For a year, NVIDIA's price per share has risen from $40 to $168.
"NVIDIA's share price will rise." In early June, in the gap between Shenzhen and Ubisoft Robotics, Tao Dacheng told Tencent Technology. Tao is a well-known expert in the field of artificial intelligence. He is now a member of the International Association of Recognition Patterns and is the first scientist of the company.
Among the upright walking robots developed by You Bi, the application of NVIDIA is being considered. The rise of artificial intelligence, graphics processing is of paramount importance, and the development prospects of NVIDIA's GPUs that specialize in image processing are gaining more and more recognition.
In addition to industry experts like Tao Dacheng, many people from NVIDIA also include some large US investment banks. They are bullish on Nvidia's share price, from the first quarter of 2018 to $180.
The investment of Alibaba's Softbank Sun Justice is also optimistic about Nvidia. In May of this year, Softbank spent about $4 billion to buy a 4.9% stake in Nvidia, probably the fourth largest shareholder of Nvidia. Sun Zhengyi, the founder of Softbank, is a technology-based investor. He regards artificial intelligence as the next wave of technology after the Internet and mobile Internet. He actively deployed and operated in the upstream chip field and purchased the famous chip design company ARM. .
NVIDIA, founded by Jen-Hsun Huang, sells its products to the data centers of major Internet companies, car manufacturers, and various types of artificial intelligence, because of the leading technology in artificial intelligence. Innovation and startup companies are in the midst of enthusiasm.
On May 24, Huang Renxun pointed out that NVIDIA GPU computing has become an indispensable tool for contemporary Darwin and Einstein. Over the past year, the number of startups that have built GPU-driven AI services has grown by more than four times to 1,300. No one wants to miss the next major technological breakthrough."
Artificial intelligence outletAt the beginning of June, during the Taipei Computer Show, NVIDIA founder Huang Renxun made a debut and gave a speech. This lecture held at the Grand Hyatt Taipei, the process management level is quite different from that of similar conferences in China. There is no real-time voice conversion text display, no super-string, super cool large electronic screen, and time is almost near lunch time. Channels from the industry, various media people, etc. were all stopped outside the elevator on the first floor, patiently waiting for admission. “Like a pilgrimage,†says a local channel.
When entering the third floor venue, people scrambled to find a seat, and finally many people had to stand and listen. Born in Taiwan, Huang Renxun has a Chinese face and speaks English in his mouth. He was dressed in a bright black leather jacket and looked powerful.
The opening is Huang Renxun's statement that the CPU is developing slowly, and Moore's Law has stagnated after 30 years of progress. Electronic transistors, which have been pushing hard on hardware in the past, now grow at a rate of 50%, and CPU performance is only 10% higher.
In Huang Renxun's view, the CPU (Central Processing Unit) is entering a difficult position, and the GPU (Graphics Processor) will replace it and lead the scientific and technological community. "NVIDIA GPU computing has taken the industry a big step - up to 1000 times by 2025." He also cited NVIDIA's CUDA architecture, groundbreaking speed-up capabilities for artificial intelligence.
Taking deep learning neural networks as an example, two years ago, training Microsoft ResNet modules required 7 million megabytes of operation. Now, module training has increased by 15 times, and this demand is constantly expanding and deepening. In the future, the software will be composed of many neural networks, and the demand for processors and software will continue to rise.
Why do you need GPUs for deep learning? In the past, software engineers were responsible for developing programs and carefully writing algorithmic code. Now, various algorithms can learn from a huge number of examples, and the software can write code by itself. Deep neural networks are deployed in data centers and smart devices to reason and predict next steps.
All of this is achieved by training and deploying deep neural networks on the GPU. GPU-driven deep learning is a new computing paradigm in which deep neural networks are trained to recognize patterns from massive amounts of data, and this new model also puts humans into the era of artificial intelligence computing.
