While Nvidia is currently a leading player in the deep learning inference chip market, powered by its GPGPU (general-purpose computing on graphics processing unit) processors, other vendors are offering FPGA-based solutions to contest for market shares.For example, Alibaba Cloud has chosen Intel Altera Arria 10 FPGAs to power its F1 instance as well as Xilinx KU115 FPGAs for its F2 instance.Microsoft's FPGA-based Project Catapult servers are also designed to improve its Bing search engine services and Azure cloud computing services, further highlighting increasing influence of FPGAs in the deep learning inference processor market.While the inherent ability of an FPGA to be reconfigured and reprogrammed at any time allows FPGA-based products to be developed rapidly to shorten time to market, the use of Hardware Description Language (HDL) also enables FPGAs to accelerate deep learning inference processes and adapt to the rapid evolving of different algorithms.Meanwhile, Google is also developing its custom ASIC chips for deep learning inference processing. Such ASIC-based solutions will be able to optimize algorithm configuration for specific tasks and also allow streamlined hardware designs and minimized chip sizes to enhance the advantage of terminal-end devices, Digitimes Research believes.
Google has been regarded as the leading force for the development of artificial intelligence (AI) technology thanks to its AlphaGo AI computer program. While Google still holds the advantage in the cloud-AI sector, Apple appears to have outraced Google in the development of on-device AI products with the launch of its iPhone X and iPhone 8 devices.The built-in dual-core neural engine chip incorporated within the A11 processors, which power the iPhone X and iPhone 8 devices, integrates functionality such as machine learning, inference model and related algorithm to perform as an application-specific integrated circuit (ASIC) chip for image recognition and achieve hardware acceleration. The face ID application allows iPhone X users to actually experience on-device AI applications. Through the use of the TrueDepth camera, users can build up a depth map of face and the neural engine can identify and confirm the face and unlock the iPhone.With regard to AR development, although Google has rolled out its Tango AR platform and also cooperated with hardware vendors such as Lenovo and Asustek Computer to launch AR-enabled smartphones, these AR-enabled smartphones have failed to wage an AR campaign because their prices are too high due to the adoption of high-priced components such as depth measuring sensors, and the number of users of Tango-based AR phones is scarce due to limited cooperation smartphone partners, as well as immature integration between related AR software and hardware.In comparison, Apple's ARKit supports current and previous iOS-based devices launched up to two years ago through upgrades. Without using hardware components to build a 3D mode, the ARKit can utilize software algorithm to stack up AR scenes and objects to enable users to ride on AR experiences.To compete effectively with ARKit, Google launched ARCore at the end of August, using a technology similar to that of ARKit. However, since hardware devices supporting ARCore are still limited to only Android 7.0 and above smartphones such as Google's Pixel and Pixel XL, and Samsung Electronics' Galaxy S8, the number of ARCore-based devices will lag far behind those of ARKit-based products, Digitimes Research believes.
