mythos-the-leaked-ai-model-that-poses-unprecedented-cybersecurity-risks)
---
#2Mistral Borrows $830M to Build a Sovereign European AI Supercluster
France's Mistral secured $830 million in debt financing from seven banks — including BNP Paribas, HSBC, and BPIFrance — to build a 44 MW data center near Paris powered by 13,800 Nvidia GB300 GPUs. It's the company's first debt raise and part of an aggressive push toward 200 MW of European compute capacity by end of 2027, with a Swedish site already in the pipeline.
Source: [TechCrunch](https://techcrunch.com/2026/03/30/mistral-ai-raises-830m-in-debt-to-set-up-a-data-center-near-paris/) | [CNBC](https://www.cnbc.com/2026/03/30/mistral-ai-paris-data-center-cluster-debt-financing.html)
---
#3EU Parliament Votes to Delay AI Act Compliance Until December 2027
The European Parliament overwhelmingly voted to push back the EU AI Act's high-risk compliance deadline from August 2026 to December 2027, handing major tech firms — Google, Microsoft, Meta, OpenAI — nearly 18 months of extra runway. The same vote banned AI "nudify" apps outright, signaling regulators will compromise on timelines but not on the most egregious harms.
Source: [PYMNTS](https://www.pymnts.com/cpi-posts/eu-parliament-votes-to-delay-ai-act-compliance-deadlines/) | [JDSupra](https://www.jdsupra.com/legalnews/ai-today-in-5-march-30-2026-the-delay-64700/)
---
#4Super Micro Co-Founder Arrested in $2.5B Nvidia Chip Smuggling Scheme
Federal prosecutors unsealed an indictment charging Super Micro co-founder Yih-Shyan Liaw and two associates with conspiring to route high-performance AI servers through Taiwan and Southeast Asia into China, circumventing U.S. export controls on over $2.5 billion in restricted hardware. A separate case charged three others for a smaller parallel scheme running through Thailand.
Source: [Bloomberg](https://www.bloomberg.com/news/articles/2026-03-19/three-charged-by-us-with-plot-to-illegally-send-ai-tech-to-china) | [Al Jazeera](https://www.aljazeera.com/news/2026/3/20/three-charged-in-the-us-with-smuggling-ai-chips-into-china)
---
#5$100M Political Operation Bets Big on Deregulatory AI Midterm Candidates
A new PAC called Innovation Council Action is preparing to spend more than $100 million in the 2026 midterm elections backing candidates who support a deregulatory AI agenda. The move signals that AI policy has fully entered the electoral arena, with industry players now investing at the scale of traditional corporate lobbying blocs.
Source: [TechStartups](https://techstartups.com/2026/03/30/top-tech-news-today-march-30-2026/)
---
#6AI-Driven Advertising Projected to Hit $57 Billion in 2026 — Up 63%
New market data shows AI-targeted and AI-generated advertising is on track for 63% growth this year, reaching $57 billion globally. The surge is driven by programmatic ad automation, AI creative generation, and personalization at scale — reshaping agency models and media buying in real time.
Source: [AI Business](https://aibusiness.com/generative-ai/what-anthropic-mythos-means-ai-lab-businesses)
---
#7Apple Poaches Google VP to Lead AI Product Marketing
Apple quietly hired a veteran Google VP to head up AI product marketing, the latest signal that the company is moving to sharpen its external narrative around Apple Intelligence after a slow and rocky rollout. The hire follows Apple's lag behind rivals in the generative AI race and suggests a more aggressive go-to-market posture ahead.
Source: [9to5Mac](https://9to5mac.com/2026/03/30/daily-march-30-2026/)
---
**BIG PICTURE**
Today's stories converge on a single uncomfortable truth: AI capability is outpacing every system designed to govern it. Anthropic's Mythos leak crystallizes the dilemma — the most powerful models are now dual-use weapons, and the labs themselves are warning governments even as they race to ship. Meanwhile, Europe blinked on its own flagship regulation, the U.S. is prosecuting chip smugglers while a $100M PAC fights to strip guardrails, and Mistral is sprinting to build European AI sovereignty before the window closes. The infrastructure arms race, the regulatory retreat, and the security threat are no longer separate stories — they're the same story.
