Top 10 hot trends in semiconductor industry in 2022
Время обновления: 2021-12-27 15:05:25
In 2022, the global semiconductor market continues to be volatile. Jotrin has analyzed topics around chip processes, cross-border core building, DPU startups, ray tracing, AI mobile photography, DDR5 memory, third-generation semiconductors, mobile robotics, fast-charging applications, price increases and stock-outs, and meta-universe concepts. (Note: The trends described in this article are for readers' reference only and do not constitute any investment advice)
Into 2022, the global economy is still under the shadow of the new crown epidemic, but the semiconductor industry continues to grow. early August 2021, the World Semiconductor Trade Statistics (WSTS) forecast - the global semiconductor industry in 2022 will be 10.1% of revenue. However, in early December 2021, WSTS revised the figure to 8.8%, and the global semiconductor market size was adjusted to $601.4 billion.
It is worth noting that the year-on-year growth of global semiconductors in 2021 is 25.6%, and the estimated growth rate in 2022 is 17 percentage points lower than that in 2021, but still shows a positive growth trend. At the same time, industry participants also generally agree that the global trend of tight semiconductor capacity will continue throughout 2022.
In such an environment, Jotrin analyzed trends around hot topics such as chip process, cross-border core making, DPU startups, ray tracing, AI mobile photography, DDR5 memory, third-generation semiconductors, mobile robots, fast charging applications, price increase shortages, and meta-universe concepts.
4nm process chip batch commercialization
Although the complexity and cost of advanced processes below 7nm are climbing significantly, it is still very important for those manufacturers who pursue the ultimate chip performance. Because in the semiconductor process process gradually approaching the physical limits of the limitations of the chip development must be through changes in transistor architecture, back-end packaging technology or material breakthroughs, in order to continue to achieve the purpose of improving performance, reducing power consumption and reducing the size of the chip.
At the end of 2021, two top cell phone SoC chip makers, MediaTek and Qualcomm, have announced the launch of their flagship SoC platforms - Tiangui 9000 and Snapdragon 8 Gen 1 - manufactured using TSMC's 4nm and Samsung's 4nm processes. will run into the 4nm era.
5nm chips were actually mass-produced in 2020, with Huawei Kirin 9000, Qualcomm Snapdragon 888, and Apple's A14 and A15 chips, all manufactured using the 5nm process. On top of that, TSMC and Samsung are moving towards 3nm chips. Before the 3nm mass production, the 4nm process effectively fills the gap between 5nm and 3nm.
TSMC's N4 process platform is based on the N5 platform. The new process is further enhanced in terms of speed, power consumption and density, and remains compatible with N5 in terms of design rules, SPICE and IP as well, to facilitate a seamless transition between 5nm and 4nm and to form a complement when 3nm is not in mass production. According to previous plans, N4 will start trial production in Q4 2021 and achieve mass production in 2022.
Previously, Samsung regarded its 4nm process 4LPE as an evolutionary version of its 7LPP process, and added a low-power version 4LPP on top of it, which is said to achieve a 5% performance improvement and 10% power reduction compared to 4LPE, and is "the best PPAc (power, performance, area, cost) before the full surround gate transistor (GAA) architecture. ) of the 5th generation EUV node process", is expected to achieve mass production in 2022. Of course, this is also the last Samsung advanced process technology using finned field-effect transistor architecture (FinFET) process, starting with the 3nm process, the company will fully adopt the GAA process.
In general, compared to the 3nm process, the 5nm process based on the improved 4nm process technology is relatively stable and mature, the production cost is relatively low, belongs to the transitional process process, 3nm is the focus of the future competition between related companies.
But on the other hand, we also have to admit that the challenges faced by advanced processes is huge, which is why the current trend of development of integrated circuit process technology, is gradually from a single pursuit of size-dependent advanced processes, to advanced processes (More Moore), non-size dependent features of the process (More than Moore) and advanced packaging three dimensions in parallel, small chips ( Chiplet), heterogeneous integration of system-level packaging (System-in-Package), 3D stacking and other new technologies are emerging for a variety of reasons.
Cross-border self-research chip trend will be more obvious
Standard chips can no longer meet the actual needs of Apple, Amazon, Facebook, Tesla and other top global technology companies, they are developing their own customized chips. Such a trend is also reflected in China, and currently, the top 10 domestic electronics manufacturing companies have the ability to design their own chips, OPPO, Xiaomi, Midea, and even Baidu, Alibaba and other companies have been building their own teams.
