Insights from Frost & Sullivan On September 8th, the Hangzhou Economic and Information Technology Bureau publicly solicited opinions on 'Several Opinions (Draft for Soliciting Comments) on Accelerating the Integration of Technological Innovation and Industrial Innovation to Promote the High-quality Development of Hangzhou's Manufacturing Industry'. The draft proposes that by 2030, the scale of the core AI industry should reach 600 billion yuan, with over 1,000 enterprises above designated size. Will the demand for domestic substitution of AI chips accelerate in 2025? Is the main direction of substitution towards the inference end or the training end? What is the potential market for domestic AI computing power chips in the next three to five years? How large will it probably reach?   Tran Cui, Executive Director of Frost & Sullivan Greater China, was interviewed by Times Weekly to discuss the domestic substitution of AI chips and future market potential. Times Weekly *Click on the end of the article Read the original article for a complete report     Q: Will the demand for domestic substitution of AI chips accelerate in 2025? Is the main direction of substitution towards the inference end or the training end? Tran Cui Executive Director of Frost & Sullivan Greater China   From the current industrial and policy environment, it is very likely that the demand for domestic AI chip substitution will enter an accelerated phase in 2025. There are three reasons:   First, the international environment remains uncertain, and the security of the computing power supply chain has become a consensus in the industry, with domestic manufacturers and application enterprises having an urgent need to strengthen their autonomy and control.   Second, the implementation of domestic AI applications is expanding, whether it is the deployment of large models on the enterprise side or the application of generative AI on the consumer side, both driving larger-scale computing power demand.   Third, the technical capabilities and ecological adaptability of domestic chip manufacturers have significantly improved in the past two years, and they are already ready to undertake larger-scale applications.   As for the direction of substitution, the inference end will be substituted faster in the short term. Inference scenarios have higher requirements for chip cost-effectiveness, energy efficiency ratio, and ecological adaptability, and domestic manufacturers can already provide competitive solutions in these dimensions. At the same time, the domestic application market is vast, providing opportunities for rapid iteration. The substitution of the training end is also advancing, but due to the involvement of software frameworks, ecological compatibility, and the stability verification of large-scale computing power clusters, the progress will be more steady and may require a longer time window.   Q: What is the potential market for domestic AI computing power chips in the next three to five years? How large will it probably reach? Tran Cui Executive Director of Frost & Sullivan Greater China   In the next three to five years, the market space for AI computing power chips in China will be very large, mainly due to the following reasons:   First, the breadth and depth of the demand side are continuously expanding. Large models are gradually becoming the infrastructure for industrial intelligence, not only with internet companies increasing their investment but also traditional industries such as finance, energy, manufacturing, and healthcare actively exploring AI applications. This means that the long-term demand for inference and training computing power will remain high.   Second, dual-driven by policy and capital. At the national level, it has been clearly stated that an autonomous and controllable computing power system should be built, and local governments are also promoting the construction of intelligent computing power centers. At the same time, the capital market's investment enthusiasm for AI computing power has not diminished, creating an environment for the accelerated maturity of the industrial chain.   Third, ecological factors are gradually improving. With the continuous progress of domestic chips in hardware performance, software stack compatibility, and developer ecosystems, the adoption threshold for application enterprises is gradually decreasing, and the conditions for chip large-scale implementation are maturing.   Therefore, in the next three to five years, the domestic AI computing power chip market will not only grow rapidly in scale but more importantly, AI computing power chips will become one of the most strategically significant links in the entire AI industrial chain. *This interview has been published in Times Weekly, with reporter Zhu Chengcheng, and the original title was: Alibaba, DeepSeek Launch Successive Moves, Will Hangzhou's AI Industry Reach 600 Billion in Size within Five Years? Huawei's chip roadmap announced, Ascend Kunpeng advancing side by side! Xu Zhijun: The world's strongest super-node will be launched in the fourth quarter of next year

At a time when the AI industry is experiencing frequent technological breakthroughs, "AI companionship" is becoming a new track that is being bet on by capital, entrepreneurs, and industry giants alike. How can we judge the overall popularity of the current AI companionship track? Is it a phase-specific trend or has it a long-term growth logic? The current AI companion market offers a variety of product forms with varying quality standards. When is it expected that the industry as a whole will enter a mature phase? What are the key indicators of this process? In the user experience of AI companion products, long-term memory and continuous interaction capabilities are considered one of the core competencies. When is the approximate window period for technological breakthroughs in this field within the industry?   Li Qing, Director of Frost & Sullivan China, was interviewed by Time Weekly to discuss the opportunities in the AI companion track. . Time Weekly *Click at the end of the article Read the original text View the full report   Q: How can we gauge the overall popularity of the current AI companion track? Is it a short-term trend or one with long-term growth potential? Li Qing Director, Greater China, Frost & Sullivan   Nowadays, more and more families, young people, and even the elderly are beginning to accept an AI role as a companion in their lives. This is not just hype, but a real social change at work—such as smaller families, fewer children, and more elderly individuals, many of whom are facing the problem of lacking emotional companionship. AI companionship is no longer just a short-term trend, but is becoming a long-term social need and technological trend. It is moving from the technology validation phase to the scenario explosion phase. However, long-term development requires crossing three thresholds: technical capability ceilings, market demand matching, and policy ethics compliance.   Q: The current AI companion market offers a variety of product forms with varying quality standards. When is it expected that the industry as a whole will enter a mature phase? What are the key indicators of this process? Li Qing Director, Greater China, Frost & Sullivan   AI is accompanying the industry into maturity, which is expected to occur around 2028-2030. During this period, there will be several stages: currently, it is a time for technical refinement and market trial and error. Next, it will enter an integration phase, and finally, a stable industry pattern will be formed.   