Mistral Secures $830M Debt Deal for Nvidia-Powered European Data Center OpenAI Shuts Down Sora, Redirects Compute to Hig

Published: 2026-03-30 06:00 (Asia/Shanghai) Coverage: 2026-03-29 18:00 — 2026-03-30 06:00

📰 Top Stories

1. Mistral Secures $830M Debt Deal for Nvidia-Powered European Data Center

France’s Mistral AI has raised $830 million in debt financing from a seven-bank consortium to purchase 13,800 Nvidia GPUs and build a major data center near Paris. The move marks Mistral’s first major debt raise and signals Europe’s push toward AI compute sovereignty, with expansion plans targeting Sweden and 200 megawatts of compute capacity across Europe by end of 2027. Lenders now view AI infrastructure as bankable rather than purely venture-backed speculation.

2. OpenAI Shuts Down Sora, Redirects Compute to Higher-Priority Products

OpenAI has officially shut down its Sora video generation product, which had been run separately from the company’s main research effort. The decision reflects a strategic reallocation of scarce GPU capacity toward models and products with stronger traction. Sora’s closure underscores how AI product lines compete internally for compute resources — not just users — and that even high-profile consumer features can be cut when they no longer justify their economic cost.

3. Anthropic Confirms Leaked Reasoning Model — Claude Mythos Tier Emerges

Anthropic confirmed that a leaked model identified as “Claude Mythos” is real, describing it as a new tier above Opus with dramatically higher scores on software coding, academic reasoning, and cybersecurity benchmarks. The confirmation came after leaked documents surfaced online, revealing details about the unreleased model’s capabilities. This marks the first time Anthropic has publicly acknowledged a model tier beyond Opus, raising questions about the company’s release strategy for its most powerful systems.

4. DeepSeek Suffers 7-Hour Outage — Longest Since Its Breakout Rise

China’s DeepSeek chatbot experienced a major outage lasting over seven hours, its longest disruption since the service’s viral 2025 debut. The downtime affected both consumer users and developers awaiting the company’s next model update. For a lab that has built its reputation on competing directly with Western AI players, the outage highlights the growing expectation that leading AI services must deliver reliability comparable to established cloud infrastructure, not just benchmark-leading performance.

5. Starcloud Raises $170M for Orbital AI Data Center Constellation

Orbital compute startup Starcloud closed a $170 million funding round at a $1.1 billion valuation to develop an 88,000-satellite constellation powered by continuous solar energy in space. The company has previously worked with Nvidia hardware and AWS-related technology. The bet reflects a growing conviction that AI’s next bottleneck is energy and cooling — not chips alone — and that terrestrial data centers may face physical constraints that push compute workloads into orbit.

6. Sakana AI Releases AI Scientist-v2 for Automated Scientific Discovery

Sakana AI unveiled AI Scientist-v2, leveraging Agent Tree Search to automate end-to-end scientific research at workshop-paper level. The system navigates complex research spaces autonomously, generating novel scientific insights without human intervention. This release represents a significant step forward in AI-driven research automation, building on the company’s earlier work in evolutionary model composition and demonstrating how agentic frameworks can be applied to the scientific discovery pipeline.

7. Apple Pivots AI Strategy Toward Hardware-First, On-Device Intelligence

Apple is reshaping its AI approach to emphasize deeper integration with its hardware ecosystem, including new Siri extensions allowing third-party app integrations and a renewed focus on on-device processing for privacy and performance. Rather than competing head-on with cloud-heavy AI models, Apple is betting on seamless hardware-software synergy and ecosystem lock-in to differentiate its AI offerings — a strategy that could pressure rivals to innovate beyond software-only approaches and drive device upgrade cycles.

📊 Trend Watch

🔮 What to Watch

• Anthropic Claude Mythos Release Timeline: With the model confirmed and benchmarks leaked, the market is watching for Anthropic’s official launch strategy — particularly whether safety protocols will limit availability or whether it will ship alongside the Opus line.
• European AI Sovereignty Buildout: Mistral’s Paris data center is the first concrete step; watch for follow-on announcements from other European labs (Aleph Alpha, DeepMind Paris) as the EU moves from policy rhetoric to physical compute capacity.
• OpenAI Post-Sora Product Roadmap: After reallocating Sora’s compute, expect announcements on what higher-priority products benefit — likely tied to GPT-5.4 variants and agentic capabilities that OpenAI has been quietly developing.

Briefing generated: 2026-03-30 06:00 (Asia/Shanghai) Data sources: AI-curated from public technology reports and industry analysis