China tech roundup: artificial intelligence chips, policy and big tech moves

This week’s China tech coverage spans semiconductor policy, big product launches and cloud rivalry. From fresh export limits to new Artificial Intelligence models, here are the key developments.

China’s intensifying technology contest with the United States framed much of the week’s coverage. A feature underscored that the Artificial Intelligence race is increasingly fought by hyperscalers with full-stack capabilities across software, hardware and applications. In Washington, the US Senate passed a measure that limits Nvidia and AMD’s Artificial Intelligence chip exports to China, following a Trump administration deal earlier this summer that had eased some restrictions. Beijing continued to recalibrate critical materials and market oversight: new rare earth export rules will require case-by-case approvals for shipments used in advanced semiconductor design and production, while authorities targeted a consultancy behind teardown reports on Huawei chips after TechInsights claimed Huawei’s Ascend processors used advanced components from TSMC, Samsung and SK Hynix. Separately, China launched an antitrust probe into Qualcomm related to its acquisition of Israeli vehicle semiconductor firm Autotalks.

Semiconductors remained in focus across corporate moves and markets. Intel unveiled a new chip built on its 18A process in what the company cast as a pivotal moment to win back market share and attract foundry customers. TSMC posted better-than-expected sales on robust Artificial Intelligence demand, reinforcing its role as the go-to manufacturer for major accelerator designers. On the mainland, memory maker CXMT edged closer to a domestic initial public offering amid a supply crunch for Artificial Intelligence memory chips, with its listing widely watched as China’s best hope to challenge the dominance of SK Hynix, Samsung and Micron. Investor sentiment was mixed, as another report noted that soaring valuations in China’s chip stocks are testing traders’ patience.

Big tech and cloud players accelerated Artificial Intelligence bets. Google launched Gemini Enterprise to challenge Microsoft and OpenAI in the workplace, pitching a unified and open platform strategy. Chinese fintech giant Ant Group released a powerful Artificial Intelligence model that it says outperforms peers in code generation, software development, competition-level mathematics and logical reasoning. Alibaba signaled deeper ambitions in intelligent hardware as its Qwen lab set up a robotics team, and Alibaba Cloud inked a multi-year collaboration with NBA China. ByteDance founder Zhang Yiming made a rare public appearance in Shanghai to launch the Zhichun Innovation Centre, describing a continued focus on talent recruitment and development.

Robotics and hardware trends also surfaced. Unitree’s humanoid and quadruped robots briefly appeared on Walmart’s website, marking the first time Chinese robots were listed on a major US retail platform and highlighting a widening robotics gap. In autos, Tesla’s Shanghai factory deliveries hit a record for the year as a Model Y variant wooed new buyers. Multimedia reports rounded out the week with a look at China’s trials of underwater data centres aimed at lowering the carbon footprint of digital infrastructure.

68

Impact Score

The missing step between Artificial Intelligence hype and profit

Artificial Intelligence companies have built powerful systems and promised sweeping change, but the path from technical progress to real business value remains unclear. Conflicting studies, weak workplace performance, and poor transparency are leaving a critical gap between hype and evidence.

Samsung workers leaked secrets into ChatGPT

Samsung employees reportedly exposed confidential company information while using ChatGPT for coding help and meeting note generation. The incidents highlight the risk of feeding sensitive data into public Artificial Intelligence tools that retain user inputs.

DeepSeek launches new flagship Artificial Intelligence models

DeepSeek has introduced preview versions of its V4 Flash and V4 Pro models, positioning them as its most powerful open-source Artificial Intelligence platform yet. The release renews competition with OpenAI, Anthropic, and major Chinese rivals while drawing fresh attention to the startup’s technical ambitions and regulatory scrutiny.

OpenAI’s GPT-5.5 sharpens coding but trails Anthropic’s Opus 4.7

OpenAI’s latest model upgrade improves coding, tool use, reasoning and token efficiency as the company pushes deeper into enterprise adoption. Early evaluations suggest stronger security performance, but Anthropic’s Opus 4.7 still leads in some important coding areas.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.