The key indicators of maturity include:   1. Technological standardization: Multimodal interaction has become a standard feature; emotion computing algorithms achieve an accuracy rate of over 90% in cross-age and cross-language sentiment recognition; the deployment and popularization of large-end models on the client side have enabled low-latency, offline intelligent interaction.   2. Clear product segmentation: There are products for different positioning, such as 'Home AI Companion' products with long-term memory, personalized growth, health monitoring, etc.; IP co-branded smart toys that integrate education and companionship; and low-cost voice interaction toys with focused functions.   3. Mature business model: Hardware + subscription services become mainstream: Value-added services such as content updates, personalized training, and mental health counseling generate stable revenue.   4. Regulatory oversight and standard implementation; Establish an ethical framework for AI emotional interaction to prevent risks such as addiction and misinformation; Establish an industry certification system.   Q: In the user experience of AI companion products, long-term memory and continuous interaction capabilities are considered one of the core competencies. When is the window period for technological breakthroughs in this area of the industry approximately? Li Qing Director, Greater China, Frost & Sullivan   The years 2026-2028 are a critical window period for achieving large-scale breakthroughs in long-term memory and continuous interaction capabilities. In the short term, such as next year, optimization based on vector databases will enable AI to remember more details, such as users' birthdays and preferences, significantly improving cross-session memory accuracy. In the longer term, as large models become more lightweight, their capabilities are enhanced, neural-symbol systems are integrated, and edge computing capabilities improve, AI toys can store and learn user data locally, achieving cross-day memory, long-term memory, and even personality evolution. AI may naturally forget unimportant information like humans, while forming associative links with key memories. For example, mentioning rain might remind the user of a past preference for reading on rainy days. It is not just responding to you, but can predict you, care about you, and grow with you. This is the true warm companion. In even longer periods, such as five years from now, brain-computer interfaces may appear as auxiliary functions, influencing mainstream product design and user experience decisions. *This interview has been published in Time Weekly, with the reporter being Zhu Chengcheng. The original title was: The Business That Lei Jun, Musk, and Zhu Xiaohu Are Eyeing May Have a Bubble!

Frost & Sullivan insights Data from the National Post Bureau shows that by the end of 2024, the large-scale application of unmanned delivery logistics vehicles had accumulated over 6,000 units, delivering hundreds of millions of orders to users in over 100 different scenarios. How has the industry ecosystem for unmanned logistics vehicles progressed, and has it achieved a 'technical closed loop + commercial closed loop'? The industry still faces some challenges, such as unstable 'road rights', the inability of technology to meet complex scenarios like night-time, rainy or snowy weather, and mixed traffic between people and vehicles. Additionally, some manufacturers adopt low-price strategies to capture market share, even leading to 'strategic losses'. In light of this, in the next phase, where will the focus of competition in the industry be?   Yang Lei, consulting director for Frost & Sullivan Greater China, was interviewed by Securities Daily to discuss how unmanned logistics vehicles are evolving from 'able to run' to 'scalable and sustainable profitability'.   Q: What is the current progress of the industry ecosystem for driverless logistics vehicles, and has it achieved a 'technology loop + business loop'? Yang Lei Frost & Sullivan Greater China Consulting Director Currently, unmanned logistics vehicles are transitioning from single-point trials to a collaborative phase that emphasizes both 'technology' and 'platform'. The industrial chain has formed a multi-entity pattern consisting of OEMs and chassis platforms, autonomous driving full-stack systems, fleet operations and intelligent scheduling platforms, scenario operators and shippers, as well as urban V2X and energy supply systems. The supporting sensors, computing power, charging and swapping facilities, insurance, and financial solutions are continuously being improved.   Multi-vehicle convoys, remote duty, human-machine collaboration, and vehicle-cloud integration have become basic elements of operational systems. Business models are also shifting from one-time equipment sales to a service-oriented path that charges by mileage, shifts, and service level agreements (SLAs).   In terms of the maturity of 'technology closed-loop + business closed-loop', in enclosed and semi-enclosed scenarios such as ports, mining areas, park main roads, airport aprons, and sanitation operations, the technology closed-loop is available, reusable, and capable of cross-park migration. In business terms, a continuous contract focusing on transportation capacity/service has been formed, with the feasibility of generating positive cash flow; these scenarios share common characteristics such as predictable tasks, simple interaction objects, high mileage density, which make it easier to increase the effective hours per vehicle and fleet utilization rate, while the unit mileage cost curve flattens with scale.   In comparison, the end-to-end delivery of open urban roads and urban arterial roads are still in the stage of large-scale exploration due to road right approval, complex long-tail conditions, mixed human and vehicle traffic, and fluctuations in order density. The commercial closed loop mainly manifests as phased and regional 'local closed loops'.   Overall, the industry has transitioned from 'able to run' to 'able to replicate and run'. The evaluation loop should not only look at demonstration mileage but also observe replicable deployment cycles, remote takeover rates, effective hours per vehicle, the relative discount of comprehensive cost per kilometer compared to labor costs, as well as the health of SLA fulfillment and contract structure. Only when these indicators are repeatedly verified across different cities and customers does it mean that the technical and commercial loops have entered a sustainable scale-out phase.   Q: The industry still faces some challenges, such as unstable 'road rights', technology that cannot meet complex scenarios like nighttime, rainy or snowy weather, and mixed traffic of people and vehicles. In addition, some manufacturers adopt low-price strategies to capture market share, even leading to 'strategic losses'. Given this, where will the focus of competition in the industry shift in the next phase? Yang Lei Frost & Sullivan Greater China Consulting Director Against the backdrop of unstable road rights, challenges in long tail conditions such as at night and in rain or snow, and some companies trading volume for low prices or even 'strategic losses', competition in the industry will shift from 'technological feasibility' to 'scaling up and sustainable profitability'.   