Artificial Intelligence

Mech-Mind lists in Hong Kong with a blue-chip cornerstone roster, betting standard parts beat the embodied-AI impossible triangle

by

On 1 September 2026, Mech-Mind Robotics, stock code 09615.HK, listed on the Hong Kong exchange. The headline attraction is a top-tier cornerstone roster led by Baillie Gifford, the century-old Scottish asset manager known for long-term bets on Tesla and Nvidia, with Tai Kang Life, Jane Street, BYD’s industrial arm Golden Link, Invus, Ghisallo, Ruihua and … Read more

While Musk’s SpaceX chases space-based data centres, China’s Envision is already running AI power in the Gobi

by

SpaceX’s IPO drew global attention, not only refreshing the record for a first raise at US$75 billion but becoming the world’s fourth-largest listed company on its third trading day, sending Musk past a trillion-dollar fortune. AI data-centre power use is set to roughly double by 2030, nearing Japan’s annual generation. (Source: LeiFengWang) The prospectus stated … Read more

Alibaba’s Qwen 3.8-Max leans into long-horizon agents, betting execution beats raw scale

by

The author had Qwen 3.8-Max build a 3D large-model evolution universe website from scratch and iterated against GPT-5.6. Alibaba’s Qwen 3.8-Max was tested against GPT-5.6 on a 3D website build. (Source: LeiFengWang) On ceiling, Qwen 3.8-Max showed far more delivered detail than GPT-5.6. On efficiency and cost, Max’s high completion came with low efficiency, and … Read more

Moonshot’s secret Hong Kong IPO filing surfaces as China’s large-model race turns into an IPO sprint

by

After Zhipu and MiniMax listed in Hong Kong earlier this year, who becomes the third large-model company to land is still open, but the field of serious contenders is thin. China’s leading large-model startups are now in an IPO sprint. (Source: Sohu IT) New reporting says Moonshot has confidentially submitted an A1 filing to the … Read more

Weifen Zhifei closes several-hundred-million yuan A2 round as capital flows into aerial embodied intelligence

by

On 7 September, Gasgoo’s embodied-intelligence channel learned that Weifen Zhifei, a Hangzhou company, has closed an A2 round of several hundred million yuan, led by Puhua Capital, with Huihui Fund and the Yangtze River Delta Digital-Culture Fund following, and earlier backers Five YSource, Shenzhen Capital, Hongtai, Huaying, Huakong, Changshi and Baidu Ventures adding. Proceeds go … Read more

Dyna-2 abandons the VLA mainstream for a world-action model and lifts robot success from 20 to 90 per cent

by

In August 2026 Dyna Robotics released Dyna-2, a robot foundation model pretrained on over 1 million hours of first-person human video, about 170 years of continuous waking experience. Across seven benchmarks the WAM architecture hit 1.55 times the success rate of the prior VLA model. High-precision manufacturing success rose from Dyna-1’s 20 per cent to … Read more

Delin Capital dangles RMB 2 million cash and up to RMB 5 million more in equity for AI founders

by

The Delin Cup 2026 AI Startup Contest lays out its rewards first. The champion takes RMB 2 million in cash, unconditional and equity-free, a sum that dwarfs the RMB 300,000 to 600,000 typical of domestic AI contests. Up to 15 finalist teams enter a roughly 4,000-square-metre Delin Residency for ten weeks of deep incubation. During … Read more

DeepSeek to deploy 160,000 Huawei Ascend chips for inference as training stays on Nvidia

by

Citing people familiar with the matter, Bloomberg reports that DeepSeek plans to deploy at least 160,000 Huawei Ascend 950DT accelerators in a data centre under construction in Ulanqab, Inner Mongolia, one of the largest Huawei AI chip clusters known. DeepSeek and Huawei have not commented. The purpose is clear. The chips will run already-trained models, … Read more

Kunlun Wanwei’s Matrix-Game 3.5 tackles object permanence and pushes world models toward robot control

by

World models were the soul of this year’s WAIC conference, and the question running through every session was blunt: how does a model actually understand the physical world and step into it? At a dedicated forum on 19 July, Kunlun Wanwei chairman and chief executive Fang Han declared 2026 the first year of the world … Read more

OpenAI’s GPT-6 Astra rollout caps two days of launches that made model rankings obsolete in hours

by

Two days, three frontier launches, and the fastest overtake took five hours. On 2 September Google and Meta each released a new model: Gemini 3.8 Flash, Google’s third Flash iteration in six weeks, and Muse Spark 1.3, Meta’s fourth Spark in five months. Within hours Meta’s model had overtaken Google’s same-day release on the Artificial … Read more