1. Japan Pours Another $4 Billion Into Rapidus In High-Stakes AI Chip Bet
Tokyo has approved ¥631.5 billion (around $4 billion) in fresh subsidies for Rapidus, lifting total public backing for the fledgling foundry to roughly ¥2.6 trillion ($16.3 billion) by the end of fiscal 2026. The money is earmarked to underwrite Rapidus's first commercial contract, a 2-nanometer order from Fujitsu, after an external committee signed off on the company's progress at its Hokkaido fab. Rapidus still aims to hit pilot production late this year and reach a 2031 IPO, but it trails TSMC by several generations and is now also competing for engineering talent with Elon Musk's new Intel-partnered Terafab venture.
Why it matters: Japan is effectively betting the bank that a domestic 2nm foundry is essential national infrastructure, not a private gamble. For Asia-Pacific buyers, this tips the odds that a second credible advanced-node supplier emerges outside Taiwan within the decade, easing the supply-chain concentration risk that boards in Singapore, Seoul and Sydney have been flagging since 2024. It also signals to Southeast Asian governments that very large state capital, not tax incentives alone, is now the price of entry for sovereign AI capacity.
Read more: https://www.taipeitimes.com/News/biz/archives/2026/04/13/2003855475^
2. Stanford's 2026 AI Index Declares The US-China Performance Gap All But Closed
The Stanford HAI 2026 AI Index, published yesterday, finds Chinese and American frontier models have swapped the top slot on leaderboards repeatedly since early 2025, with the best US model ahead by just 2.7 percent as of March. China leads on publication volume, citations, patents and industrial-robot installs, while the US keeps an edge on top-tier flagship models and private investment, which at $285.9 billion last year was 23 times China's reported $12.4 billion. The report credits China's open-source push, led by DeepSeek, Qwen and Zhipu, for the rapid catch-up.
Why it matters: For enterprises across Asia, the takeaway is that dual-stack AI strategies are no longer a hedge, they are the default. Procurement teams in Japan, Korea and ASEAN can credibly plan to run open-weight Chinese models for cost-sensitive workloads alongside US frontier models for regulated use cases, and expect comparable quality. Expect regional cloud providers and systems integrators to accelerate multi-model reference architectures this quarter.
Read more: https://hai.stanford.edu/ai-index/2026-ai-index-report^
3. Hong Kong Opens World Internet Conference Asia-Pacific Summit With AI Governance Front And Centre
The 2026 World Internet Conference Asia-Pacific Summit opened in Hong Kong on Sunday, running two days at the Hong Kong Convention and Exhibition Centre under the theme of digital and intelligent empowerment. Six sub-forums today cover AI agents, AI security and governance, digital finance, digital health, AI for social good, and digital transformation, with IT ministers from Samoa, Madagascar, Turkmenistan and Burundi on the speaker roster alongside the GSMA chief executive and the Chairman of the International AI Governance Association. Organisers plan to release three reports covering AI for public welfare, digital finance and AI governance before the summit closes.
Why it matters: Hong Kong is positioning itself as the neutral convening ground for AI rules in Asia, at a moment when Singapore, Seoul and Tokyo are each pushing their own frameworks. For enterprise buyers and policy teams, the sub-forum outputs and the three reports will shape the next round of procurement and compliance language for cross-border AI deployments in the region, particularly for firms selling into mainland China through Hong Kong entities.