Guizhou’s power system is becoming a testing ground for an AI problem that is easy to describe but difficult to solve: how to coordinate renewable sources that rise and fall on different schedules.
China’s domestic AI chip industry is receiving a powerful demand signal, but the challenge is changing. It is no longer enough for a local accelerator to demonstrate that it can run an existing model at inference.
A small company without an established borrowing record can spend days waiting for a lender to verify basic information. A new Guizhou system is designed to compress that stage into minutes by joining datasets that normally sit in separate public-data systems.
A 21-year-old graduate washing dishes and putting objects away in a Guiyang apartment may not look like part of China’s AI infrastructure. At this year’s data expo in Guiyang, however, that work became a public example of how embodied-intelligence companies are building training material for robots.
A hybrid AI model promises users a choice: deliberate longer when a hard problem demands it, or answer quickly when speed matters more. Tencent researchers argue that the trade-off cannot be measured by latency and accuracy alone.
For much of machine learning’s history, the central assumption has been convenient: training data describe the world a model will later encounter. A research group at Nanjing University has spent years working on the opposite premise.
China’s attempt to turn data into a more usable AI input is moving from national rhetoric into city-level trials. On August 30, the National Data Administration said it would place a new batch of data-annotation pilots in 32 cities, a program intended to direct labeling capacity toward manufacturing, medical care and transportation.
Chinese AI companies have captured eight of the top ten positions on a text-to-video ranking that includes audio, according to a new comparison reported by the South China Morning Post. Alibaba’s Wan 3.0 sits first on the Artificial Analysis leaderboard, ahead of Google’s Gemini Omni Flash.
ChangXin Memory Technologies has taken an unusual route through the U.S.-China technology dispute: a lawsuit in Washington. China’s leading producer of dynamic random-access memory chips, known as CXMT, has sued the U.S. Department of Defense over its designation as a Chinese military company.
Beijing’s HICOOL 2026 entrepreneurship competition closed with 200 winning projects selected from 10,209 applications submitted by 13,472 entrepreneurs. Artificial intelligence was one of the fields represented, alongside biomedicine, new energy, quantum information and high-tech equipment.
Kingsoft Office is testing a more ambitious role for the business spreadsheet. Instead of asking an employee to choose among prebuilt templates or write code, WPS now proposes that a user can describe a need in ordinary language and receive a working business application inside the company’s existing table environment.
Machine translation can render a sentence into another language in seconds, but speed is not the same as political understanding. A Chinese scholar is urging greater human oversight when artificial intelligence translates Beijing’s policy language for overseas audiences, arguing that literal renderings can unintentionally make China’s messaging appear more confrontational than intended.
China’s artificial-intelligence buildout is often measured in chips, model releases and national capacity figures. In Guizhou, it is also visible in a quieter question: what does a data-center strategy change for the places that host it?
China’s national data authorities say the country had reached 2.45 million PFLOPS of intelligent-computing capacity, measured at FP16 precision, by the end of July.
Artificial intelligence is moving quickly into the interpretation of CT, magnetic-resonance and ultrasound images, but China’s newest ethics guidance draws an unambiguous boundary around who remains responsible when the technology is used in research or clinical care. The rules say an AI output cannot serve as the final diagnosis.
China’s embodied-AI investment cycle has often been described through the machines that attract attention on a stage: humanoids, robot dogs, warehouse arms and autonomous vehicles. Songying Technology’s latest financing points to a less visible layer of that stack.
Amap’s latest artificial-intelligence release addresses a problem that becomes acute when machines leave controlled demonstrations and operate for hours in changing spaces: how to keep a usable 3D account of the world without allowing memory use and accumulated localization errors to grow without limit.
MiniMax has sharply increased the cloud-computing capacity it is prepared to buy from Alibaba, providing a concrete measure of how expensive China’s leading AI model race has become. South China Morning Post reported that the Shanghai-based company raised the ceiling of its three-year Alibaba Cloud procurement agreement by 220% to US$1.2 billion through 2028.
Lightstandard is betting that the next bottleneck in artificial intelligence will not be only the number of available accelerators, but the speed and energy cost of moving data between them. The Chinese optical-computing company said it had completed multiple financing rounds during the past year and was approaching a post-money valuation of RMB20 billion.
Greater Bay Area AI Computing Tech has reported a sharp first-half shift from a mixed infrastructure business toward AI computing services. In an HKEX filing released on August 28, the Hong Kong-listed company said revenue for the six months to June 30 reached RMB2.55856 billion, an increase of 1,020.9% from a year earlier.
