AI News

AI News Brief: Taiwan’s tech supply chain continues to benefit from the AI infrastructure boom this year, with several companies seeing sharp market cap gains, but the structural bottleneck of supply-demand imbalance remains unresolved. Unimicron, a key Nvidia and Intel supplier and Taiwan’s largest chip substrate and PCB maker, has seen its stock price climb roughly 350% this year and has already raised its capex plan twice to meet surging demand. However, new chairman Chien Shan-chieh says expanding revenue…

AI news brief: The White House is drafting voluntary standards for frontier AI models, expected to be released as early as next week. This stems from the Trump administration’s recent direct intervention in OpenAI and Anthropic’s new model release pace — the government asked OpenAI’s GPT-5.6 to launch in batches, and once imposed export controls on Anthropic’s latest model citing cybersecurity concerns before lifting the restrictions…

AI news flash: Indian serial entrepreneur Bhavin Turakhia (46) self-funded $30 million to found enterprise AI platform Neo. His core argument: workplace software designed before the AI era cannot be upgraded by simply adding chatbots; it must be rebuilt from scratch. He uses the analogy ‘you can’t build an iPhone using Nokia parts’ to illustrate this point. Neo began internal use in April…

AI News Flash: Japanese pharma giant Takeda and Hong Kong biotech Insilico Medicine partner to grant Takeda access to Insilico’s generative AI platform for screening and identifying potential new drug candidates. The deal could be worth up to $600M per the source URL headline. Insilico’s AI drug development platform has multiple track records, significantly shortening the R&D cycle from target identification to molecular design.

AI News Flash: AI privacy protection platform Venice AI completes new funding round, officially entering unicorn territory; Venice Token rises 6% on the day. CEO Erik Voorhees says the funds will prioritize expanding its own data center infrastructure—directly purchasing GPUs that power the platform, instead of continuing to rent computing power from external cloud providers at high prices. The remaining funds will be used for customer scaling, new market expansion, talent acquisition, and strategic acquisitions that create multiplier effects.

AI News Flash: Ashton Kutcher announces departure from Sound Ventures, co-founded with Guy Oseary, to launch a new VC firm with partner Morgan Beller. Sound Ventures is known for concentrated high-conviction bets, focusing its investment strategy on leading AI labs across categories, with a core style of selecting few targets and deep positioning.

AI News Flash: WSJ reports Musk’s SpaceX showed investors a ‘phone-like’ AI device prototype, described as slimmer and more elongated than an iPhone, with a style between a compact touchscreen phone and Rabbit R1. SpaceX privately demonstrated to investors and stakeholders before public disclosure, noting it’s still early stage and design may evolve. Musk himself directly denied the report, calling it ‘completely untrue’.

Google releases multiple AI updates in June 2026, focusing on seamless integration of AI into everyday hardware and workflows. The open-source Gemma 4 12B model runs locally on laptops with just 16GB memory, using a unified architecture that supports visual understanding and native speech processing, enabling advanced reasoning on consumer-grade hardware while maintaining computational privacy and speed.

AI news: Kawasaki Heavy Industries plans to raise approximately 200 billion yen (about $1.23 billion) through issuing new shares and convertible bonds, funds to be used for Physical AI and other growth areas. Physical AI refers to deploying AI capabilities into tangible hardware like robots and automation equipment, distinct from pure software or cloud AI applications. Kawasaki Heavy Industries has already been actively布局 related fields.

AI news flash: Hugging Face partners with Cerebras, Google DeepMind, and Alibaba to launch a fully open-source real-time voice dialogue pipeline based on WebSocket. The entire system uses a modular design with the following flow: after voice input, Nvidia’s Parakeet model performs speech recognition to convert audio to text; then Cerebras…