.png)
Published: 2020
Author: D.Y.
Operates on the edge Edge Computing Part 2
Edge computing concept stock integration
Industrial structure
Taiwan’s industry is very important to edge computing because it is no longer an oligopolistic market. Edge computing cannot be solely owned by a few large cloud platform manufacturers, but will be supported by many related supply chains and competitors, especially in Taiwan. Strengths, including server factories, Netcom factories, and other hardware manufacturers are also striving to seize this pie; in addition, semiconductor factories have not forgotten the business opportunities of edge computing chips, including IC design factory MediaTek, wafer foundry TSMC Participate indirectly; in terms of software and hardware integration, there are even more industrial computer manufacturers Advantech and ADLINK.
Introduction to Edge Computing Concept Stocks
1. Chip
1. ARM
As a large silicon intellectual property company, ARM provides solutions that combine CPU, GPU, NPU and open source software to establish a global AI edge computing ecosystem. ARM launched the Zeus platform, a new generation of Neoverse edge computing solutions in 2019, to help improve its machine learning performance with ARM-based CPUs.
Intel is a well-known IC design manufacturer, and naturally it will not be absent from edge computing business opportunities. In the past, Intel’s market share in the server industry was as high as 97%. Therefore, multiple Intel Xeon D series chips have been introduced in edge computing. Xeon D is A system-on-chip (SoC) designed specifically for servers, especially the Xeon D-2000 chip has many edge computing applications. According to Intel, companies currently using this chip include HP, Dell, NEC and other companies.
3. Nvidia Huida (NASDAQ: NVDA)
Nvidia, a major graphics card manufacturer, launched the EGX edge super computing platform. Currently, vendors using this edge computing platform include P&G, BMW, Walmart, and Samsung Electronics. Nvidia provides a complete picture of the EGX platform. We understand that the underlying hardware uses the EGX server, and the software includes EGX Stack, application frameworks and software development kits, container images provided by Nvidia, and other partner products.
Image source: Nvidia official website
Nvidia also announced its partners in EGX. Among them, major vendors include Microsoft. Microsoft owns Azure's current second largest cloud provider. Through the cooperation of EGX and Azure, the cloud and the edge can complement each other and achieve effective serial collaboration. Another thing that we think is worth paying attention to is that Quanta's large data center manufacturer Yunda is also on the list of cooperation.
Image source: Nvidia official website
According to MediaTek’s announcement that it will launch the i700 AIoT platform in 2020, the i700 is mainly used for image recognition and is equipped with two ARM Cortex-A75 CPUs to join the ARM ecosystem.
5. Creative Electronics (TPE: 3443)
Taiwan's IC design service company focuses on special application chip ASICs. According to the company's announcement, AI chip service cases including edge computing and model training accounted for 5% of overall performance in 2018. Although the current proportion is still low, it will grow strongly in the future. Because in edge computing, more ASICs will be required for different industries, the company estimates that the proportion will increase from 7 to 8% in 2019 to double digits, which is expected to drive performance improvement.
Among the chips used in edge computing, most of the chips used in edge computing are undoubtedly the responsibility of the foundry veteran TSMC. According to TSMC’s annual report, in response to the Internet of Things and edge computing, TSMC provides customers with the most competitive and complete variety of RF Special process technologies such as radio frequency, embedded flash memory, emerging memory, sensors and display chips, and a variety of advanced packaging technologies including leading integrated fan-out (InFO) technology. Because edge computing needs to rely on high-performance computing chips, which require the most advanced 7nm process, TSMC estimates that high-performance computing chips will account for 25% of turnover in 2020, becoming a new wave of growth drivers.
2. Industrial Computer
As the world's leading industrial computer, Advantech assists customers to integrate software and hardware and quickly introduce AIOT. In order to achieve this goal, Advantech cooperates with international chip manufacturers such as Intel, Xilinx, and Nvidia. Advantech has launched the Packetarium XLc box-type micro data center, an edge computing device, as a small endpoint, which also uses Intel's processing chips inside.
ADLINK is also an industrial computer factory and has chosen to cooperate with cloud provider Google. ADLINK provides IoT solutions for Google Cloud. With the blessing of Google Cloud, it will greatly help ADLINK’s performance. According to Deutsche The bank estimates that Google Cloud will reach a compound annual growth rate of 64% by 2022. However, it is worth noting that Google Cloud’s market share in the cloud business is close to 10%, which is still lower than the other two AWS and Azure. It should be used for industrial computers. Pay attention to which one has the ability to grab more large customer orders. In addition, ADLINK is also actively cooperating with another 5G edge computing manufacturer Nokia, which will benefit from the US's strategy of actively excluding Chinese-made equipment.
1. Wiwynn Technology (TPE: 6669)
Wivy is a large-scale data center server supplier and a subsidiary of Wistron, a major foundry manufacturer. Wivy, which was listed in 2019, has not reached a new high once its stock is listed. Wiwynn has launched a number of edge computing server platform products to assist telecom operators in setting up edge nodes at base stations or computer rooms in various locations to perform "far" edge computing. Many of them are co-launched with Nokia, like OCP OpenEdge servers suitable for outdoor use, but it is worth noting that according to the news, most of Nokia's high-end products are cooperating with ADLINK, and the cooperation with Wisdom has a low gross profit but a large amount Orders.
