Trump, Tech Leaders Sign ‘Morally Binding’ AI Safety Accord 
Polymarket Traders Bet on Trump AI Rename to ‘Super Intelligence’ 
Morgan Stanley Launches Digital Asset Lab for Stablecoins and Tokenization 
OpenAI, Anthropic CEOs Summoned to Australia AI Inquiry 
Robinhood Chain Meme Coins Slide as Stock Token Locks Rise 
Apple CEO John Ternus Plans Major Overhaul 
Grayscale Sees Zcash Gaining Market Share Against Bitcoin 
Samsung Invests $1 Billion in KKR’s Helix AI Infrastructure 
AMD Hits $1 Trillion as BofA Raises Price Target to $720 
TSMC Accelerates 2nm Expansion as AI Chip Demand Surges 
QNT Whales Move $10M to Exchanges After 400% Rally 
Aztec Revives zk.money for Private Ethereum Payments 
CSL Partners With AWS to Bring AI Into Drug Research 
Goldman Sachs Private Credit Fund Sees Redemptions Ease in Q3 
China Says U.S. Shares Responsibility for AI Governance 
Binance Pay Goes Live at PayPay Merchants in Japan 
TikTok Reaches $100 Million Alabama Settlement Over Child Safety Claims 
Add as a preferred source on Google
DeepSeek has partnered with Huawei Technologies to develop new programming tools for Huawei’s Ascend AI chips, strengthening China’s efforts to build domestic alternatives to Nvidia hardware.
DeepSeek said Wednesday in a WeChat post that it is developing and open-sourcing a programming backbone designed for Huawei’s Ascend chip lineup. The initiative is intended to improve the software infrastructure available to developers using Huawei AI processors for model development and high-performance computing.
Huawei provided full support for the development of the programming infrastructure, according to DeepSeek. The two Chinese technology companies have also made progress on a “supernode” project built around 128 Ascend 950 chips, highlighting their push toward larger-scale AI computing systems.
The partnership comes only weeks after Huawei unveiled a new generation of Ascend AI processors aimed at computing superclusters and data centers. Huawei expects its AI chips to become widely used for artificial intelligence model training by 2027.
Huawei has emerged as a key player in China’s campaign to develop homegrown AI semiconductors and reduce dependence on Nvidia. Beijing has introduced a range of regulatory and economic measures over the past year to encourage the adoption and development of domestically produced technology.
DeepSeek, considered one of China’s leading frontier AI laboratories, has used hardware from both Nvidia and Huawei to train and operate its artificial intelligence models. The company initially depended heavily on Nvidia processors, particularly the H800, but has increasingly shifted toward Chinese alternatives during 2026.
That transition has become more significant as access to Nvidia AI chips in China remains heavily restricted. DeepSeek’s V4-Flash AI model was optimized to run on Huawei hardware, further demonstrating the company’s growing reliance on domestic computing infrastructure.
The DeepSeek-Huawei collaboration could help strengthen the software ecosystem surrounding Ascend processors, an important factor as Chinese technology companies seek to establish a competitive domestic AI hardware and software stack.