Taiwan-based Arbor Technology has been in the industrial PC (IPC) secor for over 20 years and has always been an important upstream partner of worldwide IPC brand vendors. With automated production lines and Industry 4.0 gradually becoming the new trends of the IPC industry, Arbor also is keenly developing solutions for clients.Digitimes recently interviewed Arbor's CTO Karl Chen, who talked about the company's latest progress on machine vision product development.Q: Is Arbor's machine vision equipment for automated production lines accounting for a heavy portion of the company's overall shipments?A: In the past few years Arbor was in the stage of market cultivation, focusing mainly on sending out samples and providing compatibility tests to first-tier machine vision equipment brand vendors worldwide.We started seeing surging demand for related applications in 2017 thanks to the rising popularity of the Industry 4.0 concept. Other key growth drivers include rising labor costs and clients' increasing demand for equipment's precision and yield. To cut down expenses and satisfy clients, more and more product manufacturers are considering adopting machine vision equipment to replace the traditional manual labor.We are seeing our brand equipment clients aggressively looking to satisfy such demand and therefore we expect our machine vision product shipments to continue enjoying growths in the next couple of years.Q: What is Arbor's vision about the machine vision market's future?A: Currently, factory manufacturing is the main application for machine vision. However, with the releases of Nvidia's deep learning platform, the Jetson TX2 and Intel's deep learning VPU solutions, edge computing will grow to become a new trend for the automated production market.Since all the calculation and logic analysis can be instantly processed at the edge side, in the future back-end servers will no longer need to process every single calculation request from equipment and can focus mainly on collecting calculated data done by the edge side to conduct big data analysis for Internet of Things (IoT) applications.With more hardware and software players joining the deep learning development, machine vision's applications will gradually expand from factory manufacturing to the smart home and more products such as autonomous driving vehicles, smart surveillance electric fencing, service robots and drone, will feature the machine vision functionality.Q: How can machine vision benefit IPC?A: Machine vision is actually part of the IPC and can be treated as a type of sensor. With cameras growing smarter, the development of controllers is now marching to a direction of featuring a smaller form factor and more sophisticated functions.For example, during the traditional image analysis process, IPCs were required to equip with a frame grabber card for gathering images, but now cameras are already capable of processing some simple analysis and image capturing, leaving controllers and IPCs with less work to do and they can have a more simplified and refined structure.However, this does not mean IPCs are less important than before; the systems are still responsible for in-depth analysis and calculation of the captured images and via system settings IPCs are able to handle actuators' controls based on users' demand. Smart cameras detecting defective products and sending information to the IPC for it to control robotic arms to remove bad items is an example of the process of the combination.With more software designers joining the industry to develop solutions for sensors and AI-related applications, the supply chain will be able to expand their product coverage to consumer applications such as autonomous driving vehicles, smartphone recycling detectors, service robots and smart surveillance electric fencing to further enlarge the market.Q: What are the advantages of Arbor's machine vision technology?A: To satisfy high-end machine vision equipment vendors' demand, Arbor is able to install transmission interfaces such as PoE/USB 3.0 in its image inspection equipment and DIO/COM ports in its related industrial equipment. Via its expandable modularized I/O board design, Arbor is able to quickly create customized solutions to answer each client's demand.In the modern world, machine vision technologies' required hardware configurations are no longer as simple as in the past. We usually see our clients' complete machine vision solutions have an overall power consumption surpassing 200W with multiple POE/USB 3.0 cameras lens and a specified image capture card.To dissipate the heat generated from the high power consumption in fanless equipment has always been an issue for IPC players, but Arbor's experience from its ODM/OEM support to brand vendors in Europe and North America in the past several years has given the company advantages in heat dissipation technology and high-performance IC designs.Q: What other emerging technology or products is Arbor currently developing?A: We have products for automated optical inspection (AOI) applications. Eurotunnel has adopted Arbor's industrial-level fanless embedded IPCs to improve the efficiency of its automated and manual tolling system at Channel Tunnel. The system also features machine vision technologies to detect car plates and control gates for better management.Arbor also has a mobile device recycle inspection system that can be used for recycling and inspecting handsets.The company is also developing autonomous driving solutions with deep learning capability. The high-end solutions will use Nvidia's Jetson TX2 platform, while entry-level applications are all based on Intel's VPUs.Arbor CTO Karl ChenPhoto: Company
