🚀 Alibaba Shocks the World with a Smaller, Faster AI Model!
In the fierce race between tech giants, Alibaba has turned the tables with a new AI model that’s smaller, faster, and cheaper—yet outperforms its …
China is making strides to unify its scattered data centers into a comprehensive national computing network to support artificial intelligence applications, according to data center developers. By the end of June, the country had built more than 10.85 million AI standard bundles, a 30.7% increase compared to last year.
The Open Data Center Committee of China (ODCC) hosted a three-day conference discussing how to convert remote data centers into a unified network, with participation from major investors such as Alibaba, China Telecom, China Mobile, China Unicom, Tencent, and Baidu.
Despite restrictions on access to advanced U.S. chips, Chinese data center operators are innovating to enhance computing performance and share best practices.
Significant progress has been made in improving inter-chip communication within server clusters that use a mix of domestic and foreign semiconductors—critical for AI model training and large-scale data processing.
ODCC has established a laboratory to test communications between locally manufactured chips and network switches, laying the foundation for future development of GPUs and network chips domestically.
Officials emphasized the importance of planning data centers with flexibility and sustainability in mind, considering land, energy, water, and climate resources, including the use of more green energy.
Founded in 2014, ODCC is a nonprofit organization representing about 100 technology companies and infrastructure organizations. Its mission is to promote openness and innovation in servers, networking, and data centers. This initiative comes amid the intensifying AI race between China and the United States, with both countries planning to expand AI data centers and provide the computing resources necessary to train and operate advanced AI models.
In the fierce race between tech giants, Alibaba has turned the tables with a new AI model that’s smaller, faster, and cheaper—yet outperforms its …
Innosilicon’s new Fenghua No.3 GPU, built on RISC-V, targets gaming, AI, and medical use with ray tracing, 112GB+ HBM memory, 8K display support, and …
Blender 5.0 has officially released, bringing a cleaner interface, faster performance, improved asset management, and upgraded shading tools for a smoother, more modern creative …
Apple is reportedly in final talks to acquire talent from Prompt AI, aiming to strengthen its AI capabilities and close the gap with competitors.
A leaked memo claims OpenAI’s CEO urged teams to improve ChatGPT after Google’s new Gemini 3 model surpassed it in key benchmarks.
Qualcomm has officially announced its latest flagship processor for high-end Android smartphones — the Snapdragon 8 Elite Gen 5 — bringing major improvements in …