On April 16, the U.S. government issued a ban on ZTE’s exports and prohibited US companies from selling components to ZTE. This ban evokes the panic that lies deep within the hearts of China's telecommunications and computer makers. If Americans do not sell chips, what should we do?
Although there are still many parts imports will be affected by the ban, but the chip is always the core one. At present, the domestic chip occupancy rate is extremely low, and more than 90% of the domestic high-level chips are dependent on imports, which cost more than US$200 billion annually. Once foreign countries have completely blocked the chip trade, the domestic technology industry will collapse.
However, in the face of the ferocious chip crisis, there are not many domestic pessimistic voices. On the contrary, many people believe that the "ZTE Incident" may force changes in China's domestic related industries. This may be an opportunity for entrepreneurs and investors.
The reason for this confidence is that, first, the domestic semiconductor industry has made great progress. Currently, domestic manufacturers can already mass-produce processors with a 28-nm process, although compared to Qualcomm and TSMC, the 7-nm manufacturing process has several years or even more than a decade of gaps. However, compared to the late start of 4,50 years, Closer to many.
Another point must be attributed to the changes in the chip industry itself. AI chips are replacing traditional chips. As long as you keep an eye on foreign news, you will find that the most reported in the chip field is technology giants such as Amazon, Apple, and Google. Amazon is designing AI chips for its Echo speakers, and Apple and Google have developed and applied their own AI chips. Even Facebook has announced high-profile recruitment of chip engineers.
On the contrary, the traditional chip makers ARM and Intel now have a day that is not as good as one day...
All this shows that the dominant chip industry in the future may not be a chip company, but a technology giant with strong Internet colors such as Apple, Google, and Amazon. Then, the domestic BAT Big Three, compared to foreign counterparts, did not fall out of the air, whether they have the opportunity to share a slice of the future AI chip trend? If China really has its own "Qualcomm," then it bothered now The chip problem of the industry is naturally solved
In fact, BAT has already been laid out in the field of AI chips. In order to facilitate readers to understand the layout of the three companies, here we briefly introduce the trend and direction of the development of AI chips.
At present, there are four main schools in the field of AI chips, namely GPUs, FPGAs, ASICs, and partial brain-like chips.
Since NVIDIA took GPUs from graphic computing to general-purpose computing, GPUs have become synonymous with high-performance computing by virtue of the inherent advantages of parallel computing architectures, and have once become the hope of the AI ​​industry. But now, in the field of AI chips, GPUs are being marginalized. On the one hand, Nvidia and AMD monopolize the GPU market; on the other hand, the GPU is too expensive and it consumes electricity.
The FPGA is an architecture created by Xilinx. It is called “Field Programmable Gate Arrayâ€. Users can change their functions by updating FPGA configuration files. It is very flexible. Coupled with low power consumption and low power consumption, FPGAs have replaced GPUs in recent years. However, the disadvantages of FPGAs are poor peak performance and the need for users to understand hardware and software programming languages ​​at the same time.
Therefore, there has been an ASIC architecture developed for the mass-consumption field. The full name is an ASIC designed specifically for AI. Although the time and technical requirements are high, the performance is excellent. With the SoC+IP module solution, it is not without flexibility. After using IP modules designed by specialized vendors, the performance is not too bad, and the comprehensive advantages are the greatest. Now many internet giants are rushing to develop and many of them are such chips.
The brain-like chip is a system that is capable of simulating the human's ability to sense, understand, act, and communicate by reducing the computational power of the brain. However, the development of this kind of chip is subject to the study of the brain, so it is relatively unpopular.
Understand the development chip of AI chip, we go back and look at the layout of domestic manufacturers.
First of all, as mentioned above, GPUs are monopolized by the United States. Therefore, unless major breakthroughs are made in technology, it is almost impossible for domestic manufacturers to break through these two aspects. While FPGAs are 90% owned by Xilinx and Altera, Microsemi and Lattice partition the remaining 10%. Although 4 companies have jointly built a patent wall consisting of 9,000 patents, other companies can be used for compiling. ASICs and brain-like chips are also optional.
