Nvidia's Jensen Huang leaves Tokyo with deals spanning Japan's entire tech ecosystem

Title: Global AI race reshapes US-Japan-China partnerships and ambitions

01Nvidia's Jensen Huang leaves Tokyo with deals spanning Japan's entire tech ecosystem

Jensen Huang concluded his Tokyo visit with comprehensive partnerships across Japan's technology sector, signaling deeper US-Japan AI cooperation. The agreements span multiple layers of Japan's tech ecosystem, from established semiconductor firms to emerging AI startups. Nvidia's push into Japan reflects broader geopolitical calculations as the US seeks to strengthen allied nations' AI capabilities against Chinese competition. The deals reportedly include GPU supply commitments, joint research initiatives, and workforce development programs. Industry observers note the timing aligns with increased restrictions on advanced AI chip exports to China, making Japan a critical market for Nvidia's growth. The visit underscores how AI infrastructure has become a cornerstone of US-Japan diplomatic and commercial relations.

02Apple lawsuit threatens OpenAI's hardware ambitions and IPO plans

A legal dispute between Apple and OpenAI is raising questions about the startup's ability to execute on its hardware strategy and planned public offering. The lawsuit centers on intellectual property claims related to technology that became central to Apple's recent product announcements. OpenAI has been vocal about ambitions to move beyond software into dedicated hardware devices, positioning itself as a vertically integrated AI company. However, the legal challenge could complicate partnership discussions and investor confidence ahead of any IPO. Analysts on the Equity podcast debated whether the lawsuit represents a significant obstacle or a negotiating tactic. The outcome could determine whether OpenAI can maintain its valuation and attract the capital needed for hardware development.

03Alibaba releases Qwen 3.8 with 741 Hacker News points

Alibaba's Qwen 3.8 has emerged as a significant contender in the global LLM race, earning 741 points on Hacker News from 524 comments. The open-weight model represents China's latest push to compete directly with Western AI providers in capability and accessibility. Qwen 3.8's performance metrics position it competitively against leading models from Anthropic, Google, and OpenAI on several benchmarks. The release includes token pricing details for commercial API access, signaling Alibaba's intent to monetize enterprise adoption. The strong community reception indicates growing acceptance of Chinese AI models among developers outside China. This launch intensifies competition in the open-weight model space, where Meta's Llama series has dominated Western offerings.

Also today

04
Claude Code quietly switched to Bun written in Rust — Anthropic's coding assistant now runs on a Rust port of Bun, improving Linux startup time by 10%.
05
AI advice makes people 3x more wrong but 2x more confident, study finds — Research reveals that AI-generated advice suppresses critical thinking, leading to more incorrect answers delivered with greater certainty.
06
Christopher Nolan calls AI an obvious 'Trojan horse' — The Oppenheimer director warns that artificial intelligence poses hidden risks despite widespread adoption.
07
Nonprofit Current AI aims to build World Wide Web of AI free for all — New nonprofit initiative seeks to create open AI infrastructure that serves all cultures without commercial barriers.
08
Moonshot AI suspends new Kimi K3 subscriptions due to overwhelming demand — Chinese AI startup pauses new user signups after its latest model release crashes servers within hours.
09
AI mania overwhelming corporate decision-making, consultants warn — Industry analysts report that executives are committing billions to AI strategies without understanding the technology.
10
OpenAI reduces Codex context window from 372k to 272k tokens — Codex code generation model receives significant context reduction in latest update, affecting long-file handling.