The main benefits of building their own chip design team, one is the hope that the fate in their own hands, especially in the last two years in the face of the outbreak of the industry chain crisis, this strategy has withstood the test, received the effect; secondly, OEMs have systematic knowledge, know what type of chip they need; again, is the formation of a certain amount of scale, it can reduce procurement costs. In addition, companies can also develop their own independent technology, do some differentiation, proprietary technology and products.
Surge C1 is Xiaomi's self-developed ISP image processor, the person in charge has said, "If we do not invest in chip design, then the future may not hold the key to the digital world. Xiaomi will use this as a starting point to get back to the core SoC device design for cell phones."
Looking at autonomous driving again, the automotive industry in the traditional sense has a complex supply chain, with automakers at the top and multi-layer suppliers such as raw material suppliers, secondary suppliers and primary suppliers providing them with various hardware and software. But with the emergence of smart driving and a large number of new car makers, they are starting to work directly with software/hardware design companies, thus bypassing some of the steps in the traditional supply chain. This means that market players that provide software and digital technology were originally in the Tier 2 and Tier 3 tiers, but may soon jump to the Tier 1 position under current trends.
But in our view, these companies are also facing challenges, including: what kind of volume they can achieve, the ability to design their products, whether the price/performance ratio can meet the demand, etc. At present, most of the companies are in the "burning" stage at the beginning, and it is difficult to develop. Therefore, relatively speaking, large companies will act more aggressively because they are in a better financial position, and they often receive government subsidies or other support for investing in semiconductors in the domestic environment.
DPU becomes the newest SoC startup
In October 2020, Nvidia officially named its Mellanox-based SmartNIC solution as Data Processing Units (DPU) and called CPU, GPU and DPU the "three pillars of future computing". Since then, the DPU concept has become a hit and many competitors have come from all over the world.
The emergence of DPU is not a surprise, it is another milestone of heterogeneous computing. As we all know, the slowdown of Moore's Law has made the marginal cost of general-purpose CPU performance growth rise rapidly, but the demand for computing has always shown explosive growth, and this phenomenon of "scissor difference" between performance growth and data growth is an important reason for the development of dedicated computing chips represented by DPUs.
The global demand for computing power doubles every 3.5-4 months, far exceeding the current growth rate of computing power. Driven by this, the global computing, storage and networking infrastructure is also undergoing a fundamental shift, with some complex workloads that cannot be handled well on general-purpose CPUs. As a result, DPUs are gradually replacing CPU chips in network data processing, network security, and network interconnection protocols to reduce the burden on CPU cores and improve computational efficiency.
Although a "new species" in the digital infrastructure, DPUs are not born out of nowhere, but have been in the making for a long time. From the early offloading of network protocol processing, to the subsequent addition of various functions such as AI, security, storage and network on the basis of SmartNIC, it can complete performance-sensitive and universal task acceleration, better support CPU and GPU upper layer business, and leap to become the central node of the entire network and a new generation of computing chips. It can be said that the DPU chip has previously been in function, but only "lack of name".
Intel, Broadcom, Nvidia, Ceres, Marvell, all familiar international companies, are important players in the DPU market, but in the past two years, Chinese DPU start-ups, represented by Centric Harness, Core Qiyuan and Nebula, have also joined the hot DPU track, and for a while, DPU has become the hottest concept for SoC start-ups in 2021.
According to International e-Business, China's DPU market size will be close to RMB 20 billion by 2023; meanwhile, other analysts also predict that China's DPU market size is expected to exceed USD 3.7 billion, or about RMB 24 billion, by 2025.
In the future, the volume of DPUs used in data centers will reach the same level as data center servers, and it can even be said without exaggeration that "each server may not have a GPU, but there will definitely be one or several DPU/IPU cards, which will be a 100 billion dollar market."
DDR5 memory batch commercialization, price drop in 2-3 years
In July 2020, the JEDEC Solid State Storage Association officially announced the DDR5 SDRAM standard (JESD79-5). DDR5 memory has twice the pin bandwidth (frequency) of DDR4, changed to an independent dual 32bit channel, the main frequency rises from 2400MHz of DDR4 to 4800MHz, about 50% higher than DDR4 3200MHz frequency The total transmission bandwidth has increased by 38%, and will reach 6400MHz or even around 8400MHz in the future.
