China’s smart computing gets upgrade as AI demand, tech restrictions drive expansion

The world’s second-largest economy is unifying its computing resources to speed up tech development as these and similar curbs escalate. By the end of May, China had over 10 smart computing centres with high-performance clusters.

Beijing is also eager to achieve technological self-reliance as it searches for new sources of economic growth, seeing great potential in the dividends up for grabs on the bleeding edge.

According to estimates from CAICT, every 1 per cent increase in computing power contributes around 0.2 per cent to the country’s economic growth, and 0.4 per cent to growth in the digital economy. The institute also calculated that every yuan invested in the industry will generate three to four times that amount in economic output.

05:03

How does China’s AI stack up against ChatGPT?

How does China’s AI stack up against ChatGPT?
The race for tech parity has become more urgent following the US release of large language model ChatGPT and the text-to-video model Sora. These developments, coupled with the increasing likelihood of Donald Trump’s re-election as US president, could intensify the country’s tech blockade against China.

As the AI boom fuels a surge in demand for computing power, local governments have supercharged the construction or deployment of their own clusters.

For example, the southwestern city of Chengdu – home to nearly 900 AI companies valued at a cumulative 78 billion yuan (US$10.8 billion) – put a new smart computing centre into operation on July 2. Local authorities said they expected the centre to enhance efficiency by up to 80 per cent and save costs for research institutes as well as tech firms.

China released its first computing interconnectivity platform last week, which will identify, register and test computing resources nationwide and provide AI firms access to computing power from across the country.

Analysts had said the country’s fragmented computing market was a bottleneck preventing the flow of data across regions, hindering the progress of China’s AI industry and increasing costs for tech companies training large models.

On the national level, China plans to build eight computing hubs and 10 data centre clusters in a megaproject called “Eastern Data and Western Computing”, projected to drive around 400 billion yuan in investment each year.