HTC, which built the world's first Android-based smartphone, the HTC Dream, in 2008, is facing new challenges after it signed a US$1.1 billion deal to hand over its ODM team or half of its R&D unit to Google.It is generally considered as a good buy for Google to spend only US$1.1 billion to scoop up a hardware team as compared to the company's generous offer of US$12.4 billion to purchase Motorola's mobile business in 2012.On the other hand, HTC is expected to book a one-time non-operating gain of US$1.1 billion when the deal is completed in the first quarter of 2018, translating into an EPS of NT$40 for the quarter.The deal will help enhance HTC's financial structure given that the company had only a cash position of about NT$25 billion (US$827.3 million) at the end of the second quarter of 2017, down from NT$41.5 billion of a year earlier.HTC CFO Peter Shen also said that the sale of its ODM unit, or the "Powered by HTC" team, will help the company save its operating expenses by 30-40%.But based on HTC's financial results for the past six quarters, a 30-40% reduction in operating expenses is not sufficient enough for the company to make a turnaround.Cher Wang, chairperson and CEO of HTC, said that the company can utilize the new funds to continue to build its VR ecosystem, to grow its Vive business, while investing in other next-generation technologies, including Internet of Things (IoT), augmented reality (AR), artificial intelligence (AI) and 5G products.However, Wang did not mention that a substantial cut of its R&D team will also undermine the company's R&D capability needed for its strive to develop IoT, VR, AR, AI, 5G products.As a short-term impact, HTC is expected to see its revenues continue to wane, particularly after the first quarter of 2018, as the "Powered by HTC" team currently serves as a major source of income.While HTC has said it will continue its branded smartphone business, the handover of half of its handset engineers to Google certainly is not smart strategy that will improve the prospects of its beleaguered smartphone business. And helping Google become a smartphone maker that may compete with HTC's own brand in the future is hardly a smart strategy.In the long term, weakening R&D capability and diminishing market share is likely to force HTC to phase out from the smartphone market or transform itself into a white-box smartphone vendor.
From the control of household electric appliances at home to the monitoring of satellites in the outer space, grasping more data and expanding management depth and width are the exact benefits scientific and network technologies have brought to the mankind.In the development of Internet of Things (IoT) applications, various LPWAN (low power wide area network) communications technologies are capable of two-way communications. But due to the limitation of transmission frequency bandwidth, and in order to prevent remote control from occupying the bandwidth, most LPWAN applications are confined to one-way data collection. But Taiwan industrial PC maker and industrial IoT specialist Arbor Technology can utilize edge computing and LoRa two-way communication capability to meet remote smart monitoring needs.Chen Jung-chang, Arbor's chief technology officer, said the emerging of LPWAN wireless communication technologies enable long-range data transmissions in smart cities. Particularly, Chen continued, LoRa systems are very suitable for multiple outdoor IoT applications, as they have a transmission range of up to 10 kilometers and require terminal sensors to transmit data packets only at regular intervals, involving very low power consumption with batteries able to last for years before replacement.Although Taiwan is a densely-populated island with a high coverage rate for network signals, there are still many IoT application scenarios where various wireless signals and equipment maintenance personnel can hardly reach. LPWAN solutions can be applied to such scenarios and other harsher environments such as desert areas and offshore oil rigs, according to Chen. Arbor's LoRa solutions ideal for special application situationsARBOR has developed a set of LoRa transmission solutions dedicated to special application situations. Lin Yang-cheng, a senior manager at the firm's automation department, said the whole transmission architecture includes sensor nodes and gateways, with a mobile cloud app also developed to facilitate remote control by users. Lin added that terminal sensor nodes are built with ambient light sensors and can send temperature and humidity data packets to LoRa gateways at regular intervals, with a single gateway able to collect data transmitted by all the sensor nodes within the transmission range. Users can simultaneously manage numerous gateways, as they can monitor and conduct remote equipment control t after the data are sent to the cloud through the gateways. Most of LoRa solutions focus on data collection, as two-way communications could affect the transmission quality due to limited transmission volume. Against this, Arbor's LoRa solution can allow every terminal node to execute the on-and-off switch function by building relays in terminal devices.Citing a concrete application case, Lin said street lights are one of the indispensable devices that consume relatively higher energy, and therefore if street lights are built with multiple sensing technologies, not only various data of local areas can be collected through street lights, but energy consumption by street lights can also be reduced through remote smart sensing settings.Lin continued that users can send commands to terminal devices through gateways to switch on or off the devices. In addition, the "edge computing" concept can also be applied to control the switch, in that configuration logics can be built in terminal devices to automatically switch on or off relays if the set conditions are met. For instance, when sensors find ambient light in a dark state, street lights will be automatically turned on.