In response to Tencent Technology, Nvidia stated that the GPU they invented can effectively solve some of the most complex problems in the field of computer science, while simulating human intelligence, by running deep learning algorithms, becoming computers that can perceive and understand the world. , robots and the brains of self-driving cars.
What do you think of Google (microblogging) for developing TPU for Go (AlphaGo) to beat the world champion? NVIDIA's reply is: "Google is a very good customer, and their use of artificial intelligence and complex models is constantly increasing. Like Google, there are many other customers who are increasingly demanding acceleration, and our GPUs serve These markets include all architectures, industries, and all cloud service providers."
"Now, the world's major Internet and cloud service providers are using NVIDIA's GPU chips," NVIDIA Chief Financial Officer Colette Kress said after the release in the first quarter of 2017.
In the past year, large Internet technology companies such as IBM, Amazon, Google, Tencent and Alibaba have been deploying cloud computing and big data, which constitutes a huge customer base of NVIDIA. This power is so powerful that mobile phone manufacturers can feel that memory prices have been pulled up.
NVIDIA is still continually upgrading its GPU accelerators, and the latest is the seventh generation architecture. On May 11th, NVIDIA released the new Tesla V100, also known as Volta, at the GPU Technology Conference in San Jose, USA, which is said to be the strongest GPU accelerator in history. NVIDIA claims that Volta will become the new standard for high performance computing, with outstanding performance in computing science and data science.
Volta uses 21 billion transistors and its deep learning performance is equivalent to 100 CPUs. By consolidating the CUDA(R) core and the new Volta Tensor core into a unified architecture, a single server with a Tesla V100 GPU will replace hundreds of traditional HPC commercial CPUs. The Tesla V100 GPU has broken through the deep learning of 100 trillion floating point operations per second.
Various new applicationsIn early May, Nvidia released its first quarter earnings report as of April 30, 2017. Among them, the artificial intelligence part of the data center revenue increased by 186% compared with the same period of last year. The latest quarterly income was 1.94 billion US dollars, and artificial intelligence accounted for more than 21%. This ratio was only 6% two years ago.
What are the products that help NVIDIA achieve artificial intelligence growth? NVIDIA told Tencent Technologies, including NVIDIA Tesla GPU, GRID software and NVIDIA DGX, which includes the DGX-1, the world's first single-chassis artificial intelligence supercomputer, and the personal supercomputing workstation NVIDIA DGX StaTIon.
These things are not boring code. In front of some developers, it can solve real problems, and some have finally been applied to mobile apps.
Aipoly is a smartphone app that instantly recognizes more than 4,000 objects in a home environment, such as tools, bathrooms, etc., which can then be displayed on the screen and can be named to the user. Its vocabulary is equivalent to a five-year-old child. At present, the vocabulary of this app is constantly increasing, and the objects that can be recognized are constantly updated. This applies to image recognition training to provide accuracy, and reaching the corresponding quality can help visually problematic people.
Constantly accurate image recognition requires a lot of training. NVIDIA's Nvidia GPUs are the key. “The time required to train NVIDIA products is compared to competing products. The former is like baking a piece of bread, while the latter is like a whisky that is aged (microblogging),†said Aipoly founder Rizzoli.
Brad Folkens, co-founder of CloudSight, a Los Angeles-based company, agrees. He made a free, open source TapTapSee for the visually impaired. This app allows the user to double-tap the screen, take the previous events, and then the app can tell the subject. Folkens believes that image recognition and deep learning behind the app are key, and they are inseparable from the technology provided by NVIDIA.
“With DevBox, we can process large numbers of images, sort them, and train the neural network with samples.†Folkens said he was particularly excited about the NVIDIA DGX-1 supercomputer used by Cloud View.
"It allows us to process a large number of images, which was previously impossible," he said, and time is also controllable. A visually impaired user gave feedback to Folkens. After using the cloud-view app, he didn't need help from others, and for the first time, he went to the department store.
