Jack Ma-backed Ant Group cuts AI training costs using Chinese chips | Technology News

Chinese fintech giant Ant Group, backed Alibaba founder Jack Ma, has achieved major success in reducing AI model training costs using Chinese chips made Alibaba and Huawei. The company has developed methods that cut AI training costs 20 per cent, Bloomberg and CNBC reported Tuesday, citing sources.
Many are calling it a breakthrough, as the combination of chips not only reduces the time and cost of training AI models but also limits Chinese tech companies’ reliance on American suppliers such as Nvidia.

While Ant Group has used chips from Alibaba and Huawei for training AI models, it has also used Nvidia chips. In fact, the Chinese-made chips performed about as well as Nvidia’s H800 chips, Bloomberg sources claim. Though Ant Group continues to utilise Nvidia’s hardware for certain AI development tasks, the company is relying on alternatives—particularly chips from AMD and Chinese manufacturers—for its latest models.
Story continues below this ad

Earlier this year, a small Chinese AI startup, DeepSeek, stunned the world achieving a technical breakthrough with a large language model that rivals AI models from OpenAI and Google—without spending billions to train it. It was a major milestone for Chinese tech companies, bringing China closer to the US in AI supremacy. Chinese artificial intelligence firms are driving down costs to create competitive models as they grapple with US chip restrictions and significantly smaller budgets than their Western counterparts.
The US has sought to slow China’s AI development limiting access to the most advanced semiconductors, such as those from Nvidia, used for training models. However, Nvidia can still sell its lower-end chips to China.

DeepSeek’s advancement have sparked a nationwide push in China to deploy its large language models everywhere from hospitals to local governments
Ant Group recently announced “major upgrades” to its AI solutions for healthcare, with a new healthcare-specific AI model. The large model is integrated into “All-in-One Large Model Machines for Healthcare.” Seven hospitals in cities such as Beijing, Shanghai, Hangzhou, and Ningbo have adopted this solution. The healthcare AI model is built on DeepSeek’s R1 and V3 models, Alibaba’s Qwen, and Ant’s own BaiLing. Ant’s healthcare-specific model can answer questions about medical topics and also help improve healthcare services.
© IE Online Media Services Pvt Ltd

Expand