Firstly, security and reliability remain fundamental hard constraints. Enterprises need to steadily reduce takeover rates and unsafe event rates in complex scenarios such as at night, during rain or snow, against backlight, and in dense pedestrian and vehicle traffic. They should solidify capabilities into auditable metrics through systematic certifications such as functional safety, expected functional safety, and cybersecurity, trading quantifiable security performance for more stable road rights and customer trust.   Secondly, the cost reduction of the entire vehicle and full-stack operations will determine the economy per unit mileage. By solidifying and integrating sensors, centralizing domain control with vehicle-cloud collaborative computing power, standardizing platform-based chassis and upper installations, BOM can be reduced. At the same time, on the operational side, strengthen management of energy and consumables, predictive maintenance, insurance, and financial leasing optimization, as well as service network layout, to build a predictable and controllable full life cycle TCO.   Secondly, the portability of ODD and 'template-based city opening' will become key to magnifying advantages. Relying on high-fidelity simulation and data flywheel, different cities and parks can be abstracted into reusable scenario packages and policy libraries. This compresses the cycle from signing to stable operation into weeks, forming barriers with replication speed and marginal cost advantages.   Meanwhile, fleet-level operational capability will become the decisive factor. Intelligent scheduling and formation management need to reduce idling and waiting times, deeply integrating with the cargo owner's TMS, WMS, and OMS to achieve end-to-end online order, capacity, and settlement processes. Under a compliant framework, through one-to-many remote monitoring, human input can be precisely directed towards a very small number of long-tail instantaneous tasks.   Governance coordination will also directly affect the scale cap. The path from demonstration zones to gradual gray release requires enterprises to use KPIs such as takeover rate, availability, and mileage density as a basis for policy formulation, enabling the road right-of-way, insurance, and liability determination mechanisms to more smoothly connect with business operations.   In terms of business models, the industry needs to shift from simply offering low prices to value pricing linked to service levels. Contract structures such as cost-saving sharing, minimum guarantees tied to performance, and other improvements can be adopted to enhance cash flow quality and gross profit structure, while avoiding price wars from eroding R&D and safety redundancy investments.   Finally, the assetization of data and models, along with ecological collaboration, will determine long-term compound interest rates and margin widths. Enterprises that can solidify data governance and training/simulation platforms across vehicle models and scenarios into stable engineering capabilities, and build a solid partnership network with OEMs, energy supply, mapping and V2X, insurance finance, and scenario operators, have more opportunities to achieve a positive cycle of 'running safer and cheaper'.   Overall, the leader in the next phase will not necessarily be the one with the lowest hardware prices, but rather a company that can continuously deliver fleet economics under measurable security metrics, possesses the capability for rapid cross-city replication, and has a healthy contract structure of 'technology + platform + operations'. *This interview has been published in Securities Daily Reporter: Wang Jingru, original title: Driverless Cars Reconstructing Express Logistics “Capillaries&rquo;

Frost & Sullivan insights In the face of physical obstacles, the ability to re-establish a bridge between "brainpower" and the torso or facial features, and to achieve the control of expression or mechanical limbs through "thought," is accelerating its integration into reality. Recently, China's first prospective clinical trial of an invasive brain-computer interface was successfully conducted, marking China as the second country in the world after the United States to enter the clinical trial phase in invasive brain-computer interface technology. What is the significance of invasive brain-computer interfaces entering the clinical phase? What were the previous difficulties that hindered their entry into the clinical phase? From the current company layout, it seems that the US market has chosen the invasive route more frequently. What are the reasons? In terms of both the number of brain-computer interface companies and the technology of some companies in the US, what are the relative advantages over Chinese companies? Currently, brain-computer interfaces are divided into three routes: invasive, semi-invasive, and non-invasive. What are the significant differences in the implementation scenarios of these three technical routes? Does this mean that the commercial progress of the three routes will also vary significantly?   Zhou Mingzi, Executive Director of Frost & Sullivan Greater China, was interviewed by the 21st Century Business Herald to discuss the technological leaps behind the clinical breakthroughs of invasive brain-computer interfaces.   21st Century Economic Report   Q: Recently, our country successfully conducted the first prospective clinical trial of an invasive brain-computer interface, becoming the second country in the world to advance to this stage. What is the significance of the entry of invasive brain-computer interfaces into clinical trials? What were the previous difficulties hindering their clinical advancement? Zhou Mingzi Executive Director, Greater China, Frost & Sullivan The entry of invasive brain-computer interface technology into clinical trials in China is of great significance, marking a breakthrough in this field. From three dimensions: scientific research attitude, technical assurance, and future development: Firstly, China's scientific research attitude towards invasive brain-computer interfaces has shifted from cautious to proactive. In the early stages, due to technical risks, research efforts mainly focused on non-invasive technologies. However, in recent years, there has been a significant increase in investment in invasive technologies, with industry and other sectors listing them as a key technical direction since 2023. Secondly, the successful conduct of this clinical trial proves that China has the capability to ensure clinical application of invasive technologies, reaching international leading levels in key indicators such as biocompatibility and signal stability, laying a solid technical foundation for subsequent medical applications. Frost & Sullivan believes that this breakthrough will greatly boost scientific research enthusiasm and promote the formation of a development pattern that equally emphasizes invasive and non-invasive technologies. Invasive technologies will develop deeply in professional fields such as medical rehabilitation, while non-invasive technologies will target broader consumer applications. The two complement each other's advantages to jointly promote the industrialization process of brain-computer interface technology. In the future, China is expected to achieve more original breakthroughs in this frontier field. Historically, brain-computer interface (BCI) technology has encountered numerous challenges in clinical applications: Technically, long-term implantation issues such as biocompatibility (e.g., inflammation caused by electrode materials) and signal stability (signal weakening after implantation) are major obstacles that urgently require the development of softer electrode materials and anti-interference algorithms; ethically, there are disputes involving the protection of thought privacy and the definition of the boundaries of autonomous consciousness; clinically, it is necessary to establish standardized surgical procedures (e.g., precise implant positioning) and establish long-term postoperative monitoring systems; from the patient's perspective, it is essential to weigh the risks and benefits of surgery. Currently, this technology mainly serves specific groups such as those with severe disabilities and faces expensive costs. These factors collectively hinder the progression of BCI technology from laboratory to clinical application.   