Three, cloud server
Quanta's cloud server business is mainly in charge of its wholly-owned subsidiary Werda. Werda has also launched a variety of edge computing servers and cooperated with Rakuten Telecom and Intel. In the future, Rakuten Telecom base stations in Japan will install Quanta's edge computing servers to provide users with more timely services. If more telecom operators enter the base station market smoothly, this will greatly increase the amount of servers and increase the gross profit margin. , Because the server has a higher gross profit margin than laptops, it can break away from the fate of the foundry of "three to four gross".
4. Netcom equipment
Netcom equipment manufacturer Cisco is the first company to propose Fog Computing. Fog computing is a concept between edge and cloud computing. Therefore, it has also launched a number of edge computing related products, such as the launch of Cisco Live at the 2019 Cisco annual conference. Cisco Catalyst IE3400 Heavy Duty series switches and Catalyst IW6300 Heavy Duty series base stations are mainly waterproof and dustproof equipment that can still be used in harsh environments. These equipment will also promote the development of the Internet of Things and edge computing applications.
Dell has launched a number of edge gateways for IoT. The Edge Gateway series of gateways allow users to perform analysis and calculations at the edge. In addition, almost all of these gateways use Intel processor chips. Dell has more Telecommunication companies have become partners such as Orange and Telenor in Europe, which will effectively extend the technology and equipment to practical applications and expand the ecosystem.
HPE, a spin-off from HP, is responsible for enterprise-related server products and launched the HPE EDGELINE converged edge system, which provides enterprises with setting up their own data centers on the edge. HPE is considered an advanced among edge cloud computing vendors and has been applied In factories and drilling platforms, it will enjoy the advantages of advanced players to compete with Dell and Cisco.
Five, cloud platform
Many people think that cloud giants and edge computing are mutually exclusive, but compared to the confrontation with edge computing, the three major cloud companies choose to expand their computing capabilities to the edge, and effectively enhance the collaboration between the cloud and the edge, and make use of The person can have more experience.
Amazon's cloud AWS has the highest market share. AWS has also launched several products specifically for IoT edge computing, including AWS IoT Greengrass and Amazon FreeRTOS. AWS IoT Greengrass allows the edge to operate offline, store data on the edge device, and perform Analyze and process, and allow customers to adjust the frequency and data uploaded to the cloud by themselves. At this time, the cloud is responsible for management, analysis, and durability storage. AWS has recently cooperated with Verizon and announced that it will cooperate with Vodaphone in Europe in the future. South Korea will also cooperate with SK Telecom and KDDI, which is the fastest commercial transfer of 5G. This will accelerate AWS's business application expansion on the edge.
Microsoft’s cloud platform Azure and AWS have launched multiple edge computing products Azure IoT Edge, allowing customers to transfer part of the cloud computing load to the edge to reduce transmission pressure and response time. It is worth noting that Microsoft has also launched the move to Azure machine learning. Edge-side services, which will effectively speed up data processing and analysis. Microsoft chose to cooperate with Nvidia to jointly develop collaborative AI computing between the cloud and the edge. With the blessing of AI chip manufacturer Nvidia, Microsoft will have greater momentum in improving performance.
3. Alphabet Google (NASDAQ : GOOGL)
Google's cloud GCP also launched a self-designed chip ─ ─ Edge TPU tensor processor chip. This type of chip pursues power saving and uses low power consumption to allow AI to operate at the edge. Even Google, known for its software in the past, has personally jumped down. The purpose of making ASIC chips is to seize the business opportunities of AIoT in the future of the Internet of Everything. According to Taiwan's leading industrial computer manufacturer, Ling Hua, they have products that are cooperating with GCP edge computing.
Six, telecommunications company
Telecom operators are a very important link for manufacturers to push edge computing to commercial use. One is because they can add servers to existing base stations, making them the edge closer to users, even in the 5G era. The telecommunications industry also needs to install more small base stations to meet the signal coverage. If these small base stations can be combined with servers to have computing performance, they will effectively promote the popularization of AIoT. In the development of self-driving cars, This type of edge computing device has become a crucial part, enabling self-driving cars to respond more timely than cloud computing.
Verzion, a US telecommunications company, not only cooperates with Amazon, but also cooperates with Intel and Taiwan's major industrial network security company Liduan to launch the uCPE network edge computing platform, gradually expanding the ecosystem.
2. Chunghwa Telecom (TPE: 2412)
Chunghwa Telecom and a number of Taiwanese manufacturers have jointly developed a 5G smart edge data center. The actual application will still be announced after the completion of 5G. However, Chunghwa Power is also the operator with the most 4G base stations among the five major telecommunications industries in Taiwan. The current 5G base station construction also relies on 4G base station reconstruction, which will be a great weapon for the deployment of edge computing IoT.
summary
In edge computing, various vendors are constantly seeking partners, expanding the ecosystem, or joining a larger ecosystem, tapping the latest and most widely used applications, and identifying larger edge business opportunities. Therefore, all vendors are actively developing. Therefore, when we look at investment targets, we should not only look at individual companies, but also pay attention to the competition and cooperation of the larger ecosystem.