Today's sharing is the artificial intelligence AI enterprise research report: "2023 artificial intelligence enterprise research report".
The report totals: 138 pages. The full PDF file can be obtained at the end of the article.
A summary of the study is provided below
Since the 20th century information technology revolution, human beings have entered the network era, we have experienced the PC Internet, mobile Internet, and the future will also enter the Internet of Things, Internet of Vehicles, and artificial intelligence era.
You will find that the network era has a very obvious feature, for example, computers, mobile phones are composed of hardware layers (such as CPU), software layers (operating systems) and upper application layers (social, e-commerce, travel, etc.), so the Internet of Things, Internet of Vehicles and artificial intelligence also have similar logic, and we will also find that different levels have given birth to different companies, such as PC Internet: Intel, NVIDIA, Apple, Microsoft, Tencent, Facebook, Alibaba, Google, etc.; Mobile Internet: ARM, Qualcomm, Google (Android operating system, IOS operating system), Apple, WeChat, Twitter, YouTube, Bilibili, ByteDance, etc.
Therefore, the future investment logic is to look for companies that may become the industry leaders at different levels according to the current development situation, in addition, the peripheral companies of these companies should be added.
NVIDIA is the leader in the era of artificial intelligence, as data becomes larger and larger, unstructured data becomes more and more, and the computing power required is increasing.
For example, the ChatGPT that caught fire in 2023, it is a large language model, requiring a large amount of data for training, and the computing power required for training is amazing, and the GPU is more suitable for a large number of parallel calculations than the CPU, his computing time will be much shorter, with a familiar example to explain is: in 2012, Ng Ender led Google Brain to identify a cat from 10 million images, and the whole process used 1,000 computers and 16,000 CPUs.
Then it replaced Intel's CPU with an NVIDIA GPU, using only 16 computers and 64 GPUs to complete the same identification work. Therefore, in the future, GPUs may replace CPUs as the core of AI computing power.
Speaking of which, according to the above logic, then Nvidia has not become the hardware layer of the AI era?
Yes, it is so, but it is more than that, if you look at the stock prices of companies at all levels in the PC Internet era, you will find that although Intel dominates the global computer CPU market, its market value is far lower than Microsoft's market value, and Microsoft controls the operating system of the PC precisely because it controls.
In the information technology industry, there is a law called Andy Beale's Law, which is what Andy gives, bill takes away. Bill's operating system eats up profits from Intel hardware.
Therefore, it is not enough to do hardware, if you can unify the operating system of the AI era, there may be an opportunity to rule the entire AI era.
Excerpts from the original text of the report are as follows:
(This article is for informational purposes only and does not represent any investment advice from us.) For information, please refer to the original report. ）
The report totals: 138 pages. Due to space limitations, only some of the contents are listed.
Featured reports from the public: [Phantom Vision], reply to the keyword "202306" to get the full PDF electronic version