Although 2021 is the first year of DDR5 commercialization, the current price of DDR5 remains high, and in the second half of 2021, Kingston, SanDisk, Jiahe Jinwei, Corsair, Micron, TEAMGROUP, Innodisk, MaxSun and other manufacturers released DDR5 memory sticks. Take TEAMGROUP's memory stick kit for example, its 4800MHz 32GB DDR5 is priced at $399.99 (about Rmb 2,578). The current price of 32GB DDR4 memory sticks on mainstream e-commerce platforms is mainly between RMB 1,000-1,500. With the same capacity, DDR5 is about two-thirds to twice as expensive as DDR4. Some industry insiders say that DDR5 memory will only gradually drop in price when it accounts for 30% of the overall memory market.
Observing the characteristics of each previous generation of memory conversion, there will generally be a 2-3 year conversion process between two adjacent generations, during which the market share of new products will become higher and higher. Based on this feature, we expect DDR5 memory prices to see significant downward adjustments in the second half of 2023 to the second half of 2024.
As DDR5 products will gradually enter mass production, NAND Flash stack technology will move beyond 200 layers. In terms of DRAM, Samsung, SK Hynix and Micron will gradually mass produce DDR5 products, while 5G cell phone demand is also stimulating a continued increase in LPDDR5 market share. In addition, SK Hynix and Samsung will gradually mass produce 1anm process products, and this market visibility is expected to increase quarter by quarter in 2022.
In terms of applications, on October 28, 2021, Intel released the 12th generation Core desktop CPUs, bringing DDR 5 memory to the PC DIY market. Of course, to experience the full power of Intel's 12th generation Core, it must be paired with come DDR5 memory. As DDR5 is facing a shortage crisis, some consumers are currently unable to experience the advantages that DDR5 memory brings to the PC.
Third-generation semiconductor fabs increase 8-inch capacity
As the first commercial third-generation semiconductor materials, silicon carbide (SiC) and gallium nitride (GaN) have attracted much attention in recent years. By 2025, the global market for power discrete devices and modules will reach $27.4 billion, with SiC and GaN increasing their market share to 17 percent.
Global wafer makers are actively promoting the commercialization of 8-inch SiC wafers. SiC manufacturing difficulties are concentrated in substrate growth, substrate cutting and processing, and oxidation processes, and the SiC substrate link holds the industry chain discourse, with some downstream epitaxial and device manufacturers acquiring SiC substrate plants, and silicon wafer makers cutting into the SiC substrate business.
In recent years, Cree, Infineon, SiCrystal, II-IV, Nippon Steel & Sumikin and Dow Corning have expanded their 6-inch production lines, and are actively promoting SiC to 8-inch development. Among them, Cree, II-VI, SiCrystal and ST already have 8-inch SiC substrate technology. The domestic manufacturers are upgrading from 4-inch to 6-inch SiC substrate link, with CEC 55, CNMC, Sanan Optoelectronics, China Resources Microelectronics, and Jack Tower already having 6-inch SiC production lines.
In addition to automotive inverters, automotive OBCs, slow charging chargers, and fast charging piles can all use SiC products. This leads to the current SiC production capacity cannot meet the market demand, and the big manufacturers lock the substrate capacity in advance by virtue of expansion/merger, such as Wolfspeed (under Cree) and Infineon, ST, ON Semiconductor, SiCrystal (under Rohm) and ST, GTAT (under ON Semiconductor) and Infineon have signed long-term supply agreements respectively.
Epitaxial wafers are a very critical part of GaN, and their epitaxial wafers are usually made of heterogeneous substrates, such as sapphire, SiC, silicon (Si), etc. In theory, GaN homogeneous substrate is the best substrate for growing GaN epitaxial layer, but the traditional melt method cannot be used for the growth of GaN single crystal, so the progress of GaN single crystal growth is slow.
GaN-on-SiC has better performance but high price, but the current size is limited to 4-6 inch wafers; GaN-on-Si has faster growth speed and easy to expand to 8 inch wafers, but the performance is slightly inferior to GaN-on-SiC.
