CoinWorld News: The Guangdong-Hong Kong-Macao Greater Bay Area domestically produced intelligent computing super-ten-thousand-card cluster was launched and put into operation in Shaoguan on July 9, primarily serving the AI industry of the Greater Bay Area, marking the official entry of domestic large model R&D into a new stage of "computing power autonomy." The cluster has a total investment of 5.5 billion yuan, deploys 11,520 Ascend domestic AI accelerator chips, achieves a total computing power of 9,000P, and delivers overall performance at an internationally advanced level. It is part of the national integrated computing power network's Guangdong-Hong Kong-Macao Greater Bay Area hub. By integrating upstream and downstream enterprises such as domestic chips, operating systems, and algorithm frameworks, it improves the complete computing power ecosystem from infrastructure to industrial applications, promoting the large-scale adoption of domestic chips and domestic training frameworks.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 2
  • 2
  • Share
Comment
Add a comment
Add a comment
TheWindBeneathTheCyberBridge
· 57m ago
Domestic computing power has finally stood up. Seeing the 9000P number gives a sense of security—no longer being neck-choked.
View OriginalReply0
VolatilityOfToastingBread
· 1h ago
With 5.5 billion yuan thrown in, 11,520 Ascend cards—this scale is competitive even globally. I’m just curious: how much do training efficiency and NVIDIA differ in reality? Can anyone knowledgeable share their thoughts?
View OriginalReply0
  • Pinned