Traditional telecom is suitable for long-range coverage but involves high power consumption, and the coverage capability is also largely affected by the density of base stations deployed. Now in the IoT (Internet of Things) era, many application scenarios require transmission of small data volume, creating an opportunity for Low Power Wide Area Network (LPWAN) to emerge rapidly for extensive applications in smart manufacturing plants, smart cities and the agriculture, fishing and livestock sectors.As long-range transmission of small data volume with lower power consumption does not involve sophisticated technologies, many LPWAN solutuions have been rolled out, with the most notable ones including NB (narrowband)-IoT solutions using a licensed spectrum, and LoRa and SigFox solutions operating in an unlicensed spectrum.Kevin Yang, managing director of Taiwan wireless communication products maker GemTek Technology, said that with a transmission speed lower than than 200kpbs, LPWAN has a transmission range of 3-5 kilometers in urban areas and 10-15 kilometers in suburban areas, which can fill up the deficiency of short-range wireless local area networks WiFi and Zigbee. Further, its long-range transmission with low power consumption can also address the shortcoming of high cost seen in traditional cellular communication networks.Lin Yu-hsiung, a senior product manager at Taiwan industrial PC maker Advantech, pointed out that low power consumption is extremely crucial to IoT applications, as sensors deployed in many IoT application scenarios are widely scattered, with some even on high mountains or in uninhabited areas. Initial deployment of IoT devices in such sites would be hard to complete, Lin continued, and follow-up inspection and maintenance of the devices would also involve high personnel cost. Accordingly, he stressed, low-power consumption of data transmission modules plays a critical role in determining whether the installed IoT systems, usually with long life cycles, can effectively utilize the LPWAN's advantages to collect data at low cost.SigFox not on the same par with LoRa, NB-IoTGemTek's Yang opined that it is a wrong practice for people to put LoRa and NB-IoT on the same par with SigFox, reasoning that the latter is just a global LPWAN service platform operated by French IoT startup SigFox, instead of a common transmission protocol. By partnering with telecom firms around the world, SigFiox renders paid IoT connectivity services to terminal users through cellular wireless networks.Just like Apple setting specifications and regulations for partners in the entire iPhone supply chains to follow, SigFox has worked out a complete set of standards governing its entire LPWAN ecosystem, including software and hardware, for cooperation partners to observe. Accordingly, all the participants in SigFox LPWAN service platform should adopt the same software and hardware architectures to facilitate integrations and connections, according to Yang.In contrast, like Android as an open-source mobile operating system, LoRa is an open, long-range, low power wireless platform allowing any companies interested in the technology to make operating deployments, Yang said, adding that though LoRaWAN is a protocol specification built on top of LoRa tech developed by LoRa Alliance, the open platform can help to inspire innovations by all participating parties to make LoRa technologies become more advanced.LoRa remains the mainstream solutionAt the moment, LoRa remains the mainstream LPWAN solution and enjoys significant market demand, Yang continued. He analyzed that SigFox LPWAN is a closed system unilaterally dominated by the French company, limiting its technological development. Although also using unlicensed spectrum, LoRa can offer diverse technological integrations through cooperation partners to achieve better anti-jamming performance and higher transmission efficiency, according to Yang.On another front, as NB-IoT, based on standards set by International Telecommunication Union (ITU), adopts licensed spectrum that can better resist jamming, it is more suitable for the application scenarios that require highly reliable transmission and wider service coverage. But the cost for incorporating NB-IoT solution will be higher, as the solution has to be brokered by telecom firms, Yang noted.While some wonder whether the competition in LPWAN specifications will repeat that in 4G specifications, Yang said that different specifications have different advantages and are suitable for different application scenarios. For instance, he said, SigFox boasts the advantages of standardized service platforms and wider-area applications, while LoRa can perform better in dedicated application networks and allow users to build their private and closed application networks at lower cost than other LPWAN solutions.Some Taiwan telecom firms have started to operate LPWAN data transmission services using NB-IoT technology, but such technology is not applicable to many IoT application scenarios where telecom signals become quite weak. In this regard, LoRa Gateway can be applied to help users build private telecom networks to collect data at lower costs. In addition, LPWAN will not necessarily have to compete with LTE cellular networks and local area networks (LAN), and instead, they can complement each other to better adapt to multiple IoT application scenarios, Yang concluded.