GaN-on-Si is widely used in power electronics, and foundries such as SunEdison, Sembcorp, Silicon Industries Group and IDM factories such as TI, ST, Infineon, ON Semiconductor, Innosec, China Resources Microelectronics and Silan Microelectronics are making efforts in this track; GaN on SiC is suitable for RF applications, and the heat dissipation of SiC substrate is better, but its wafer size is small and has not yet exceeded 6 GaN on SiC is suitable for RF applications and has better heat dissipation on SiC substrates. GaN on SiC track suppliers are Wolfspeed, Dow Corning, Roma, II-VI, Nippon Steel & Sumikin, Norstel (acquired by Chinese capital), Tianke Heta, Shandong Tianyue, etc.
It is worth noting that on June 5, 2021, Innosec GaN production base was officially put into operation. By the end of 2021, the GaN wafer production capacity of Innosec will reach 6,000 wafers/month, and when the project is fully put into operation, the annual production capacity will reach 780,000 wafers and the annual output value will exceed RMB 10 billion. We predict that GaN will be accelerated for commercial use in China starting from 2022.
The growth of applications such as electric vehicles, 5G communication and data centers will also push up the production capacity of SiC and GaN wafers. At present, SiC and GaN wafers are mainly limited to 4-6 inches, with the head supplier in 8-inch wafers, the third generation of semiconductor 8-inch wafer capacity will increase in 2022, the trend of wafer makers to increase 8-inch capacity will continue in the next few years.
Ray tracing has become a must in GPU applications
The ray tracing mentioned here refers to real-time ray tracing. Real-time ray tracing has been very popular in the graphics computing field for the past two years, even though ray tracing technology has been in existence for quite a few years. Ray tracing is now considered by the industry to be a must-have technology for graphics computing, as it can create a more realistic light-shadow relationship between 3D objects in the virtual graphics world, making the picture visually more realistic.
In the desktop, data center and professional visual fields, the main push for ray tracing comes from Nvidia. both AMD and Intel's GPU products are actively following up in 2021. The mobile segment, on the other hand, will be a little later because of the insatiable arithmetic demand for ray tracing. However, Imagination is already ready for ray tracing technology implementation in mobile in 2020 in terms of IP; Arm, MediaTek and other players are currently doing pre-deployment in the market.
At the application level, it goes without saying that games and professional vision are important scenarios for ray tracing pre-application. But what ray tracing really represents is a class of simulation and emulation technology that maps the real world in the virtual graphics world. Among the popular concepts in the past two years, "digital twin" and "meta-universe" both put high demands on the physical simulation of the real world, including light, particles, liquids, materials, springs, cables, and other physical characteristics of simulation.
A representative example of the application of ray tracing to the digital twin is Ericsson's use of Nvidia's Omniverse to create a digital twin of a city. This virtual city has physical-level accuracy in materials such as buildings, vegetation and foliage, enabling accurate simulation of signal reflection strength. Using ray tracing technology, it is possible to calculate and visualize the quality of the 5G signal at every point in the city, ultimately designing an efficient and reliable network.
The simulation and emulation technologies represented by ray tracing will further drive the development of technologies and applications such as graphics world, digital twin, and meta-universe.
AI becomes standard for cell phone photography and deeply penetrates more imaging/vision markets
AI is probably the hottest technology direction for computational photography right now, and cell phone photography is one of the applications that will benefit the most. Not only in the direction of photography, automotive, industrial, security and other markets, AI in the field of machine vision is becoming an increasing proportion. AI-specific processors or units, are entering the image processing process across the board.
The hardware level is quite representative of the trend: cell phone chip makers generally began to image processing process, ISP and NPU to do "fusion", and to do optimization for the communication between the two units. 2020-2021, Huawei, Samsung, MediaTek and other chip manufacturers are doing this effort. AI has revolutionized the hardware in vision systems and has impacted the entire industry.
In terms of cell phone photography applications alone, AI's involvement in photography and imaging is also much more than just the icing on the cake functional features of face changing and beauty, but plays a role in key components of photography and imaging, such as deep involvement in processes such as automatic white balance for photos, noise reduction for dark photography, anti-shake for video shooting, HDR high dynamic range, and local lighting. These features are now benefiting from the development of AI hardware technology, which is also starting to fully penetrate from photo taking to video shooting - AI video shooting will become standard in smartphones.
At the same time, AI photography technology is still being explored, such as a research team using the same scenes taken by cell phones and DSLRs as a neural network mapping to train a network that can obtain "DSLR-level" imaging quality.