From the perspective of BAT's three major manufacturers, Baidu has the earliest layout of AI, but its main direction is not AI chips, and its main investment is in applications such as image, voice, and driverless applications. In February last year, Baidu set up a dedicated hardware department through its wholly-owned acquisition of Raven Technologies. On the chip side, Baidu is mainly cooperating with other companies to develop and develop application-level content. In February of this year, Baidu purchased Lightelligence, an optical AI chip start-up company, as an official entry into the AI ​​chip field.
Tencent and Baidu are similar, in its AI strategy, the proportion of the underlying chip is not large. Currently, Tencent’s investments in the chip field include Barefoot Networks, which had previously invested in Alibaba.com, and its features are the creation of a new programmable PISA architecture. However, its chips are mainly used for communications equipment rather than terminal equipment.
In addition, Tencent also invested in Bit Continental. Since the mining equipment needs strong computing power, Bitland has gained a lot of experience in developing chips during the development process. At present, the company has developed ASIC-based TPU, which is one of the hottest AI chip companies in China.
Compared with the first two companies, Ali is a lover of investment chips, and its investments in transit, cambrian, and energy-capability are all oriented towards ASICs or NPUs that foreign companies have not monopolized. In 2016, it invested in Zhongtianwei, which has already produced more than 500 million SoC chips in the past two years. In the Cambrian period, the Cambrian also became famous for developing the artificial intelligence chip Kirin 970.
It is worth noting that Ali’s investment and R&D is not a widespread gambling and luck but a clear strategy – investing in Zhongtian Weihe Investment Cambrian can open up the whole industry of “processor + operating system + terminalâ€. chain. In conjunction with Ali's existing cloud services, it is finally possible to open up such a channel as "operating system + processor chip or application chip + algorithm + terminal + application + cloud" and ensure its unimpeded.
In the long-term investment process, Ali, who is the coordinator, has also obtained a large amount of technology accumulation. Therefore, Alibaba's research institute, Alibaba Dharma, has just entered the chip industry after being established for half a year. Just two days after ZTE was banned, Dharma announced that it is developing a neural network chip called Ali-NPU.
According to reports, according to the design, the performance of Ali-NPU will be 10 times that of the mainstream CPU and GPU architecture AI chips on the market, while the manufacturing cost and power consumption are only half, and the cost performance is more than 40 times. Slightly regretful, this is a chip designed specifically for AI. It cannot compete with CPUs and GPUs for general-purpose computing. Although it is in line with the direction of the times, it is still unable to solve the domestic chip dilemma in the short term.
Overall, China has been warming up in the development of chips. Now is the time to play. The crisis may have forced China's domestic industries to change. This may be an opportunity for entrepreneurs and investors.
More stable performance
Disposable ecig have a completely enclosed design, reducing the need for charging and replacing cartridges. The no-charge design also reduces the occurrence of faults. It is understood that with rechargeable e-cigarettes, each cartridge needs to be charged at least once and the battery efficiency is extremely low, while the design of disposable ecig can solve this problem very well.Disposable ecig are very easy to carry on daily trips, no need for any operation in the use of airflow induction, direct suction can be out of smoke, meaning that you just need to suck directly on the hand can be the main feature of the small cigarette is practical portability is the main, its appearance is very small, no longer like the previous big smoke box mechanical rod nowhere to put.
The market outlook for disposable ecig
There has been a controversial topic regarding the use of disposable ecig. According to scientific studies, the long-term use of e-cigarettes can cause harm to the human body, no less than tobacco. Therefore, it is important not to overuse either type of e-cigarette - after all, smoking is bad for your health. Specifically, the PMTA requires that any new tobacco product be approved by the FDA for legal promotion and that the FDA consider all aspects of the product to be beneficial to public health. Our products can always be sold under the supervision of the PMTA at cost and are prohibited for sale to minors. We have strict guarantees regarding the quality of our products.
Disposable E-Cig,Disposable Vape Pen Oem,Best Disposable Vapes,One-Time Use Vape
Shenzhen MASON VAP Technology Co., Ltd. , https://www.disposablevapepenfactory.com