At the end of the year, Intel began to pack up and look to the future.
2016 is not a year of good for Intel. In the context of the continued sluggish PC market, Intel experienced a series of layoffs, abandonment of mobile chip business and other transformational pains, until recently announced to the media, focusing on the future transformation of eight key technology areas: artificial intelligence, no Human driving, 5G, virtual reality, robotics, sports, precision medicine and China manufacturing 2025.
The conclusion of this overall strategy, such as Yang Xu, Intel’s global vice president and president of China, is not the result of one day.
On December 13, in an interview, Yang Xu said that Intel had only turned this corner in the past year, that is, Intel is not a pure chip company, and it does not have to rely on chips to find opportunities.
Why was the previous transformation so painful, because the thinking at that time was looking for chips, and I was thinking about finding a place to get the chips up. Yang Xu said that although many terminal small devices require intelligent processors, Intel can certainly do it, but the time it takes is twice that of others. The cost is many times that of others, and the economic benefits are not good.
In this case, Intel needs to seek a more flexible approach. For example, some devices use ARM architecture chips, but Intel's IP people may need to be able to achieve faster and lower cost design. Why can't they cooperate on IP? Yang Xu said that in the past two years, I have really figured out a lot of questions.
After opening up the idea, Intel began to pay attention to the changes in the big picture.
Intel believes that the future of smart Internet will present four major trends: data flooding, digitalization of everything, computing awareness and cloud value-added.
Intel expects that there will be 50 billion interconnected smart devices in 2020, each generating an average of 1.5 GB of data per day, a smart hospital generating 3000 GB of data per day, and a driverless car generating 4000 GB of data per day. An intelligent manufacturing facility will generate 1 million GB of data per day.
Faced with such a large data torrent, how to mine, analyze, and finally add value from the cloud is a key issue that needs to be solved in the future.
To this end, Intel will improve processing capabilities from the cloud, terminal, and intermediate connections.
Take the artificial intelligence that everyone talks the most, for example, the artificial intelligence data processing category is different, there are very complex data, and there are also data that is not complicated but has a large amount of processing. How can artificial intelligence fully exploit and release the data value? One of the important factors is the end-to-end calculation. Yang Xu explained that from the generation and collection of terminal data, to the high-speed transmission and storage of data, to the analysis and mining of cloud data, end-to-end computing capabilities can help artificial intelligence fully release data value and bring more innovation. And value-added, which led to profound changes in smart cities, smart retail and Internet finance, and brought a new user experience.
Eight areas, such as artificial intelligence, driverless, and 5G, are the areas in which Intel believes it can handle its processing capabilities.
However, the outside world is more concerned about when Intel can form scale in these areas and truly produce commercial value.
For such questions, Yang Xu said that although it is difficult to accurately estimate the time, there are two implications for these areas: First, Intel is optimistic about these technological trends and believes that they will bring great changes to people's work and life. New services and new applications will also be born. Secondly, this is also the verification process of technology, including virtual reality, precision medicine, and driverless driving.
Taking driverless technology as an example, it has attracted Intel and so many companies because it is the most forward-looking technology. Unmanned applications can truly verify the maturity of technologies including artificial intelligence, robotics, and whether these technologies can meet future needs.
Yang Xu said that if the problem is found, it will be adjusted. If there is not enough capacity, it will reinforce the ability. If the cooperation method is not correct, it will change the cooperation mode, but it must be tried. The most important thing is that people in the world who have the right to speak in these fields are testing the water.
Among these areas, Intel is undoubtedly the most artificial. In the past year, Intel has acquired a basket of companies in this field, and many people do not understand Intel's layout.
At the recent Intel Artificial Intelligence Forum (AI Day), Intel showed its holistic approach to artificial intelligence, which included five levels of hardware, libraries/language, open framework, tools/platforms, and solutions. .
The underlying Intel's various chip products have the ability to process different types of data, the second layer provides algorithms and function libraries, and the middle two layers are development tools and development platforms. For example, the acquired Nervana FPGA can realize personalized chip, most The upper level is for specific applications in different industries. Yang Xu explained to the interface journalist that artificial intelligence is completely personalized, and the requirements in different fields are completely different, so it is necessary to break through these applications and provide an overall solution layer by layer.
In addition, Intel is building a broad alliance with industry leaders such as Google, establishing the Nervana Artificial Intelligence Committee to promote technology discovery and innovation, and partnering with leading global organizations to provide developer training courses.
In China, Intel has cooperated with partners in Jingdong, Tencent, Alibaba, Baidu, Nuohe Zhiyuan, Shangtang Technology, Peking University, and Tsinghua University to promote innovation and popularization of artificial intelligence.
When talking about the specific performance in 2016, Yang Xu said that since the financial report has not been officially announced, the directional things cannot be disclosed too much. However, he believes that Intel's traditional business such as PC and data center is growing steadily this year, and new areas are being explored. Now, with the directional layout basically completed, the next step is to verify and focus on strategy execution.
Keystone Jack,cat6 keystone,cat6 jack,cat6 wall jack
NINGBO UONICORE ELECTRONICS CO., LTD , https://www.uonicore.com