ComputeLabs Research

AI & Compute Infrastructure — September 20, 2026

Edition of · 15 stories

Richinfo (彩讯股份) signed an effective compute-services contract, expecting at least RMB 1.812 billion in tax-inclusive fees; orders run 60 months post-acceptance.

Richinfo (彩讯股份) announced that it signed a Compute Technology Services Agreement (算力技术服务协议) with a customer identified only as Company H on September 20, 2026. According to the announcement summarized by Financial_Express, all conditions for effectiveness had been satisfied by the disclosure date, and the contract was formally effective.

Read the full story · 1 source →

ChangXin Memory Technologies (长鑫科技; SSE: 688825) began fifth-generation process mass production; associated products have not yet achieved scaled sales.

ChangXin Memory Technologies (长鑫科技), identified in the supplied announcement summary as 688825.SH on the Shanghai Stock Exchange (SSE), announced fifth-generation process-platform mass production at the World Manufacturing Convention on September 20. It simultaneously displayed high-capacity LPDDR5X, or low-power double-data-rate 5X memory, built on that platform.

Read the full story · 2 sources →

Hygon (海光信息) plans to launch 1000-series CPUs on September 22, targeting low-power, embedded, and edge-computing applications.

Hygon (海光信息) announced through its corporate WeChat account on September 20 that it would introduce new chips in Shenzhen (深圳) on September 22. Reporting attributed to 21st Century Business Herald (21世纪经济报道) identified the products as 1000-series central processing units (CPUs) for low-power, embedded, and edge-computing requirements.

Read the full story · 2 sources →

Beijing (北京) published a 2026–2028 AI-token-economy plan covering inference-chip development, model–chip optimization, and evaluation standards for token-production infrastructure.

Beijing’s municipal economic and information technology bureau and development and reform commission issued the Beijing Action Plan for Accelerating Token-Economy Development, 2026–2028 (北京市加快词元经济发展的行动方案(2026—2028年)). The document covers both token-production infrastructure and downstream AI-agent applications.

Read the full story · 2 sources →

Qwen (千问) open-sourced Qwen-Image-2.1, unifying generation and editing with a 7-billion-parameter visual-generation component and native transparent-image support.

The Qwen (千问) team at Alibaba (阿里巴巴) announced the open-source release of Qwen-Image-2.1 on September 20. The model combines text-to-image generation and image editing, with 7 billion parameters specifically in its visual-generation component, rather than a stated total parameter count for every component.

Read the full story · 2 sources →

All ComputeLabs Research editions