Q: Judging from the current company's layout, it seems that the US market prefers an invasive approach. What are the reasons for this?   Zhou Mingzi Executive Director, Greater China, Frost & Sullivan American medical innovation companies are advancing invasive brain-computer interface technology, mainly relying on three significant advantages. Firstly, at the technical level, the United States has a profound accumulation and innovation in key core technologies such as microelectrode materials and neural decoding algorithms. For example, the flexible electrode technology developed by Neuralink is a typical representative of this technological advantage. This technology can achieve efficient connection with brain neurons, greatly improving the accuracy and stability of signal transmission. In addition, the US Food and Drug Administration has a relatively open approval process for high-risk medical innovations, which provides convenient conditions for the clinical application of new technologies. The flexibility of the approval process allows medical innovation companies to transform research findings into actual medical solutions more quickly, thereby accelerating the commercialization process of technology. Secondly, in terms of the selection of application scenarios, American medical innovation companies tend to focus on severe medical fields such as Parkinson's disease and spinal cord injury, which require extremely high precision in treatment. Invasive brain-computer interface technology, due to its high-precision characteristics, can better meet these clinical needs and has therefore been widely applied in these fields. For example, through brain-computer interface technology, patients can use their mental activities to control external devices such as wheelchairs or prosthetics, providing them with unprecedented autonomy and improved quality of life. Finally, in terms of patient acceptance, American citizens have a high level of acceptance of innovative medical technologies, and they are more willing to try and adopt new technologies. This open attitude provides a favorable social environment for the promotion and application of new technologies. At the same time, America's comprehensive insurance system can cover to a certain extent the costs of these high-risk, high-reward invasive treatment options, which undoubtedly reduces the economic burden on patients to adopt these new technologies and makes it easier for these options to be promoted and applied. In summary, the technology route chosen by American medical innovation companies not only reflects their technical strength in cutting-edge medical fields but also demonstrates their careful consideration of differentiated competitive strategies. By combining these advantages, America maintains a leading position in the field of invasive brain-computer interfaces and is expected to bring good news to more patients in the future.     Q: In terms of the number of brain-computer interface companies and the technology of some of these companies, what are the relative advantages that the United States currently has over Chinese enterprises? Zhou Mingzi Executive Director, Greater China, Frost & Sullivan Compared to other regions around the world, the United States exhibits a variety of advantages in the field of brain-computer interface (BCI) medical innovation. Firstly, in terms of talent supply, the US has established a comprehensive interdisciplinary talent training system for neuroscience and engineering through top institutions such as Stanford University and MIT, as well as the ecosystem of Silicon Valley, capable of continuously providing talents with composite R&D capabilities. In terms of commercialization, the mature venture capital mechanism and flexible FDA approval process (including the breakthrough device designation program) in the US have promoted rapid technology transformation. Industrially, this is due to the advantages of healthcare industry giants in supply chain support and clinical resource integration. Additionally, American patients have a higher acceptance of innovative therapies, and medical insurance coverage is more extensive. Finally, driven by the demand from government and military agencies represented by NASA, these factors together create a unique advantage in the innovative ecosystem.   Q: Currently, brain-computer interfaces are divided into three routes: invasive, semi-invasive, and non-invasive. What are the significant differences in the implementation scenarios of these three technical routes? Does this mean that there will also be significant differences in the commercial progress of the three routes? Zhou Mingzi Executive Director, Greater China, Frost & Sullivan Currently, invasive brain-computer interface technology has been applied in many clinical fields, especially in the following three main areas: First, the technology is widely used for treating severe motor dysfunction. For example, for amputation patients with limb disabilities, through brain-computer interface technology, they can control external robotic arms with their thoughts alone, performing a series of complex actions such as grasping objects and eating. In addition, for patients with spinal cord injuries, invasive brain-computer interface technology enables them to control exoskeletons, thereby partially restoring their mobility. Second, invasive brain-computer interface technology shows great potential in the treatment of neurological diseases. It is used for deep brain stimulation therapy for Parkinson's disease patients and neuroregulation for patients with refractory epilepsy to alleviate symptoms or control seizures. Finally, for patients with impaired language function, invasive brain-computer interface technology provides the possibility of reconstructing language function, helping them regain their communication ability with the outside world. Semi-invasive neural monitoring techniques, such as ECoG (Electroencephalogram), are widely used in neurosurgical operations that require high signal quality, especially playing a key role in locating epileptic foci. This technique monitors brain electrical activity by placing electrodes on the surface of the brain, providing surgeons with precise guidance to help them more accurately identify and handle abnormal brain tissue, with the aim of controlling or eradicating epileptic seizures. Non-invasive brain-computer interface technology has found applications in multiple clinical fields, mainly focusing on the following