CONNECT WITH US
NEWS TAGGED DEEPSEEK
Wednesday 25 June 2025
Huawei's benchmarks show CloudMatrix 384 helps put Ascend chips ahead of Nvidia H800
Huawei has released new benchmark results showing its CloudMatrix 384 AI infrastructure outperforming Nvidia's H800 GPU in running DeepSeek's R1 reasoning model, which contains 671...
Wednesday 18 June 2025
Apple turns to Alibaba for AI push in China, drawing US concern

Alibaba has open-sourced 32 variants of its latest Qwen 3 language models, designed specifically to run on Apple's MLX machine learning...

Friday 13 June 2025
Mistral AI launches Magistral series, boosting multilingual reasoning and open-source offerings
French AI startup Mistral AI has unveiled its new Magistral series models, aiming to elevate logical reasoning performance in enterprise and open-source applications. The lineup includes...
Thursday 12 June 2025
Alibaba staff worked through holiday after DeepSeek breakthrough
Alibaba Group Holding Ltd.'s engineers canceled vacations and worked through the Lunar New Year holiday to play catch up after DeepSeek shocked the global tech industry with its low-cost,...
Monday 9 June 2025
Quantum computing sidelined amid AI hardware boom
While ChatGPT and DeepSeek have already demonstrated the remarkable results that generative artificial intelligence applications can achieve using graphics processing units, the outlook...
Thursday 5 June 2025
Huawei cloud unit claims AI model breakthrough with new training method
Huawei's cloud division said its Pangu large language model achieved a breakthrough in training architecture with a new "Mixture of Group Experts" technology that outperforms competing...
Sunday 1 June 2025
DeepSeek’s stealth upgrade inches closer to OpenAI in code race

Chinese AI firm DeepSeek has quietly released a minor update to its R1 model, uploading the latest version to open-source platform Hugging...

Monday 26 May 2025
Weekly news roundup: Huawei's 5G lead, DeepSeek's LLM moves, and TSMC's packaging gap
These are the most-read DIGITIMES Asia stories from the week of May 19 – May 25. Highlights include China's accelerating push for tech self-sufficiency with new AI...
Monday 26 May 2025
Taiwan researchers step up development of distilled AI models
Taiwan has geared up efforts to develop distilled versions of large language models (LLM) in a bid to accelerate the development and penetration of AI applications.
Friday 23 May 2025
DeepSeek's next move? Wenfeng Liang stays silent on R2, releases V3 study instead
On January 20, 2025, China's DeepSeek unveiled its R1 model, marking a strategic breakthrough in the global race for large language models (LLMs). Spearheaded by founder Wenfeng Liang,...
Thursday 22 May 2025
China’s anti-Altman: Liang Wenfeng and the stealth strategy reshaping AI
As OpenAI CEO Sam Altman toured the Middle East with Donald Trump, promoting a regional "Stargate Project," a quieter but equally profound AI revolution was unfolding in China. At...
Monday 19 May 2025
China's AI appliance chaos: DeepSeek boxes flood market as prices swing 100x between full and distilled models
Since its viral breakout in February, DeepSeek has triggered a rush for all-in-one AI appliances across China's enterprise and government markets. But the boom has also unleashed...
Wednesday 14 May 2025
South Korea's Moreh aims to outpace China's DeepSeek with cost-efficient alternative
South Korean AI software company Moreh is developing "K-DeepSeek," a model designed to surpass China's DeepSeek while slashing development costs, according to TheElec and...
Friday 9 May 2025
WITS initiates acquisition plan to seek second growth curve
Wistron ITS (WITS) chairman and CEO Ching Hsiao pointed out that the model launched by DeepSeek brings a dawn of AI popularization, driving more AI applications. This will be a key...
Friday 9 May 2025
Alibaba's Qwen3 AI models dethrone DeepSeek's R1 in open-source ranking
The Qwen3 family of AI models recently released by Alibaba have taken over the top spot from DeepSeek's R1 model on LiveBench's open-source model ranking.