Along with the progress of science and technology, the continuous development of society, AI Intelligence is a hot topic with high frequent comment. AI Intelligence is no longer a rare thing that needs to be farsighted, but now widely used at home appliances. And it plays a widespread application in robots, economic and political decision-making programs and control systems.

This also makes AI chip store speculation growing rapidly. Artificial Intelligence (AI) chip stores will reach $16.6 billion by 2022, with a compound annual growth rate of 62.9% from 2016 to 2022, according to the Industrial Science and Technology Information Room. The base year for this query is 2015 and the guess period is from 2016 to 2022. AI is the integration of the most advanced skills for developing products similar to human intelligence.

The reason that there is such a market change is because the global data becomes more and more. The data computing will rely on the CPU's operational functions, thus create a global AI chip war.

NVIDIA: leader of GPU, king of voice chip

When it comes to AI computing chips, of course, NVIDIA is talked about dutifully. Founded in 1993, NVIDIA is a Fabless American IC semiconductor company and a leading global GPU manufacturer. Its GPU is the industry leader in image processing, and the synchronization operation of GPU chip is very suitable for deep learning neural network of artificial intelligence.

In practice, NVIDIA is particularly outstanding in floating-point computing, parallel accounting and other parts of the accounting, and its GPU can provide dozens or even hundreds of times the CPU function. And now the world's GPU shopping malls have been occupied by NVIDIA and AMD two major companies. What’s more exaggerated, NVIDIA market share is now as high as 82% in the second quarter of 2015. The planned Terga (4-core CPU + 256-core GPU) is now the chip of choice for Google's unmanned sedan.

In 2016, NVIDIA's share price rose by 228%. It can be said that NVIDIA has the highest profit in the industry because of the $50 billion.

Intel:  Grasp Nirvana NVC Series Music Chip skills, FPGA

Intel bought AI entrepreneur Nirvana in November 2016 for 400 million dollars, and Nirvana has been discussing how to embed in-depth learning algorithms near chips rather than simply building software that runs on many graphics processors. Intel has always said that Nirvana skills can make the in-depth learning system run 100 times faster and the program sells chips that incorporate Nirvana skills in 2017. And a few days ago, Microsoft and Intel jointly launched the enterprise-class block chain services. Now they have selected the ether currency and JP Morgan, which is very well in prospect! And this time Intel and Microsoft collaborated on the launch of Brainwave, and was so cool on the function and development potential. Finally it completed the mission that combine the cloud to run AI to achieve real-time supply computing results.

Published by Joseph Nicholls