The value of emerging smart IoT (Internet of Things) applications can be shown based on low power wide area networks (LPWAN) that boast high safety, low cost and easy management, and among LPWAN solutions, the LoRa technology has been kicking off.Global demand for LPWAN solutions has been on the rise, and they can be separated into two groups: solutions operating in an unlicensed spectrum, such as SigFox and LoRa; and those operating in a licensed spectrum, including LTE CAT-M and NB-IoT promoted by 3GPP.LPWAN solutions have attracted much attention mainly because its lower power consumption is a feature very crucial to the IoT applications that have long life cycles. In many application scenarios, sensors are widely scattered, such as those for monitoring offshore oil rigs or tracking wild cattle on pasturelands. Once initial deployments are completed, follow-up inspections and maintenances will involve high costs. If the power capacity of the data transmission modules can cover the whole equipment life cycle, the maintenance cost and efficiency of the equipment will lower significantly.Lin Yu-hsiung, senior product manager at Taiwan industrial PC maker Advantech, pointed out manufacturing plants have to often collect data about water, electricity and gas consumptions. In the past, the machine-to-machine (M2M) transmission of such data, though small in volume, still relied heavily on 2G and GPRS networks with high power consumption. The large batteries installed at the transmission devices entail higher follow-up maintenance cost, limiting the areas of their application.LPWANs, however, are quite suitable for long-range and small-data-volume management application scenarios, as they can handle long-distance transmission of small data volume with low power consumption. Through wireless transmission, the data can be sent to the cloud. At the moment, LPWANs are often used for street light management, parking space management and data collection by farming, fishing and livestock industries.LPWAN boasts wide applicationsCiting parking space management as an example, Lin said that if a parking lot can conduct remote collection of data concerning the usage situation of parking spaces, the parking lot can guide drivers to quickly park their cars at vacant spaces. In this regard, Advantech has developed smart parking sensors based on the LoRaWAN standard, which can transmit small data packet at regular intervals to update the state of parking space usage.As LoRa technology can lower battery maintenance cost and can collect data from hundreds of sensor nodes within the transmission range, Advantech has assisted many operators in boosting the utilization efficiency of parking spaces on roadsides.Lin pointed out there are many aquaculture farms in Taiwan, such as those for sturgeon farms that involve highly complicated farming technologies. Farmers have to regularly monitor the water's quality, oxygen level, pH value and temperatures in order to prevent undesirable environmental factors from affecting the growth of the fish.But most of Taiwan's aquaculture farms are spread across the southern coast and some are even located in areas on high mountains. The transmission nodes in these scenarios are widely scattered, and their managers usually fail to regularly inspect transmission equipment. In this aspect, LPWAN systems can be well applied to conduct outdoor IoT data collection.LoRaWAN helps users quickly build private transmission networksAlthough some local telecom firms have started to offer LPWAN data transmission services, their signals, however, are relatively weak when reaching coastal aquaculture farms. This can be addressed with Advantech's LoRa IoT Gateway, which can help users quickly build their own private communication networks without going through cloud servers, and can allow them to collect data from 500 sensor nodes at one time, according to Lin.In terms of benefits for users, the single LoRaWAN IoT Gateway boasts a transmission and reception range of 10 kilometers in open-air environments, allowing users to reduce their monthly cost of renting telecom base stations as they can directly transmit data through their private LoRA networks instead of the base stations. Moreover, data transmission safety can be further secured as the data will not have to be stored in the cloud through telecom networks.Advantech integrates software and hardware to accelerate LPWAN developmentWhile LPWAN is suitable for outdoor data collection, Advantech's LoRa IoT Gateway is mainly for indoor use, but it can also be well applied outdoors, according to Lin. Advantech