The new year is expected to see more applications and features of AI in photography; the main battlefield of image quality is even gradually shifting to back-end processing as a result. Continued penetration into more machine vision applications is a core component of AI technology development.
Mobile robots conquer the market with "AI
The world has a long history of innovative research on robotics applications, which have been widely used in the automotive industry, 3C electronics, logistics and other industries. It was not until the arrival of the "black swan" New Crown pneumonia epidemic that people rediscovered the business value of robots to replace human beings in "no-touch" behavior, which led to a flurry of entrants from different fields and a proliferation of new applications for different scenarios. The mobile robotics boom continues to rise. According to growth trends, the global mobile robotics market is expected to exceed $25 billion in 2022, with a year-over-year growth rate of about 30%, the highest in the next five years!
What is it about mobile robots that is gaining such great expectations in the market? It has to do with the rise of AMR applications.
Autonomous Mobile Robot (AMR) is a robot that has the ability to understand and move independently in a scenario environment with the three main advantages of flexibility, intelligence, and compounding to improve operational efficiency, increase speed, ensure accuracy, and increase safety of internal transportation processes.
AMR differs from its predecessor, the Automated Guided Vehicle (AGV), in that its applications do not depend on tracks or predefined routes, but use sensors, artificial intelligence, and machine learning technologies to calculate path planning and use navigation technologies such as collision avoidance to slow, stop, or re-route around obstacles and then continue on with the task. This reflects the flexibility and intelligence of AMR robots.
AMR is also known as a "composite robot" because it can bring together a variety of top module solutions to expand multiple functions. From practical examples, AMR can replace AGVs to meet the basic needs of logistics transportation in large-scale scenarios such as large manufacturing plants and third-party logistics transportation by assembling top modules such as display screens, interactive modules, and sterilization facilities, and AMR can be extended to segmented scenarios such as medical and health care, supermarkets, and restaurants. Typical cases include welcome robots in shopping malls, food delivery robots in restaurants, logistics robots in parks, and temperature measurement/disinfection robots in buildings. ...... It is easy to see that from traditional AGVs to AMRs, the "vehicle" attribute of mobile robots is being weakened, and the "vehicle" attribute is becoming more prominent. In turn, the business value of "robot + AI" is more prominent.
Fast-charging industry accelerates technological innovation and application landing
As people's dependence on electronic products gradually increases, "battery anxiety" has also become common. From small smartphones, TWS headphones, electric toothbrushes, rechargeable batteries, to large laptops, tablets, and even electric cars, users want faster charging, more energy-efficient power consumption, and longer battery life. In view of this demand, without revolutionary changes in battery technology, fast charging/super fast charging technology is back in vogue, accelerating development in both technology innovation and industrial landing.
In terms of technological innovation, fast charging products are developing in the direction of high power, small size, low cost, multi-port, wireless, etc. Accordingly, new requirements for the degree of integration of power adapters, circuit topology design, anti-electromagnetic interference, heat dissipation management and other performance have been put forward. At the same time, fast-charging manufacturers are also more concerned about the integration of product quality, structure and cost control advantages of the integrated solution.
In addition, GaN switching characteristics will be far better than MOSFET, which provides the feasibility of high frequency to reduce the size of the charger. And with the improvement of GaN shipments, GaN cost is also gradually optimized, the price of some models has been close to MOSFET. future GaN has a good chance to become the mainstream of fast charging chargers.
Of course, the shortage of power class chip market in 2021 to a certain extent affect the development of the fast charging market. But in the long run, for the factor of consumer demand for fast charging, the future demand for fast charging will certainly increase sharply, especially recently some cell phone manufacturers canceled the distribution charger, which brings huge potential demand for the whole industry.
In terms of application landing, fast-charging technology has been gradually covered from the three segments with the highest popularity, namely cell phones, notebooks and tablets, to emerging markets such as TWS headphones, new energy vehicles, power tools, personal care and outdoor power, which is expected to expand market capacity rapidly.
Especially under the trend of carbon neutrality and environmental protection, electric vehicles have been a clear trend in the future. We estimate that the total number of electric vehicles will reach 5.4 million units in 2021, accounting for about 5% of the overall vehicle market shipments. And the goal is that by 2030, the proportion of electric vehicles in the overall car market will reach about 30%, and it seems that the speed may be faster, which is equivalent to the next ten years the number will show more than 6 times the growth.
