CONNECT WITH US
May 20
Alibaba's T-Head doubles down on AI infrastructure with Zhenwu M890
Alibaba Cloud has unveiled the Zhenwu M890, a new proprietary AI training and inference chip developed by its semiconductor subsidiary T-Head Semiconductor, as the Chinese cloud group accelerates its push into full-stack AI infrastructure for the agentic AI era.
Philip Johnston, co-founder and CEO of Starcloud, opened the second day of the Plug and Play Silicon Valley May Summit in Sunnyvale with a proposition that would have sounded implausible three years ago: the most economical place to build AI data centers may soon be in orbit.
South Korea made its case to Silicon Valley investors and startups on the first day of the Plug and Play May Summit in Sunnyvale, positioning Seoul not as an emerging market but as a fully operational launch pad for global technology products.
Nvidia used its first quarter of fiscal year 2027 earnings call on May 20, 2026 to lay out a three-tier silicon cadence that should make any rival roadmap look thin: a Blackwell ramp the company calls the fastest in its history, the first production silicon of Vera Rubin in the second half of this year, and a brand-new Arm CPU, Vera, that opens a US$200 billion TAM Nvidia has never touched.
Nvidia continues to exclude Chinese data center compute revenue from its outlook, with CFO Colette Kress citing uncertainty around whether H200 imports will be allowed into the country despite recent US export license approvals.

OpenAI will commit more than S$300 million (approx. US$235 million) to expand Singapore's artificial intelligence (AI) ecosystem, choosing the city-state for its first applied AI lab outside the US as Singapore pushes to strengthen its position as a regional AI hub.

South Korea has launched a five-year public-private project to build a domestic AI humanoid robot platform, bringing together KIST, LG affiliates, universities, and a hospital as the country looks to narrow the gap with the US and China in next-generation robotics.

Moonshot AI, the Beijing-based generative AI unicorn behind the widely popular Kimi chatbot, has informed its shareholders that it intends to dismantle its offshore corporate framework to clear a regulatory path for an initial public offering (IPO), likely in Hong Kong.

Orient Semiconductor Electronics (OSE) said that strong memory market demand is lifting its outlook and expanding its role in the memory supply chain, while also strengthening its importance in surface-mount technology (SMT) for AI server boards with major US clients.
Iron Force Industrial saw short-term pressure in its operations during the first quarter of 2026, due to tariff swings in the US-China trade war, adjustments to its product mix, and foreign exchange losses from a stronger Chinese Yuan. The automotive safety parts and thermal solutions maker also pointed to new progress in its efforts to enter the AI server cooling market, with liquid-cooling products set to become a new growth driver as shipments begin in the second quarter.
Lens Technology's bid for control of Ju Teng International Holdings is putting renewed focus on changes in the notebook supply chain, as the Chinese supplier seeks to reduce its reliance on Apple and broaden its product portfolio.
Artificial intelligence is undergoing a fundamental shift. Generative AI — passive, prompt-dependent, inert without input — has given way to agentic systems that reason, plan, and act on their own. The change is not incremental. It is architectural.