has assisted customers in utilizing LoRaWAN for parking lot management by installing the gateway inside an on-site large-sized ticket processing system, which can be operated under industrial environments with wide temperature ranges.Currently, there are many LPWAN solutions available in the market. Compared with other suppliers, Advantech boasts strong advantages in technology coverage and system integration. While most suppliers concentrate on transmission modules, Advantech's software and hardware integration kits can help customers accelerate their LPWAN applications. For instance, the firm's LoRa IoT Gateway can be bundled with WISE-Paas platform to help customers conduct remote equipment management and provide a data flow logical editor that can immediately present the data in various forms of diagrams for reference in developing self-use applications.Lin concluded that LPWAN is ideal for application at cost-conscious application scenarios to handle long-range transmission of small-volume data with low power consumption, although it is unable to replace all wireless transmission applications, such as high-speed, real-time transmission of images and machine visions that require the support of other transmission technologies.Lin Yu-hsiung, senior product manager at AdvantechPhoto: Company
Worldwide top-5 notebook brand vendors' combined shipments grew on month, but decline on year in August. With the consumer notebook sector continuing to lack new innovations to stimulate demand, the overall notebook market may see difficulties growing further by only relying on the contribution from the enterprise area.Hewlett-Packard (HP), which performed strongly in the 12 months prior to August, only had a single-digit percentage on-year shipment growth in August. Lenovo returned to number two, but its shipments went down by over 10% on year.Dell achieved on-year growth in August shipments thanks to stable demand from the enterprise sector. Asustek Computer and Acer both saw an increase of over 30% on month in August shipments, but Asustek had weaker output than in August 2016.Within the top-5 ODMs, Compal Electronics had the highest on-year growth in August thanks to orders from Lenovo and Acer, while Quanta and Pegatron also had on-year shipment growths.Wistron saw shipments decline in August as the company has been cutting its notebook manufacturing business and looking to gain profits from non-PC product lines. Inventec also suffered an on-year shipment drop because its orders from HP have been gradually taken over by competitors.
Despite the fact that Japan Display (JDI) has been operating in the red for a long time, the Japan government is unlikely to let JDI go simply because Japan's automotive electronics industry still dominates the global market, and maintaining its own panel production capacity will continue to be a market niche for Japan's electronics industry, according to Digitimes Research.For this reason, JDI has recently launched an expansion project, aiming to ramp up the production capacity of its 4G plant in Tottori, which focuses on automotive panels, to 54,000 substrates a month from 45,000 units currently.JDI and Sharp have been dominating the automotive panel market in Japan and have also captured a relatively high level of market share in the US and Europe, respectively. However, the two firms have also seen their global market shares gradually diminishing as Taiwan-based AU Optronics (AUO) has actively stepped into the markets in Japan and Europe, and some Japan-based automotive electronics companies have reduced their orders to Sharp after Foxconn Electronics took over the management of Sharp.Meanwhile, LG Electronics is also stepping into the automotive electronics market and is offering support to LG Display to make its display subsidiary its main sources of panel supplies. LG Display has begun making inroads into the European market along with its improved technology.Foxconn has its own automotive electronics plants, using panels mainly from its subsidiary Innolux. Sharp has been increasing its purchases of panels from Innolux, further enhancing the supply capacity at Innolux.Although China-based panel makers including BOE Technology, Tianma Micro-electronics and Infovision Optoelectronics are also strengthening their automotive panel business, these firms are currently focusing their marketing on the domestic market.While Samsung Display is switching its focus to AMOLED panels, the AMOLED panels are unlikely to make significant inroads into the automotive panel market before 2020 due to the reliability of such panels.