three important scenarios: Firstly, it shows great potential in neurorehabilitation therapy. For example, stroke patients can control external robotic hands by wearing electroencephalogram (EEG) caps to perform rehabilitation training actions such as grasping. In addition, non-invasive brain-computer interface technology also demonstrates unique value in the intervention treatment of mental illnesses. By regulating the patient's brain state, non-invasive brain-computer interface technology has been used to treat mental illnesses such as depression and attention deficit hyperactivity disorder (ADHD), providing patients with new treatment options. Due to different clinical application scenarios, the commercial progress of invasive, semi-invasive, and non-invasive brain-computer interfaces will inevitably vary, but there is no superiority or inferiority among them. Although invasive brain-computer interfaces can obtain the most accurate neural signals, providing strong support for the treatment of some severe neurological diseases and in-depth research on neural functions, factors such as high surgical risks and expensive costs limit their large-scale commercialization. Currently, they are mainly concentrated in some high-end scientific research and specific clinical treatment projects. Semi-invasive brain-computer interfaces have achieved a certain balance between signal quality and safety and have a stable market demand in specific fields such as neurosurgery. With continuous technological maturity and gradual cost reduction, their commercial scope is expected to gradually expand. Non-invasive brain-computer interfaces, with their non-invasive and convenient features, have been widely used in multiple clinical fields such as neurorehabilitation and mental illness intervention, and also have great development potential in the consumer market, such as in scenarios like gaming and smart home control, with a very broad commercial prospect. Different types of brain-computer interfaces play unique roles in their respective suitable fields, jointly promoting the development and application of brain-computer interface technology.   Q: One view is that brain-computer interfaces will be the ultimate method of human-computer interaction. What catalysis or assistance will the trend of large AI models bring to the development of the brain-computer interface industry? Zhou Mingzi Executive Director, Greater China, Frost & Sullivan Frost & Sullivan has pointed out that digital innovation technologies such as artificial intelligence will drive the rapid development of brain-computer interfaces (BCI) at three levels: Breakthroughs in clinical treatment: Large models can construct personalized disease prediction models by analyzing large amounts of neural signal data (such as EEGs from epilepsy patients), thereby improving the accuracy of treatments such as deep brain stimulation (DBS). Metaverse technology can provide virtual rehabilitation training environments to accelerate the recovery of neural functions. Leap in Interaction Capabilities: Large models' powerful capabilities in natural language understanding enable them to decode more complex brain signals and intentions, elevating BCI from simple command control to semantic-level interaction (for example, allowing paralyzed patients to generate coherent text directly). In the metaverse, digital twins can provide real-time feedback on BCI operation effects, thereby optimizing the training process. Enhanced innovation in the human body: Combining cognitive enhancement algorithms with large models, BCI may achieve memory reinforcement (such as rapid learning of foreign language vocabulary), while the immersive environment of the metaverse will expand the application scope of BCI in spatial perception and multimodal fusion. Currently, the integration of technologies is beginning to show signs of progress. For example, Neuralink's implantable BCI is exploring voice synthesis solutions that work in conjunction with large models, while a leading international AI company has integrated non-invasive BCI technology into its metaverse innovation research, enabling the control of virtual objects with thoughts. Frost & Sullivan believes that this cross-disciplinary innovation will reshape the new paradigm of human-machine coexistence over the next five to ten years. *This interview has been published in 21st Century Economic Report Reporter: Luo Yiqi, original title: Brain-computer interface implementation race: From technology to clinical trials

Insights from Frost & Sullivan During the World Artificial Intelligence Conference (WAIC) in Shanghai, the city announced plans to create a high-level autonomous driving pilot zone. What significance does this hold for industry development? Is there an obvious shift in policy attitude compared to previous deployments of Robotaxis in specific areas by major cities such as Guangzhou and Shenzhen (for example, from cautious exploration to active promotion)? This year, the implementation and advancement of intelligent driving solutions providers like Pony.ai and WeRide are accelerating overseas markets. Coupled with Tesla's attempts, why does it seem that the Robotaxis market is advancing more actively this year? Previously, the main challenges for Robotaxis deployment were mass production costs and operational maintenance costs. Have domestic manufacturers achieved significant cost reductions in these two areas this year? What are the reasons? Besides the relatively positive attitude in the United States, which markets abroad have a higher acceptance of new business models like Robotaxis and show more visible commercial progress?   Yan Yiyang, an analyst at Frost & Sullivan's Greater China region, was interviewed by 21st Century Business Herald to discuss the underlying logic of Robotaxis entering the fast lane.   21st Century Business Herald Q: During the WAIC in Shanghai, the city announced plans to create a high-level autonomous driving pilot zone. What significance does this hold for industry development? Is there an obvious shift in policy attitude compared to previous deployments of Robotaxis in specific areas by major cities such as Guangzhou and Shenzhen (for example, from cautious exploration to active promotion)? Yan Yiyang Analyst at Frost & Sullivan's Greater China region This year, Shanghai released the 'Shanghai High-Level Autonomous Driving Pilot Zone "Model Speed Smart Mobility" Action Plan" (hereinafter referred to as the 'Action Plan') at the World Artificial Intelligence Conference (WAIC) in 2025. Its main contents can be summarized into the following key points: 1) Large-scale autonomous driving, with clear targets for passenger capacity and road testing of Level 4 autonomous driving, far exceeding previous pilot operations in small areas and limited capacity; 2) Networked application scenarios, aiming to build a regional comprehensive operation network, breaking the previous limitations of point and block testing areas; 3) Industrial ecosystem, promoting the construction of an autonomous driving industry ecosystem that includes complete vehicles, core components, algorithms, data, and operational services. From the perspective of industrial development, the release of the 'Action Plan' marks a determination to accelerate the further scale-up and commercialization of China's autonomous driving and Robotaxis industries. The 'Action Plan' conveys confidence in the continuous improvement of autonomous driving technology maturity and reveals an urgent need for the commercial implementation of Robotaxis. At the same time, achieving large-scale commercial operation of Robotaxis in Shanghai and forming a complete 'Shanghai solution' that includes technical standards, operational norms, safety supervision, and regulatory applications will provide a model for further promoting the implementation of Robotaxis nationwide and greatly accelerate the development progress of the Robotaxis market. Compared with previous small-scale commercial pilot deployments of Robotaxis in specific areas by cities such as Guangzhou and Shenzhen, the 'Action Plan' clarifies the policy attitude towards truly integrating autonomous driving into urban transportation ecosystems and the urgent need to break through profit bottlenecks and operational models brought about by regional pilots. This also indicates that China's autonomous driving policy will develop in a more systematic and open direction.     