Artificial intelligence (AI) refers to the engineering and deployment of machines to enable them to perform tasks with intelligence similar to humans. Since its inception 60 years ago, AI has witnessed significant growth in recent years. Initially, AI was considered as a topic for academicians. However, in recent years, with the development of various technologies, AI has become a reality and is influencing many lives and businesses. Additionally, the evolution of various other supplementary technologies such as cloud computing, machine learning, and cognitive computing is collectively paving the growth of the market for AI.Some the prominent players in AI market in the recent times are Intel Corporation(US), Google Inc.(US), Microsoft Corporation (US), Amazon.com, Inc.(US), Baidu, Inc.(China),and NVIDIA Corporation (US).

Machine learning algorithms often consist of matrix operations. These calculations benefit greatly from parallel computing, which leads to model training performed on graphics cards (rather than only CPU). The major types of chipset considered as AI chipsets for AI applications are as follows:

CPU & GPU: Before 2001, general computing was executed by CPU and computation for rendering graphics was done by GPU. The development of faster and more efficient matrix multiplication and factorization techniques led to the creation of programming languages that allow general-purpose computing on GPUs, including CUDA and OpenCL.

GPU: Major manufacturers such as Nvidia Corporation (US) are looking to address the concern of high power consumption by devices while solving complex machine learning algorithms and programs on a priority basis.

FPGA: FPGAs are integrated circuits whose logic blocks can be programmed and reconfigured using a hardware description language (HDL). Series of FPGA-based mining systems provided increase in throughput performance as well as energy efficiency. The cost of electricity created a break-even, favoring low-power systems. Efforts are being taken to implement machine learning models using FPGAs. As of now, FPGAs only match GPUs on throughput performance, but consume less energy for the same workload. Thus, they are more feasible and most preferred in low-power environments such as self-driving cars.

ASIC: The demand for deep learning and statistical inference is driving the hardware industry toward Machine Learning (ML)-

specialized hardware. Google leads with ML-based ASICs,especially the TPUs; where as its competitors (Amazon, Microsoft, Baidu etc.) run FPGAs.Popular ML libraries include TensorFlow and Caffe.

The increasingly large and complex data sets, as well as the adoption of AI-enabled products to improve consumer-centric services are the two major growth drivers for the AI chipset market

Evolution on the Hardware Front:
Growth in AI applications would lead to a demand for ML-specialized devices. This would drive the hardware into the next phase of evolution. The major impact of these technologies once applied /deployed would be witnessed in healthcare and medicine, automotive and transportation, and robotics (for various applications). The majority of start-ups globally have emerged in the past two years owing to significant technological advancements.

However, the ratio of the software start-ups to that of hardware start-ups (exclusive to machine learning) within AI ecosystem varies drastically because of the cost of developing the ML chip hardware.

According to Mr.Sachin Garg, AI chipset is likely to gain traction with the amalgamation of various technologies such as deep learning, robotics, digital personal assistance, querying, natural language processing, and context-aware processing to develop an AI-featured product. In the near future, AI is expected to have a crucial impact on multiple end-use applications such as driverless cars, healthcare diagnostics, and physical assistance in elder care.

Exponential Evolution Expected to Continue:
AI is likely to disrupt every business segment across the world. The end-use industries of AI include agriculture, BFSI, manufacturing, healthcare, oil & gas, media & advertising, transportation and automotive, and retail. Among these verticals, transportation and automotive sector was the largest contributor to the AI chipset market in 2015. This was because of significant investments made by federal governments and venture capitalists in the development of connected and autonomous vehicles. This funding consequently resulted in some of the tangible recent developments in autonomous projects. For instance, in October 2016, Otto (US) (now acquired by Uber Technologies, Inc. (US)) shipped its first consignment from Denver, Colorado, US to Denver Springs, U.S. using driverless truck over a 120-mile journey. However, the healthcare industry is estimated to exhibit the highest growth in the near future. The growing penetration of AI in healthcare assistance and medical management are the two key factors driving the market for the healthcare industry. Moreover, AI technologies are also used in the healthcare industry for drug discovery and formulation processes, decision-making processes, extraction of information from the data collected from patients, and further processing for testing and simulation of new treatments, scenarios, and devices.

The increasingly large and complex data sets,known as big data, as well as the adoption of AI-enabled products and services to improve consumer-centric services are the two major growth drivers for the AI chipset market. However, the scarcity of low-cost and energy-efficient hardware, and the lack of skilled workforce for the development of AI algorithms and tools restrains the growth of the AI market.

In essence, the AI chipset market offers significant growth opportunities for all the stakeholders involved in its ecosystem. The development of human-aware AI systems and the growing widening scope of AI technologies in niche markets are the prominent opportunities existing today. In addition to this, the deployment of AI chipset at the edge of the network is likely to become a massive industry trend in coming years. Thus, AI technologies driven by the AI chipsets are the future for many industries that are striving for efficient, low-cost operations and high profitability in their markets.