TL;DR

BRICS backs UN-led AI governance; Samsung warns on AI chip delays; Capgemini acquires WNS for $3.3B.

Highlights

• BRICS Rio Declaration calls for AI data safeguards, fair compensation for right-holders, and a UN-centered governance framework, while affirming national sovereignty and interoperable standards 1.
• ByteDance to launch a U.S.-only “M2” TikTok app by September to comply with divest-or-ban law; Oracle cited as a potential buyer, pending U.S. and China regulatory approval 2.
• Capgemini acquires WNS for $3.3B to expand generative and agentic AI business-process services; integration and pricing scrutinized by investors 3.
• Samsung forecasts a 39% profit drop, citing delays in Nvidia Nvidia Corp HBM chip certification and ongoing U.S. export controls; SK Hynix and Micron lead in AI memory shipments 5.
• DeepSeek’s open-source AI model drives Hangzhou startup growth, attracting global pilots and narrowing China-U.S. AI gap; U.S. chip curbs and state involvement remain constraints 7.
• Alphabet’s Alphabet Inc. Class C Capital Stock Isomorphic Labs to begin first human trials of AI-designed cancer drugs, claiming major reductions in development time and cost 6.
• Groq opens its first European AI data center in Helsinki, targeting local inference workloads and compliance with EU data-sovereignty rules 8.
• Huawei denies plagiarism allegations over Pangu Pro LLM, following claims of code reuse from Alibaba’s Qwen 2.5; debate grows over IP and transparency in China’s AI sector 4.
• OpenAI disavows tokenized “shares” listed on Robinhood’s new platform, raising concerns over digital-asset securities and regulatory oversight 10.
• Musk’s xAI Grok chatbot faces criticism for antisemitic outputs after July 4 update, highlighting ongoing LLM content moderation risks 12.
• Nikkei finds hidden AI-praise prompts in 17 arXiv papers, exposing attempts to manipulate automated peer review and prompting calls for stricter academic AI oversight 13.
• Cursor reverses usage-based pricing changes for its AI coding assistant after user backlash, reinstating legacy plans and issuing refunds 15.

Commentary

AI regulatory divergence is intensifying, with the BRICS Rio Declaration advocating for AI data protections, compensation for right-holders, and a UN-centric governance model, while upholding national rulemaking 1. This approach increases compliance complexity for global AI firms and signals a growing bloc influence over international tech standards, especially as new BRICS members expand the group’s reach 1.

The hardware supply chain remains under pressure. Samsung’s profit warning, driven by delays in Nvidia Nvidia Corp HBM chip certification and ongoing U.S. export controls, underscores the fragility of AI memory supply 5. SK Hynix and Micron are outpacing Samsung in shipments to AI data centers, while Groq’s Helsinki data center launch reflects rising demand for regional inference and compliance with EU data-sovereignty requirements 58.

Enterprise AI adoption is accelerating, as shown by Capgemini’s $3.3B acquisition of WNS to deepen generative and agentic AI services 3. In parallel, DeepSeek’s open-source model is catalyzing a surge in Hangzhou startups and global pilot projects, but U.S. chip restrictions and heavy state involvement continue to limit China’s international scaling 7.

AI research and product integrity remain under scrutiny. Huawei’s public denial of plagiarism in its Pangu Pro LLM and Nikkei’s uncovering of hidden prompts in arXiv papers highlight ongoing challenges in IP protection, transparency, and ethical standards 413. Meanwhile, OpenAI’s distancing from unauthorized tokenized shares and xAI’s Grok chatbot controversy reinforce the need for robust content controls and clear regulatory frameworks for both AI outputs and financial instruments 1012.