Q: This year, the implementation and advancement of intelligent driving solutions providers like Pony.ai and WeRide are accelerating overseas markets. What are the reasons? Besides the relatively positive attitude in the United States, which markets abroad have a higher acceptance of new business models like Robotaxis and show more visible commercial progress? Yan Yiyang Analyst at Frost & Sullivan's Greater China region Firstly, technical capabilities are the foundation for Chinese intelligent driving solution providers to successfully expand overseas markets. From a technical capability perspective, Chinese intelligent driving solution providers represented by Pony.ai have formed a relatively mature intelligent driving technology solution and service system through years of research and development, testing, and verification, and possess the ability to quickly deploy operations in new regions. Secondly, from an enterprise strategic perspective, going overseas is an operational choice for Chinese intelligent driving solution providers to develop revenue growth points and seek commercial scale expansion. Furthermore, operating Robotaxis in different countries, with different regulations and traffic environments, gives Chinese intelligent driving solution providers the opportunity to participate in or even lead the formulation of some regional and even global autonomous driving technology standards, safety standards, and operational norms. At the same time, cooperation with global or regional transportation giants is a practical exploration of the 'technology empowerment + localized operation' business model by Chinese intelligent driving solution providers, which helps them understand user habits, payment systems, and regulatory requirements in different markets and accumulate experience for building replicable and scalable global business models. Currently, looking at the global market, except for the United States, regions such as the Middle East, Southeast Asia, and Europe show a high market acceptance and commercial progress visibility for new business models like Robotaxis. The Middle East region, represented by the UAE, has made the development of smart cities and future transportation an important national strategy. In terms of commercial progress, Pony.ai has also reached an agreement with the Dubai Roads and Transport Authority to start road testing this year. In the Southeast Asian market, Singapore, as the world's first country to carry out public Robotaxis trial operations, has always maintained a globally leading development process in terms of relevant policies and regulatory systems; Pony.ai and Baidu's RoboRover are both expected to land in Singapore this year for commercial operations. In Europe, countries such as Germany, France, and the United Kingdom are also accelerating the development of the Robotaxis market. German automakers such as Volkswagen and Daimler have participated in the demonstration operation of Robotaxis projects in their respective countries; WeRide will carry out pure driverless commercial operations in France through strategic cooperation; the United Kingdom will launch a Robotaxis pilot operation in London in 2026.   Q: Coupled with Tesla's attempts, why does it seem that the Robotaxis market is advancing more actively this year? Previously, the main challenges for Robotaxis deployment were mass production costs and operational maintenance costs. Have domestic manufacturers achieved significant cost reductions in these two areas this year? Yan Yiyang Analyst at Frost & Sullivan's Greater China region The main reasons for the accelerating advancement of the global and Chinese Robotaxis markets this year can be seen from three dimensions: technical maturity, commercial verification, and policy promotion: At the technical level, the maturity of Level 4 autonomous driving technology has further improved. The 'multi-sensor fusion + modular system' solution represented by Waymo has verified through large-scale testing that the safety and reliability of Robotaxis have gradually reached the commercial threshold; Tesla's 'end-to-end model-driven pure vision solution' has significant economies of scale, iterative efficiency, and expansion capabilities, with extremely high commercial potential. In terms of commercial verification, Chinese Robotaxis companies represented by Baidu Apollo, Pony.ai, and Baidu RoboRover have demonstrated the feasibility of Robotaxis commercialization after multiple verifications from technical testing to fully driverless commercial demonstrations. From a policy perspective, the US Department of Transportation issued relevant policies in April 2025 to simplify unmanned driving technology testing and reporting; first-tier cities such as Beijing, Shanghai, Guangzhou, and Shenzhen have further opened up Robotaxis commercial operations this year; countries and regions including the Middle East, Europe, and Japan are also actively following up on policy support for Robotaxis. Under the resonance of multiple factors, the Robotaxis market in 2025 showed a 'spring after rain' trend. The downward trend in the mass production and operational maintenance costs of Robotaxis is highly predictable. Take lidar as an example. In the past, lidar was a core bottleneck restricting the cost of Robotaxis, mainly due to limited production capacity in the early stages of technology development. With the technological upgrading and large-scale mass production of related industries, lidar prices have dropped significantly in recent years. According to public information, the cost of lidar in Pony.ai's seventh-generation autonomous driving system will be reduced by 68%. Other core components, such as automotive-grade chips, are also expected to see significant cost optimization with further optimization of single-chip computing power and the gradual mass production application of domestic automotive-grade chips. It is estimated that in the next two years, the market will see significant reductions in the mass production and operational maintenance costs of Robotaxis.   