A Chinese research institution approved by the central government has been working on creating a platform that integrates all of China's computing resources and artificial intelligence prowess to enable shared access for everyone, according to the institution's founder, who said it should be made as "easy as using electricity."
It comes as China's tech sector is experiencing an unprecedented period of growth, with the country's AI industry having been energized by the emergence of a new cost-effective model released by Chinese start-up DeepSeek, which has gained significant international attention since its release early this year.
The launch of the DeepSeek-R1 large language model back in January caused a global sensation with its performance hailed for rivaling that of other top international models, despite its development cost being only one-thirtieth of that of similar products.
The Peng Cheng Laboratory, a research institution headquartered in the southern Chinese tech hub of Shenzhen, has made the DeepSeek series of models available in the open-source community for domestic developers to test and validate.
Gao Wen, the institute's founding director who is also an academician at the Chinese Academy of Engineering, shared his insights on the impact of DeepSeek on developers. He particularly pointed to the openness of the program which allows broad access for all users, rather than a select few tech giants dominating the market.
"I believe the emergence of DeepSeek will accelerate advancements in this entire field by many years. Previously, large models like ChatGPT, LLaMA [by Meta], and those from Google or Facebook required substantial resources from leading enterprises, including sufficient computing power, talent, and data for their development and application. Currently, AI applications abroad are primarily in the hands of a few major companies like Microsoft and Google. However, DeepSeek changes this dynamic by democratizing access; it's no longer limited to a select few top enterprises or players—now all startups can engage with it. This significantly broadens the possibilities, which is its greatest contribution," said Gao in an interview with China Central Television (CCTV).
The introduction of DeepSeek has triggered a series of chain reactions in various sectors, while the lowering of application barriers has stimulated increased demand for advanced computing power across society.
Gao said that AI acts as the 'fuel' driving the progress of computing power, which in turn will help AI to develop to its full potential. He noted that the laboratory will need to accelerate its plans to cope with the increased demand for computing resources.
"We did not initially anticipate such rapid growth in demand, but now we need to adjust quickly. Much of our work must be expedited," he said.
One of the Peng Cheng Laboratory's main tasks is to lead the research and construction of the "China Computing Network (C2NET)" in line with major national strategies, such as the East-to-West Computing Capacity Diversion Project and the Digital China initiative, which Gao said aims to ensure ease of access to all.
"Our goal is to allow users to access computing power as easily as they use electricity. If they need it, they can purchase it, and they can buy the cheapest computing power available," said Gao.
He explained that the purpose of constructing C2NET is to both better integrate and ensure computing resources are more fairly distributed across the country, including supercomputing centers, data centers, and cloud computing platforms.
The integration program aims to create a unified pool of computing resources, establishing a digital economy infrastructure that facilitates easy access to resources, unified task scheduling, and a sustainable operational model. This will advance autonomous computing resources across the country into an era of "shared access for all," he said.
"C2NET aims to follow a similar path. We are building a national superhighway for computing power using new technologies while also hoping to introduce new management models—departing from conventional communication management styles. We are currently collaborating with the National Data Administration (NDA) to develop a new model based on data association," said Gao.
He stated that his team hopes to enhance the utilization efficiency of all invested resources through C2NET, drawing parallels to the initial concept behind the country's power grid.
"We are quite confident in our technologies. By progressing step by step, we believe we can achieve significant advancements. Naturally, we aspire to obtain the best resources, including the finest craftsmanship, the most advanced software, and the fastest, highest-quality machines available globally," he continued.
Chinese institution aims to change tech landscape by provide supercomputing access to all
