The race for the domination of AI chips
With AI and advanced semiconductor technology an integral part of Industry 4.0, the impact of AI chips on the global technology landscape will gradually evolve in the coming decade
With the emergence of new applications of semiconductors, the concept of using artificial intelligence (AI) algorithms on high-end chipsets has opened an entirely new market for these devices, also known as AI chips.
New and emerging technologies such as 5G and Internet of Things (IoT) have created a need for devices with increased computational power and capabilities. Technologies such as machine learning and deep neural networks, which are part of the AI ecosystem, have a tremendous workload that cannot be fulfilled by traditional chipsets. AI algorithms work on parallel processing or parallelism, which is basically the ability to multi-task and simultaneously run different computational processes.
AI chips, in recent times, have tried to incorporate the needs of AI algorithms into chipsets that can be used both in the cloud (servers and data centres) as well as at the network edges (in smartphones, tablets, and other consumer devices).
Applications such as Biometrics and Image Recognition need AI chips in the cloud or servers for maintaining a large amount of data. The use of AI chips remains integral in data centres which eventually reduces operational costs and improves information management. But the role of AI chips has also diversified with applications in various industries such as robotics and autonomous vehicles. These technologies need AI algorithms for efficient and effective working, with the computational power of the chipsets needing incredibly fast processing speeds due to the need for decision-making in certain situations.
The market for AI chips has consistently increased in the last decade, with AI chips projected to account for 22% of the global AI revenue by 2022. A strong compounded annual growth rate of 51% has been projected for the AI chips market in the next four years, with technologically-advanced regions dominating the market in the future.
AI chips are also a unique requirement of different entities, ranging from smartphone manufacturers such as Apple, Samsung, and Huawei, to traditional chip designers like Qualcomm and MediaTek, to intellectual property (IP) license providers such as ARM. With most of the major semiconductor companies across the world in the business, AI chips look to be the next big thing for the industry.
Semiconductor companies have already thrown their hats in the AI ring with the development of advanced AI chips such as Graphical Processing Units (GPUs). NVIDIA has developed a dedicated application programming language called CUDA used specifically in parallel computing on GPUs. Other targeted AI chips such as Field Programmable Gate Arrays (FPGAs) and Application Specific Integrated Circuits (ASICs) have also been designed for specific applications of AI technology. Companies such as Apple and Google have also invested in the manufacturing of these chipsets keeping in mind their devices or softwares’ specific needs such as image recognition and speech processing units.
Unlike traditional chipsets, the difference between the leading and trailing edge nodes in AI semiconductor chips, especially with respect to cost-effectiveness, remains negligent. In fact, trailing edge nodes remain both efficient and cost-effective in the long run.
The semiconductor design market is concentrated in the United States (US), with top-end designs dominated by US companies. Chinese semiconductor design firms still remain behind the US companies in high-end semiconductor design as they are heavily reliant on Electronic Design Automation (EDA) tools, which are owned by top US companies. Apart from this, US companies also have an oligopoly over the AI chips market with the GPU market dominated by NVIDIA and AMD, while FPGAs are concentrated with Intel and Xilinx.
However, experts believe that the need for leading nodes is diluted with AI chips, which China has exploited. China also has the capability to manufacture trailing edge node chips, which can be used for AI applications with its own foundries rather than relying on Taiwanese firms for the manufacturing process. Though the US has a significant advantage in the AI chips market with its high-end designs, EDA tools, and leading-edge manufacturing, China has utilised a mix of both leading and trailing edge technologies to come up with its own chips with AI capabilities.
With increasing global economic revenue and a large market ripe for capture, the presence of China in the AI chips business has also been increasing. AI chip funding activity in China has been active with the hope of creating industry-leading capabilities in machine learning, deep compression, pruning, and system-level optimisation for neural networks. Chinese technology companies such as Alibaba and Huawei have invested heavily in the manufacture of AI chips for smartphones and other devices. Some Bitcoin mining equipment manufacturers in the country are also getting into the AI optimisation game. With domestic AI research in China still playing catch up vis a vis the capabilities of Western countries, these local manufacturing companies have relied on tweaking existing algorithms to create modified AI models. But increased investments along with State support and financing, similar has been done for the semiconductor industry in China, has made AI chips an important technology worth pursuing in all technically adept states.
With AI and advanced semiconductor technology an integral part of Industry 4.0, the impact of AI chips on the global technology landscape will gradually evolve in the coming decade. The race for the domination of the global AI chips market is something to watch out for in the not-so-far future.
Arjun Gargeyas is a research analyst at Takshashila Institution
The views expressed are personal