Flash

The news, the context, and what changes.

RSS
Flash

US AI Inference Chip Unicorn Groq Completes $650 Million Financing

US AI inference chip unicorn Groq announced the completion of a $650 million financing round to accelerate global AI inference cloud expansion, targeting 200 megawatts of computing power by 2027. The company previously reached a massive licensing agreement with Nvidia, with its core team joining the latter. It now focuses on inference cloud operations, running 13 data centers and serving over 5 million developers, while appointing a new leadership team.

88Z AI editorial1 min read
Flash

WeChat Begins Limited Beta Test of Native AI Assistant 'Xiao Wei'

WeChat has launched a limited beta test for its native AI assistant, 'Xiao Wei,' accessible via a green eye icon in the top-left corner of the main interface. The assistant supports text and voice interactions to perform native functions such as adjusting settings, sending messages, launching mini-programs, and generating images. It also features the ability to create mini-programs from a single sentence. The core model is WeChat's self-developed WeLM, with some responses powered by DeepSeek.

88Z AI editorial1 min read
Flash

Zhipu AI's Market Cap on HK Stock Exchange Surpasses 1 Trillion HKD

Zhipu AI's total market capitalization on the Hong Kong stock exchange has surpassed 1 trillion HKD for the first time, rising over 2000% this year to reach 1.27 trillion HKD. Its open-source GLM-5.2 model ranked first among usable models in frontend development evaluations.

88Z AI editorial1 min read
Flash

Tencent QQ Mail Launches Agently Mail, a Dedicated Email Service for AI Agents

Tencent QQ Mail has launched Agently Mail, a standalone email service designed for AI Agents. Users can obtain a dedicated address via WeChat authorization, ensuring isolation from personal inboxes while offering full email management capabilities. The service is now available on Tencent SkillHub.

88Z AI editorial1 min read
Flash

Alibaba HappyHorse 1.1 Video Model Launches on Duiyou

Duiyou platform officially launches Alibaba's HappyHorse 1.1 video generation model. The new version achieves systematic optimization in five dimensions: dynamic expression, consistency, visual quality, instruction following, and camera language. Users can experience it directly via the official website.

88Z AI editorial1 min read
Flash

Claude Code Launches Artifacts, Turning Session Processes into Live Web Pages

Anthropic announced that Claude Code now supports the Artifacts feature, which can real-time generate shareable interactive web pages from session content such as code modifications and troubleshooting, facilitating team progress synchronization. This feature is automatically generated based on full session context, with pages continuously refreshing as work progresses. It is private by default and limited to sharing within the organization. Currently, Artifacts are available in Beta to Team and

88Z AI editorial1 min read
Flash

Alibaba Fully Opens World Model HappyOyster 1.0

Alibaba Cloud launches HappyOyster 1.0, an open-world model enabling real-time interaction. It generates explorable digital worlds via natural language and supports multimodal input with real-time responsiveness.

88Z AI editorial1 min read
Flash

EVOKEN Completes Nearly $300 Million Series B+ Financing

EVOKEN has completed a Series B+ financing round of nearly $300 million, reaching a post-money valuation exceeding $2 billion. The round was co-led by Granite Asia and Shunwei Capital, with continued participation from existing shareholders such as Gaorong Ventures and Ant Group. The company operates three major platforms: LibLibAI, LibTV, and Xingliu. AI video platform LibTV saw revenue grow more than 13-fold in its first two months, contributing to a year-over-year group revenue increase of ov

88Z AI editorial1 min read
Flash

DeepSeek Image Recognition Mode Fully Launched on App and Web

DeepSeek multimodal researcher Xiaokang Chen announced that the DeepSeek image recognition mode is now fully available on both web and app platforms. Alongside fast and expert modes, this feature allows users to upload images for deep understanding by DeepSeek, offering capabilities far beyond simple text extraction.

88Z AI editorial1 min read
Flash

Doubao Real-time Voice Model 3.0 API Service Officially Launched

Volcengine has officially launched the API service for Doubao Real-time Voice Model 3.0 (Seeduplex) and opened it for invited testing. As a native full-duplex end-to-end voice large model, it features precise adherence, interference resistance, and dynamic stop detection. It supports tool invocation during real-time conversations, enabling listening, speaking, and task execution simultaneously. Stop detection latency is reduced by approximately 250ms, and the interruption rate in complex scenari

88Z AI editorial1 min read
Flash

OpenAI Launches New Scheduled Tasks Feature for ChatGPT

OpenAI is gradually rolling out a new scheduled tasks feature for ChatGPT, replacing the existing Pulse functionality. This update aims to improve speed and stability, enabling users to send reminders, handle periodic work, and monitor specific items. The new interface allows unified task management via a dedicated sidebar page, supporting both precise times and broad time windows. Monitoring tasks can perform web searches and read app data, but are limited to running once per hour and may pause

88Z AI editorial2 min read
Flash

SpaceX to Acquire Cursor for $60 Billion

SpaceX has announced an all-stock acquisition of Anysphere, the parent company of AI coding tool Cursor, for $60 billion. The deal is expected to close in Q3 2026, making Cursor a wholly-owned subsidiary of SpaceX.

88Z AI editorial1 min read