SenseTime's Artificial Intelligence Data Center looks like a chip from its shape
At the beginning of 2022, the construction of Shanghai's artificial intelligence highland has accelerated.
In the Lingang New Area, the SenseTime Artificial Intelligence Data Center (AIDC), which covers an area of about 130,000 square metres and has 5,000 cabinets in its first phase of construction, was officially opened on January 24.
From the appearance, this open, large-scale, low-carbon advanced computing infrastructure looks like a chip. Its designed peak computing process power exceeds 37.4 billion floating-point operations per second, ranking first in China and one of the largest artificial intelligence Data centers in Asia.
SenseTime responds to Shanghai's establishment of an artificial intelligence ‘highland’ with international influence. It will develop large-scale artificial intelligence technology research and development and provide momentum for industrial digital transformation, major scientific research breakthroughs and smart city construction.
The AIDC is based on the research results and technical accumulation of the prototype project construction. In July 2020, AIDC officially started construction and it only took 168 days from the start of construction to the capping of the main structure, setting a new record for port-side construction.
AIDC starts from enabling attributes of artificial intelligence to help more traditional industries improve efficiency. At the 4th World Artificial Intelligence Conference in 2021, the SenseTime AI device, which is "built" by computing infrastructure, deep learning platform, and model layer, came out.
Accelerating the localization of chips and realizing safety and control is a major goal of the development of China's semiconductor industry. At present, the development of China's AI industry is largely based on the software and hardware ecology built by foreign advanced enterprises. This dependence makes the future development of China's AI industry controlled by others, and also brings hidden dangers to the security of the supply chain. Many experts believe that on the one hand, this phenomenon stems from a certain generation difference in chip manufacturing process and manufacturing technology at home and abroad; another important reason is the strong supporting basic computing platform ecology cultivated and accumulated abroad for many years.
It is reported that at last year's World Artificial Intelligence Conference, SenseTime also established the "Artificial Intelligence Computing Power Industry Ecological Alliance" with Fudan University and Shanghai Jiaotong University, which is committed to accelerating the formation of the artificial intelligence domestic computing power ecological foundation and realising support for a group of chip companies, cultivate a software ecosystem, and prosper a group of application development.
AIDC Data Operation Centre
In the field of government affairs, the AIDC will accelerate the process of intelligent urban governance, covering a variety of automated supervision scenarios such as overflowing garbage, lost manhole covers, damaged light boxes, random parking of bicycles, and illegal road occupation.
Due to the adoption of a series of green and low-carbon technologies, the power consumption of the SenseTime AIDC is expected to be about 10% lower than the average level of data centers in China, saving about 45 million kWh per year. It is expected that around 2025,AIDC will reach the peak of carbon dioxide emissions, and the peak emissions are estimated to be no more than 350,000 tons of carbon dioxide equivalent; it is expected that around 2050, it will reach net zero emissions.
To empower the digital transformation of the energy sector, one of the major sources of carbon dioxide emissions, SenseTime is enabling smart inspections on power grids with its SenseMARS Augmented Reality platform, boosting the safety and efficiency for grid-connected renewable systems. Furthermore, SenseTime’s high-accuracy 3D remote sensing system provides comprehensive analytics for environmental management and the implementation of various carbon neutral practices.