Jack Ma's Ant Group has made notable progress in AI by using Chinese semiconductors from Alibaba and Huawei, integrating them into the Mixture of Experts (MoE) machine learning technique. This development could result in a 20% reduction in computing costs and demonstrates a shift away from reliance on NVIDIA chips, especially given U.S. export controls. The outcome suggests that Chinese companies can effectively compete in AI development without dependency on U.S. technology, highlighting a significant advancement in China's AI sector and potentially lowering costs in real-time AI service delivery.
Ant Group's breakthrough in AI model training using Chinese-made semiconductors could cut computing costs by 20%, marking a significant step in China's AI development strategy.
Integrating Chinese chips in AI training models not only reduces reliance on NVIDIA but also demonstrates how Chinese developers can innovate independently in a tech race.
With the Mixture of Experts method, Ant Group achieved performance comparable to NVIDIA's H800, highlighting the potential of domestic chip alternatives in AI.
Ant Group's reduced inferencing costs could democratize advanced AI capabilities, potentially changing the landscape of artificial intelligence in China.
Collection
[
|
...
]