Q: Currently, for major intelligent driving solution providers, Robotruck is still the main source of contribution. Can the mileage and experience accumulated in the truck field promote the faster maturity and implementation of the Robotaxi model? What are the core technical challenges between Robotaxis and Robotruck? Yan Yiyang Analyst at Frost & Sullivan's Greater China region From a technical perspective, there is strong commonality in the underlying technical architecture between Robotaxis and Robotruck, enabling a certain degree of technical reuse in basic perception and planning algorithms such as traffic rules, lane line recognition, and basic obstacle classification. The data accumulation during operation of Robotruck can also provide technical transfer value for the training of autonomous driving models (such as long-distance target recognition and lane keeping) by Robotaxis. Overall, the experience accumulation of Robotruck in the trunk logistics field is undoubtedly beneficial for promoting the development of Robotaxis. Through technical reuse and data feedback, Robotruck can provide a solid foundation for the development of Robotaxis. Compared with Robotruck, Robotaxis faces more complex operating scenarios. The main operating scenario of Robotruck is highways, where road rights are relatively clear, traffic participant types are relatively single, and the scenario complexity is relatively low compared to urban scenarios. In contrast, the urban scenarios where Robotaxis mainly operate have significantly increased complexity, facing challenges such as mixed traffic of people and vehicles, complex intersections, dynamic environments (such as temporary road construction, traffic control, roadside illegal parking), etc., which require extremely high real-time perception and planning capabilities of autonomous driving systems. In addition, the tolerance rate for misjudgment in urban environments is extremely low. Autonomous driving systems need to achieve low-threshold safe interaction and realize the full-link process of 'perception - decision - execution' in an extremely short time. These differences in factors pose more stringent technical requirements for Robotaxis autonomous driving systems. *This interview has been published in 21st Century Business Herald Reporter: Luo Yiqi, original title: Robotaxi Exploring Racing: Collective Efforts from the Supply Chain to the Ecosystem

Frost & Sullivan insights As more and more modern people choose to 'raise a furry child', the pet economy has risen rapidly, and the demand for pet insurance has also seen a surge. However, in the face of many problems in the pet insurance market, it has become an urgent task to promote the unification and standardization of industry standards. At present, what level is the development of pet insurance in China? What problems exist in the development process? Some consumers complain about difficulties in claims settlement, opaque compensation standards, and challenges in renewing policies for older pets. What are the main reasons for these problems? What is the current payout rate for pet insurance? What are the main reasons for the high payout rate? Is this also the reason why many insurance companies are cautious about entering the market?   Zhao Yuna, a research and consulting analyst at Frost & Sullivan (hereinafter referred to as 'Frost & Sullivan') Greater China, was interviewed by CBN to discuss the underlying logic behind the low penetration rate of pet insurance. *Click at the end of the article Read the original text View the full report   Q: At present, what level is the development of pet insurance in China? What problems exist in the development process? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan Although the current pet insurance market is growing rapidly, industry chaos frequently occurs, hindering its healthy development. Overall, we believe that pet insurance is currently in a relatively early stage of industry development, and there is still much to be done in the future.   On one hand, insurance companies set many restrictions on product design to control risks, such as a single payout cap, a 365-day waiting period for chronic diseases, and refusal of insurance for elderly pets. As a result, consumers actually receive far less protection than advertised.   On the other hand, due to the lack of unified standards for pet healthcare, insurance companies are skeptical about treatment costs and have strict review procedures, even refusing compensation on the grounds of 'congenital diseases' or 'non-veterinary drug labels'.   Meanwhile, some consumers engage in adverse selection behaviors (such as insuring with pre-existing conditions) or collude with pet hospitals to defraud insurance companies, further increasing the risk control pressure on insurers.   This vicious cycle has led to a decline in market trust: consumers complain about difficult claims processing, insurance companies reduce coverage due to high payment rates, and pet hospitals fall into disputes over costs.   In summary, in recent years, with changes in China's economic structure, population aging, and the development of atomized families, the number of pets kept has risen rapidly; China's pet market has also grown into one of the largest single markets in the world, leading to a corresponding rapid increase in insurance demand. However, at present, an effective trust relationship has not been established between the supply and demand sides of pet insurance, and the industry's development is not standardized. These factors have restricted the development of the industry.   Q: Some consumers complain about difficulties in claim settlement, opaque compensation standards, and challenges in renewing pet insurance policies for older pets. What are the main reasons for these issues? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan The current pet insurance market is in a vicious cycle and urgently needs industry regulation to break the deadlock. This predicament presents three contradictions:   Firstly, some irregular medical institutions engage in excessive diagnosis and treatment practices. There are also cases where users tamper with medical records to obtain insurance benefits, forcing insurance companies to set a single compensation limit to control risks;   Secondly, pet insurance faces serious adverse selection issues, with some individuals 'insuring while sick'. Coupled with the lack of pet health data and regular check-up mechanisms, this further drives up the payout rate.   These chaotic phenomena lead to a mutual loss for all parties: consumers pay high premiums but struggle to obtain high-quality services, medical institutions over-develop customer value for survival, and insurance companies weaken product competitiveness to prevent fraud. The key to breaking this deadlock lies in establishing a standardized pet diagnosis and treatment system, and by transparentizing medical practices and health data, reconstructing the market trust mechanism.   Q: Currently, what is the payout ratio of pet insurance? What are the main reasons for the high payout ratio? Is this also a reason why many insurance companies are cautious about entering this field? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan The current pet insurance industry is falling into a vicious cycle caused by dual non-standardization at the medical and insurance ends, becoming one of the types with the highest payout rates in the property insurance sector.   Data shows that additional compensation costs due to diagnostic and treatment chaos and user fraud account for more than 20%, causing some insurance companies to continue incurring losses. In the medical sector, the lack of unified standards during the diagnosis and treatment process has led to pet healthcare being like "opening a blind box," with large arbitrariness in treatment plans and dosages. Coupled with a chaotic medical pricing system, there are significant differences in the cost of diagnosing and treating the same condition, and there are even cases of insurance fraud where doctors collude with patients to prescribe high-dose medications to obtain compensation.   