However, the popularity of fast charging of electric vehicles will be significantly lower than that of consumer electronics. This is because temporarily constrained by the power battery, control and thermal management and other technical constraints, and the lack of domestic charging facilities construction-related regulations and standards, but also need the power supply sector, community centers and other links to help. Therefore, the popularity of fast charging of electric vehicles will become a hot area of the industry chain in the next five years with multiple links together.
Shortage and price increase may cover the whole year of 2022
Although "shortage" has become a "new normal" in the electronics industry, but the industry is still very sensitive to the shortage and price information. Looking ahead to 2022, we believe that: "shortage" dilemma can not be completely alleviated in a short period of time, structural shortage will be throughout the year 2022, may be concentrated in automotive MCU, power components, power management chips and other popular categories; as for the "price increase" is The specific performance of the volatile market, belongs to the regional, some categories of price increases, it is recommended that the industry chain up and down by strengthening communication, sharing market information, sharing analysis of forecasts and other means to cope.
The primary factor causing the shortage of the market is the volatility of the new crown pneumonia epidemic situation. Optimistically, if the global outbreak is fundamentally controlled in the second half of 2022, the global supply of mature process semiconductor components is expected to be largely relieved by the end of the year. Nevertheless, this is longer than originally expected until Q2 2022, and the long and short term conditions will also be extended, so we hope the industry chain will continue to pay attention to the impact of the epidemic on the overall electronics supply chain.
On the contrary, the demand side, under the normalization of global epidemic prevention, the market demand is growing steadily is the industry consensus. For example, the continued strong demand for chips for applications such as 5G, AI, IoT, and advanced driver assistance systems (ADAS) can make up for the signs of slowing market demand for some end products, and the overall market demand can still grow steadily in 2022.
As a result, Jotrin suggested that the global upstream supply chain needs to "open source" to cope with the dual pressure of capacity shortage and rising demand. Specifically, "open source" refers to actively expanding production capacity, and "cut back" is to adjust production capacity and product mix, so as to survive the special situation and meet future market opportunities.
For component distributors, it is believed that the volatile market in the past two years has given them enough time and experience to adapt to the rapidly changing uncertainties. In the coming 2022, distributors will continue to play their business agility and flexibility through active cooperation with original manufacturers, appropriate amount of stocking, and active use of alternative solutions to assist upstream and downstream to build up a stable supply chain system.
Meta-universe concept products will be difficult to implement in 2022
In 2021, a "Metaverse" frenzy swept the world - in April, Byte Jump invested 100 million yuan in the metaverse concept company Code Qiankun; in May, Facebook changed its name to Meta and said it would transform into a metaverse in 5 years. In August, VR startup Pico disclosed that it had been acquired by Byte Jump for up to 9 billion yuan; Microsoft announced its enterprise meta-universe solution at Inspire Global Partner Conference in August; VR startup Pico disclosed that it had been acquired by Byte Jump for up to 9 billion yuan... ...For a time, global high-tech companies seem to see a huge blue ocean market and have started the initial layout of metaverse industry.
Meta-universe itself is not a technology, but an idea and concept based on extended reality technology to provide immersive experience and digital twin technology to generate a mirror image of the real world. Its arrival requires the integration of different new technologies, such as 5G, 6G, artificial intelligence, big data, etc., emphasizing the integration of reality and reality.
In August 2021, Nvidia is using the 14-second "digital double" version of Jen-Hsun Huang at the launch event to "show off its muscles" for Omniverse, the cloud universe platform that is considered the current "meta-universe infrastructure". ". According to NVIDIA, AR is the way for AI to enter the real world from the Omniverse, while VR is the "wormhole" used to enter the Omniverse.
The Omniverse is designed to build a persistent virtual shared space while still maintaining the perception and experience of the real world. VR/AR devices are seen as the best metaverse "entrance" for users, in other words, the popularity of VR/AR devices will be a key to the development of the metaverse industry. It is embarrassing that the penetration rate of VR devices, which are considered as the entrance to the metaverse, is still low and considered as "immature products". This also means that we still have a long way to go from the opening of the meta-universe.
Although the metaverse concept is hot, Jotrin believes that in the coming year, wait-and-see is still the main sentiment of high-tech companies towards the metaverse concept. But one thing is certain, driven by the meta-universe concept, more manufacturers will start to enter the VR/AR market development in 2022.