China Mobile Aims to Triple AI Computing Power with Homegrown Chips by 2028
- tech360.tv
- Oct 14
- 3 min read
State-owned telecom giant China Mobile announced plans to exclusively use homegrown chips for its artificial intelligence computing network by 2028. This move aims to triple China Mobile's AI computing power, accelerating China’s drive for AI leadership and reducing reliance on foreign technology.
Chairman Yang Jie outlined this ambitious target at the company’s Global Partners Conference in Guangzhou on Saturday. He stated the telco plans to double its investment in AI.
China Mobile also intends to explore constructing a computing cluster with 100,000 graphics processing units, known as GPUs. This would bring its national total to 100 EFLOPS of AI computing power.
An EFLOP, a unit of computing speed, equals one quintillion floating-point operations per second. This unit rates the performance of the world’s most powerful supercomputers.
Using an industry standard of FP16, China Mobile would effectively triple its total AI computing power. At the end of 2024, its power stood at 29.2 EFLOPS at FP16.
FP16, also known as half-precision floating-point format, is a computer number format. It uses 16 bits of memory, half the standard 32 bits.
Yang said, "Human society has fully entered the ‘AI Plus’ era, with AI as the core engine of new productive forces." He added, "Artificial intelligence will not replace humans, but will instead handle repetitive and rule-based work, while humans will continue to lead in emotion, creativity, and moral judgement."
China Mobile plays a key role in China’s AI infrastructure push. As of the end of 2024, it accounted for 6.4% of the country’s total computing power.
The state-owned enterprise also plays a central role in the government’s “East Data West Computing” project, launched in 2022. This scheme seeks to integrate the country’s computing resources.
Under the scheme, data generated on the country’s more densely populated east coast is transported. It then goes to more energy-abundant inland regions for processing.
As of April, China Mobile had established more than 21 AI data centres across the country. Some of these centres, located in Hohhot and Harbin, have over 10,000 GPUs.
Beijing, which wants to build computing self-reliance in the face of US restrictions on advanced semiconductors, has emphasised the need for greater coordination and supply of computing resources as part of its “AI Plus” national strategy, unveiled in late August. This is part of its “AI Plus” national strategy, unveiled in late August.
The capital wants to build computing self-reliance amid US restrictions on advanced semiconductors. Accelerating AI infrastructure buildout is also expected to be a major theme in China’s 15th five-year plan in 2026.
As of May, the US led the world with approximately 75% of global AI computing power. China held second place with 15%, according to US research institute Epoch AI.
China Mobile’s latest initiative is part of its own “AI Plus” strategy. This aims to support large-scale commercial AI adoption domestically.
The strategy also seeks to support the export of AI capabilities internationally. This is particularly focused on the global south.
The state-owned enterprise, listed in Shanghai and Hong Kong, is the largest telecom operator in the world. It serves more than 1 billion customers.
China Mobile plans to triple its AI computing power to 100 EFLOPS by 2028.
The telecom giant intends to use only homegrown chips for this expansion.
This initiative supports China's goal for AI leadership and reduced reliance on foreign technology.
Source: SCMP
