- Register
- Login
32°C
- Article
- . (0)
- Related content
China-based AI (artificial intelligence) chip startup Cambricon Technologies expects to roll out hundreds of millions of AI SoCs needed for smart terminal devices and servers in the next three years, the firm’s CEO Chen Tianshi has said.
Chen made the remarks while talking the press on the side line of the 2017 Global Artificial Intelligence Innovation Summit held August 30 in Shanghai. He said while the Moore’s Law has been the engine driving the development of AI over the past decade, such an engine will slow down in the next decade, when dedicated deep-learning processors are badly needed to support massive AI applications in cloud services and terminal devices.
Chen said that supported by the Chinese Academy of Science, Cambricon has launched the world’s first IP instruction set for smart processor, and will move to expand its technical licensing to more enterprises engaged in AI applications, so as to help more AI businesses sharpen their smart processing capabilities and jointly build up an AI ecosystem. In this regard, Huawei’s HiSilicon Kirin 970 chipset, to be rolled out in September, is to carry the IP instruction set for Cambricon-1A chip instruction set, the first chip dedicated to high-performance neural network applications.
In the field of smart cloud services, Chen also revealed that Cambricon is cooperating with Dawning Information Industry, a Beijing-based company engaged in high-performance computing, cloud computing, and servers on the application of the firm’s dedicated AI chips and acceleration cards, adding that Sense Time and iFLYTEK, leading AI enterprises in China, are also among Cambricon’s potential customers.
Chen said that as Intel’s X86 instruction set architecture played a dominant role in the past PC era, Cambricon will endeavor to lead the market for AI instruction set architecture for smart terminal devices. It is now an era featuring massive explosion of smart technologies, leading to the development of multiple algorithms for different deep learning levels to support different applications of AI technologies.
Chen Tianshi, CEO of Cambricon Technologies
Photo: Company
-
Micron sells Lexar business to Longsys
Before Going to Press | 7min ago
-
Nanya disposes more Micron shares
Before Going to Press | 8min ago
-
Panel makers to maintain high fab utilization rates in 3Q17, says IHS
Before Going to Press | 35min ago
-
Orders for NOR flash, car electronics chips to buoy Ardentec revenues in 2H17
Before Going to Press | 36min ago
- China AMOLED panel capacity expansion forecast, 2016-2020
This Digitimes Research Special Report examines the China AMOLED industry, focusing on the expansion capacity of the makers, the current implementation plans of major smartphone vendors in the market and the technological hurdles faced by the China makers.
- Global AP demand forecast, 2017-2020
Digitimes Research expects global AP shipments to surpass the 1.9 billion mark in 2017, with smartphones remaining the main application. Qualcomm will be leading the market in 2017, as other players continue playing catch up and scramble for funds to invest in more diverse applications.
- Global notebook shipment forecast, 2017 and beyond
This Digitimes Special Report examines key factors in the notebook industry, including products, vendors and ODMs, that will affect total shipments in 2017 and through 2021.
