HUAWEI's "Da Vinci plan" hard solution: chip, artificial intelligence, smart phone three cart driving force

According to foreign media reports, HUAWEI is currently launching a new project named AI Project Da Vinci - Da Vinci plan. Xu Zhijun is now responsible for the rotating presidency and vice chairman of HUAWEI. What are the Xu direct army? Xu Zhijun, chairman of HUAWEI's IC design, Hass, the first reaction is that HUAWEI is going to play the role.

In a short sentence, I wonder if all the officials have captured the relevant keywords: Da Vinci plan, a handwheel value officer, Xu Jun, IC field.

HUAWEI IC field force

Xu Zhijun, who joined HUAWEI in 1993, has been the president of the company's wireless product line, the president of the strategy and Marketing, the president of the product and solution, the director of the product investment review committee, and the vice chairman of the current company, the rotating CEO and the director of the strategy and Development Committee. In 2004, HSI semiconductor was established, and Xu Zhi jun was chairman of the army. The main business was to develop CPU for his own family. At this time, it has become the blade of HUAWEI in the field of IC.

HUAWEI's "Da Vinci plan" hard solution: chip, artificial intelligence, smart phone three cart driving force

Data show that in 2017, the IC design scale reached 207 billion 350 million yuan, an annual increase of 26.1%. HUAWEI Hass semiconductor is ranked first in 36 billion 100 million of sales, ranking in second is purple light with 11 billion income, third is ZTE microelectronics, and the top ten are Hua Da semiconductor, smart core microelectronics, sink top technology, and microelectronics.

There is still a big gap between Hass and the world famous IC design company. The revenue of Qualcomm is up to 17 billion US dollars, which is more than 3 times more than that of Hass's revenue of 4 billion 700 million US dollars. The rapid rise of Heth in the field of semiconductors is due to HUAWEI's strong investment in the IC field, and HUAWEI's 2017 R & D expenditure of $13 billion 800 million, third in the world, but still needs to face the gap between itself and the world's first class IC design company.

In the chip industry, the top 20% of the industry strength has taken more than 80% of the profits. Many Chinese companies, including HUAWEI, seized the market by providing low-cost products, occupying low-end markets, accumulating capital, and then gradually climbing up. However, such a method can not work in the chip industry. Under the impetus of Moore's law, the speed of upgrading products and technology is also fast. When the price is constant, the number of components that can be accommodated on the integrated circuit will double every 18-24 months, doubling the performance, and at the same price the consumer will be able to buy more excellent performance products. The chip industry is taking the lead step by step. It is very difficult to climb from the low end to the high end.

With the increasing intelligence of the terminal products, the traditional chip can not meet the demand. The emergence of artificial intelligence has made many companies in a better starting point, and let the company which is at the bottom of the chip industry and have accumulated a certain period of economy and technology, have more opportunities to go on. The emergence of AI is equivalent to the chip industry to some extent to re shuffle the shuffle, HUAWEI just has this kind of details and ambitions.

Mobile chip dares to say "no" to Qualcomm

As we all know, the most important part of the performance strength of smart phones is cell phone chips, and high Qualcomm occupies almost half of the smartphone chip. According to relevant data, the top three high Qualcomm SoC market share in 2017 are high Qualcomm, apple and the joint hair service, and their market share is 42%, 20% and 14% respectively. HUAWEI's Hai Si chip is divided into 8%, and the 8% share, though it is said to be self-sufficient for HUAWEI, has become the only mobile phone manufacturer in the Android Department of China and in the world to say "no" to Qualcomm.

HUAWEI used high pass chips in almost all mobile phones. In the past seven years, however, it has developed its own mobile chip, and has accumulated more and more important patents related to wireless technology, which has turned itself into a strong competitor of high Qualcomm.

At the IFA 2017 exhibition in September 2nd, HUAWEI officially released the world's first AI processor - Unicorn 970. This is the world's first independent NPU (neural network unit) smart phone AI computing platform, Kirin 970 appeared directly to the standard Qualcomm Xiaolong 845.

Kylin 970, using the 8 core design, 4xA73 2. 4 G H + 4 x a 3 1.8GHz. In addition to the 10nm process, the first independent processing unit NPU for neuron computing was added to the industry, and CPU, GPU, DSP composed of HiAI artificial intelligence mobile computing platform. The AI performance of Kirin 970 is much better than that of CPU and GPU. In dealing with the same AI application tasks, the new heterogeneous computing architecture has about 50 times energy efficiency and 25 times performance advantages. In terms of GPU, Qilin 970 pioneered the commercial use of the MP12, which is 20% higher than the previous generation of graphics processing and 50% higher in energy efficiency.

Embrace artificial intelligence

HUAWEI executives will discuss artificial intelligence specific projects at the monthly summary of the Shenzhen headquarters, and embracing artificial intelligence has become the consensus of the HUAWEI top level.

The artificial intelligence has gone through four stages, each stage is not the same, like the growth process of the human baby. First, in the 1950 "the father of the computer" Alan Turing put forward a point of view, the machine and human do the same thing and not be identified, this is the intelligence, the submersible concept of artificial intelligence is born. In two and 70s, because of the development of logic, psychology, analytical science and computer science, the development of artificial intelligence technology did not meet the requirements of artificial intelligence. Three and 80s, including Internet Co, Microsoft, apple and Google.