0G’s DiLoCoX framework, in partnership with China Mobile, successfully trained a 107 billion parameter decentralized model in 2025, boasting a communication efficiency 357 times greater than traditional methods. This advancement in decentralized AI training reflects a broader industry trend where such technologies have moved beyond theory to practical applications, particularly as telecommunications providers increasingly collaborate with AI projects to enhance distributed model training efficiencies.

0G: 0G develops decentralized infrastructure solutions focused on AI and data processing. Its DiLoCoX framework supports efficient collaborative training of large AI models across distributed networks. The project recently partnered with China Mobile to apply this framework in a real-world decentralized training initiative.
China Mobile: China Mobile is a major telecommunications operator providing network infrastructure and services across China. It collaborated with 0G to test and implement decentralized AI training using the DiLoCoX framework. This involvement underscores the company’s role in supporting advanced computing applications through its telecom capabilities.

`json
{
“Industry Trend”: “Decentralized AI training has moved from theory into practical and large-scale applications.”,
“Partnership Model”: “Telecommunications companies are partnering with AI projects to improve distributed model training efficiency.”
}
`