GuShiio.com 
(Photo source: unsplash)
The DeepSeek craze has triggered a new round of discussion on the extended laws of reasoning data and computing intelligence.
On February 27, Beijing time,AI giantNvidia(NVIDIA) released fourth-quarter and full-year results. In the natural year of 2024, Nvidia’s revenue reached US$130.5 billion (approximately RMB 947.4 billion), an increase of 114% over the same period last year; gross profit margin was as high as 75.5%; under non-GAAP, net profit was US$74.265 billion, a year-on-year increase of 130%.
At the earnings conference, Nvidia founder and CEO Huang Renxun said that the reinforcement learning technology used in reasoning AI models such as DeepSeek requires the use of a large amount of synthetic data for training, especially end-to-end data processing, data training and management. Therefore, he proposed a new Multiple Scaling Laws for the calculation of inferential data:The more calculations the reasoning time or test time model uses, the deeper the reasoning process will be and the smarter the answers will be.
“Reasoning itself is a computationally intensive process. In Huang Renxun’s view,In the future, computing volume and data intelligence will be crucial to the development and implementation of reasoning AI models.
On February 25, Guo Wei, chairman of Shenzhou Holdings, said when talking about AI data that just as Ford’s assembly line completely changed the manufacturing industry a hundred years ago, today’s AI technology is actually redefining the company’s digital assembly line. By providing a series of knowledge management tools, from data distillation to data generation, large models like DeepSeek can better serve thousands of industries. rdquo;
Guo Wei, Chairman of Shenzhou Holdings
On the same day, Sun Yang, president of Shenqi Digital, told GuShiio.comAGI that through data intelligence technology, AI can accelerate the digital transformation of enterprises. Those with low degree of digital transformation, especially in some traditional industrial manufacturing enterprises, have greater competitive pressure and low gross profit levels. Enterprises ‘investment in digitalization is actually insufficient. Therefore, through a complete set of digital intelligence solutions, the company has transformed it from quantitative change to qualitative change. Through the combination of data +AI, it helps customers optimize processes and technologies with less investment, improve productivity, and reduce operating costs., the money saved is the more profits created.
“Through data intelligence technology, we feed purified and managed data to large models, so that we can pull it through and generate value. Therefore, our slogan is to let data create value. The value of the data itself must be created through industry applications and models. rdquo; Sun Yang said.
In fact, recently, DeepSeek, a domestic open source AI model, has attracted industry attention due to its low cost, high performance and open source features. More than 200 companies and related applications have adapted the DeepSeek series models. However, for most companies, especially Traditional manufacturing industries are still at a loss on how to apply large model tools such as DeepSeek, and many enterprise basic data governance issues have not yet been resolved.
In this context, on February 25, Shenqi Digital, a data intelligence brand owned by Shenzhou Holdings (00861.HK), officially released a new generation of intelligent software platform Yanyun Infinity at Peking University, and announced that it will jointly build Yanyun Infinity intelligent software with Peking University. Joint laboratory. Yanyun Infinity focuses on data-driven intelligence and allows enterprises to efficiently make correct decisions through the joint collaboration of people and AI, thereby realizing AI for Process in the entire process, improving quality, reducing costs, and increasing efficiency for enterprises.
At the meeting, Mei Hong, a professor at Peking University and academician of China Academy of Sciences, said that as an important infrastructure in the digital economic era, the integration of intelligent software and thousands of industries will promote digital transformation and empower new productivity. To this end, we first need to digitize the entire process and all elements, and then build a platform through software definition to realize organizational restructuring and process reengineering. Finally, we use various algorithms and models to conduct analysis and decisions, that is, AI for Process, to achieve quality improvement and efficiency.
Mei Hong emphasized that the establishment of the joint laboratory is an upgrade of cooperation between Peking University and Digital China. It will rely on the advantages of both parties to carry out key technology research and application innovation and provide effective solutions for the digital transformation of traditional industries, especially in the industrial field.
Zhang Wei, general manager of Shenqi Digital Technology R & D Center, cited steel companies as examples. One of its application scenarios is to intelligently control the height of iron tongs. Through data collection and analysis, prediction and prediction can be made, and ultimately reduce the operating cost of the steel furnace.
Zhang Wei said that from placing orders to completing delivery, companies need to go through a large number of data-based decision-making processes, and there is much room for improvement in the efficiency and accuracy of traditional decision-making processes. By introducing Yanyun Infinity, we can build a digital and intelligent twin of the enterprise through data aggregation, combine the condition selection, result analysis, and prediction deduction of various process links of the enterprise’s operation, and visually present each possible result and its reasoning. and the evolution process builds intelligent decisions for the enterprise, allowing people to quickly lock in the optimal path and automatically complete the execution actions corresponding to the decisions.
Digital China Holdings Co., Ltd. was established in 2000 and listed on the main board of the Stock Exchange of Hong Kong in 2001. The latest financial report shows that in the first three quarters of 2024, Shenzhou Holdings achieved operating income of 6.68 billion yuan, a year-on-year increase of 0.94%. Among them, the financial industry achieved operating income of 3.039 billion yuan, of which financial soft service income was 2.459 billion yuan, a year-on-year increase of 19.42%.
GuShiio.comAGI learned that Shenzhou Holdings is focusing on the big data +AI strategy, and continues to expand data products, accelerate the layout of company factor areas, and help enterprises digital transformation in fields including logistics, supply chain and other fields. The newly established Shenqi Digital, based on Yanyun Infinity technology, creates a one-stop AI+ data intelligent decision-making service platform for enterprises in the data layer, knowledge layer, decision execution layer, etc., to improve quality and efficiency for enterprises, including Tencent Cloud, Jiyun Technology, Hong Kong Software Association and other hundreds of companies provide ecological support.
“I hope that the Joint Laboratory will become an innovative highland for intelligent software research and application, and contribute new strength to the construction of Digital China and global scientific and technological development. ”Mei Hong said at the end of his speech.