Wu Yue, CEO of Chaichi Technology, started his business in 2015 and cut into the dialogue robot from the text. His company now has customers such as Didi, Mei Tuan and so on. Wu Yue told Tencent Technology that AlphaGo, which caused artificial intelligence, has a lot of deep learning and machine learning. The process of training Go requires no need to link so many computing processors at any time. This is one of the reasons why some mobile apps can have powerful image recognition.
NVIDIA's business opportunities in artificial intelligence are here. In addition to the above applications, NVIDIA's image processing hardware has many usage scenarios. Its official website introduces a small drone without GPS signal status navigation, which can be used to navigate through forest roads through image recognition.
Equipped with an AI supercomputer based on the NVIDIA Jetson TX1, running deep learning and computer vision functions, even in remote locations where there is no map location information, NVIDIA researchers can navigate like drones.
In the beginning, this design was designed to rescue lost hikers or to observe fallen trees. Now, even without the GPS signal, the drone can still work, and will eventually be used to search for and rescue survivors in the building, patrol the railway in the tunnel, and even check the inventory of large stores.
A variety of new applications constitute a large potential customer base of NVIDIA.
Both cash flow and futureOriginally founded in January 1993, Nvidia is an IC semiconductor company based on the design of intelligent chipset. The most famous product line is the GeForce display card series for games, and the Quadro graphics card series for professional workstations. And the nForce chipset family for computer motherboards.
At present, business revenue from games still accounts for half of NVIDIA's overall revenue. In the first quarter of this year, the company's gaming business (including the NVIDIA GeForce series) revenue increased by about 50% to $1.03 billion, accounting for 53% of the company's total revenue.
Since last year, NVIDIA has begun to provide graphics processors for data centers and autonomous driving, which has become a new growth segment that Nvidia has found after the VR field. At one time, the VR concept was fired for a while, and Nvidia, which dominated the image processing field, was smeared. However, with the VR virtual fire burned, Nvidia has had a period of transformational confusion.
Will the heat wave brought by artificial intelligence be the same outcome as VR? Will NVIDIA be happy again? Will its stock price be artificially high? Some foreign investment companies, such as the drowning water known for short-selling, jumped out to question.
However, revenues from the data center and autonomous driving fields seem to be solid. Data center business revenue in the first quarter of 2017 doubled to $409 million, far better than the expected $318.2 million. Revenue from the automotive business grew 24% to $140 million, better than the expected $132 million. The proportion of revenue in the entire company has increased.
Perhaps the real explosion of the artificial intelligence wave will take time, and the business structure like the NVIDIA in transition is relatively healthy. The revenue generated by the traditional game business enables it to survive stronger than its competitors.
In the short term, the industry has high hopes for the application of NVIDIA products in the field of autonomous driving. Earlier this year, Nvidia announced a partnership with Toyota Japan to launch autonomous vehicles into the market in the next few years.
This means that Toyota has also joined the Nvidia car manufacturer camp. Many automakers have announced that they will be based on NVIDIA technology, which is a luxury, cross-traditional and new energy vehicle manufacturer, including Tesla, Mercedes-Benz, Volkswagen and Volvo.
In the next few years, cars with autonomous driving capabilities will become very common, and NVIDIA's technology will continue to evolve into the automotive industry. At present, Nvidia official told Tencent Technology that “more than 80 partners are using our technology to develop their autonomous vehicle solutions.â€
"NVIDIA GPU-based deep learning is a groundbreaking AI approach that is helping to address the challenges of autonomous vehicles, early cancer detection and weather forecasting. We can now see that GPU-based deep learning will revolutionize each The mainstream industry, from consumer Internet, transportation, to healthcare and manufacturing, we shoulder the mission of the era of artificial intelligence." Huang Renxun said.
Battery Anti-Theft System,Vibration Anti-Theft Function Battery,Gps Anti-Theft Function Battery,Monitoring On Computer And Phone Battery
Wolong Electric Group Zhejiang Dengta Power Source Co.,Ltd , https://www.wldtbattery.com