With autonomous driving becoming a new trend in the IT market, many first-tier IT players including Google and Nvidia have been aggressively expanding their presence in related development. Taiwan's hardware providers have played a key role in the supply chain of the emerging industry.Digitimes recently had a chance to sit down with Jamie Fan, marketing specialist of Taiwan-based industrial PC (IPC) maker Neousys - a hardware supplier for autonomous driving platforms - to talk about the autonomous driving market's status and the company's product development for the market.Q: What is the latest status of the autonomous driving technology development?A: Autonomous driving has been a hot topic in the IT industry for the past few years and car vendors worldwide have also been aggressively investing in related developments.Currently, it is still a long way before we can see all cars feature autonomous driving functionality and drive all by themselves on the road. Most existing autonomous driving systems developed by car vendors are still just driver assistance systems.Before we can reach the final stage of the autonomous driving development, establishing regulations and policies is the major task for governments worldwide and a necessary transition phase for the market despite the fact that the available hardware is already able to support most of the latest technologies.Although many autonomous driving systems already feature self-driving functionality, these systems still require a driver to monitor and make decisions while on the road.Compared to consumer-related applications, we are more optimistic about the adoption of autonomous driving functions in the commercial area. The adoption of the technology from the segment is also expected to be much faster than consumer applications since there are fewer government regulations on commercial machines.We are already seeing broad implementation of autonomous driving technologies in the commercial sector such as industrial trucks and military-use vehicles. Military-use vehicles were one of the earliest adopters of autonomous driving technologies. Since military-use vehicles are often required to conduct missions in dangerous areas, the function is necessary to minimize the number of casualties.Q: What are Neousys' roles in the ecosystem? What are Neousys' advantages in the competition?A: Neousys was originally a supplier of fanless and GPU-accelerated industrial PC (IPC) products. We later crossed into the autonomous driving industry because of our experience in GPU application development.Neousys is a hardware supplier for autonomous driving platforms. All the platforms need to collect large amount of image data to simulate the road conditions. Our high-end GPU-accelerated IPC hardware is able to collect image data from cars' external sensor radars and conduct simple analysis and then send the information to a back-end processing database to share with other devices.We have recently partnered with a first-tier autonomous driving platform developer in China. The developer is looking to form an ecosystem for the market and has been aggressively establishing partnerships with software designers and other IT players.Prior to the cooperation with the China-based platform developer, Neousys also established partnerships with some enterprises in the US and Japan, but not specifically about the autonomous driving products.Because of Neousys' experience in designing fanless IPC products, our products feature better stability than those from competitors. Neousys also has several patented technologies over the hardware design that helps improve heat dissipation of its IPC products' graphics cards dramatically. One of the technologies can separate the heat generated by graphics card and by main machine so the heat does not add up together.Neousys also has a special chassis design that helps minimize the shock the GPU received while driving. Since high-end GPUs are heavily relying on its cooling fans for dissipating heat, the fans could be easily damaged if constantly experiencing vibration.To stay competitive, Neousys continues improving its existing technologies and has also been developing new fanless technologies for its GPU-accelerated IPC portfolio, looking to satisfy clients' demand.Q: How is the competition in the autonomous driving application market? What is Neousys' status in the market?A: Right now, the autonomous driving market is filled with Taiwan's IPC and IT players. Because of its popularity, the competition has been growing fiercer every year. Demand for autonomous driving applications is especially strong in China and many China-based IT players are aggressively investing resources in related application development, looking to grab a share.Because of China's large land mass, we are seeing strong demand coming from many of the country's traditional industries including agriculture. Applications such as manless cultivators have been enjoying strong sales because of their contribution in improving efficiency.The autonomous driving industry is currently still a rather closed market with car vendors all focusing on developing their own systems. To accelerate the development of the autonomous driving market, our China-based platform developing partner is trying to establish a standard so any players are able to use the platform to create autonomous driving products.As for the competition in the upcoming years, the autonomous driving market is expected to be see two groups of players: the closed platform group of mostly car vendors, and the open platform group consisting of such firms as Google. But Neousys still hopes that there will be a universal standard for the autonomous driving applications, which will truly help accelerate the development of the industry.Neousys Nuvo-5095GC IPCPhoto: Company