On the insurance side, traditional methods struggle to accurately identify the risk of similar pet breeds. Although some institutions have introduced nose print recognition technology, overall risk control remains weak. Additionally, most products adopt a uniform pricing mechanism for pets of different breeds, ages, and sizes, which not only violates the principles of actuarial science but also indirectly encourages high-risk groups to purchase insurance. Moreover, the information disclosure defect of insufficient disclosure of key terms in internet sales scenarios leads to a significant gap between consumers' expected protection and actual terms.   This medical chaos is driving up insurance costs, insurers tightening coverage, reducing consumer experience, and intensifying adverse selection, creating a chain reaction that is eroding the industry's development foundation. There is an urgent need to break this deadlock through measures such as establishing a standardized medical system, improving insurance actuarial models, and strengthening the disclosure of sales information.   Q: How can we avoid many problems emerging in the pet insurance market in terms of product design in the future? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan The healthy development of the pet insurance market in the future requires systematic optimization from two aspects: product design and service system. In terms of product design, efforts should be made to construct more transparent and precise protection plans: First, it is necessary to clearly define insurance clauses, especially by refining the criteria for defining vague concepts such as 'genetic diseases'. At the same time, the scope of coverage should be reasonably expanded to reduce unnecessary exclusions; secondly, big data analysis and AI technology can be considered to comprehensively assess risk factors such as pet breed, age, and past medical history, establishing a scientific differential pricing model to achieve precise risk pricing.   In terms of supporting service systems, it is necessary to establish a multi-party collaborative guarantee mechanism: Firstly, promote the formulation of unified pet medical industry standards to regulate diagnostic and treatment practices and charging rates, thereby reducing excessive medical care and fraud risks at the source; secondly, introduce third-party professional evaluation institutions to independently and fairly assess difficult claim cases, effectively resolving insurance disputes; furthermore, carry out insurance knowledge popularization education through online and offline channels to help consumers fully understand key terms such as waiting periods and compensation scopes, avoiding disputes due to cognitive biases.   For example, Ping An Property & Casualty Insurance has signed a cooperation agreement with the Shenzhen Institutes of Advanced Technology (Suzhou Branch) of the Chinese Academy of Sciences on nose print recognition technology. This collaboration aims to apply this technology to pet insurance, realizing commercial application of pet nose print recognition technology.   Q: The '2023 Pet Digital Insurance Industry Development Insights and Trends White Paper' shows that the domestic pet insurance market is still in its early stages of development, with an overall penetration rate of less than 1%. What is the outlook for the future development of pet insurance? What is the key to breaking through the next barrier? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan The pet insurance market in China is embracing unprecedented development opportunities. With the booming development of the pet economy and the continuous expansion of pet owners, this emerging market shows tremendous growth potential. However, compared with developed countries, the current penetration rate of pet insurance in China is still at a relatively low level, and market cultivation and consumer education urgently need to be strengthened.   In terms of channel transformation, internet platforms have rapidly risen to become the primary sales channel, leveraging their massive user base, precise big data analysis, and intelligent algorithm recommendations. This online trend has provided insurance companies with an efficient path to reach young pet owners, but it also poses higher demands on product design and customer service.   To achieve sustainable development of the industry, it is necessary to build a multi-party collaborative development system: Firstly, insurance companies should continuously optimize their product structure, improving user experience by simplifying claims processes and enhancing service response speed;   Secondly, it is necessary to deepen strategic cooperation with pet medical institutions and industry associations, jointly establishing standardized diagnostic and treatment standards and transparent pricing systems;   Finally, it is necessary to intensify market education efforts, combining online and offline methods to enhance consumers' awareness and acceptance of pet insurance value. Only through a multi-pronged approach including product innovation, channel optimization, and ecological co-construction can the development potential of the pet insurance market be fully unleashed, providing more comprehensive protection for pet owners and promoting the healthy and orderly development of the industry.   Q: Is industry standardization the key to future industry development? How should standardization be constructed? Zhao Yuna Research and Consulting Analyst, Greater China, Frost & Sullivan The innovation of products, technologies, and models is accelerating the improvement of service quality in the pet insurance industry. By establishing a sustainable cooperation model between insurance companies and pet hospitals and fully leveraging the connecting role of pet insurance, a virtuous cooperation chain can be formed: insurance cooperates with designated hospitals to attract patients; hospitals provide reasonable diagnostic and treatment services; insurance companies make timely payouts; pet owners obtain high-quality diagnostic and treatment experiences. Currently, the standardization work of pet insurance and pet healthcare in the market is gradually advancing. Recently, some insurance platforms have cooperated with pet hospitals and academia to jointly establish application norms for the pet insurance industry in terms of medical concepts and diagnostic standards, promoting the standardized development of the industry.   To establish a standardized pet healthcare industry, efforts need to be made in multiple aspects: first, standardizing the medical consultation process to ensure transparency; second, unifying procurement channel standards; third, standardizing disease classification and treatment plans. By integrating diagnostic and treatment data with insurance data, mining past data, and formulating more scientific standards, a basis can be provided for the pricing and design of insurance products. This will solve the problems of expensive and difficult pet healthcare, and build a higher-quality and suitable industry ecosystem.   Through industry collaboration, a system has been established from diagnosis and treatment norms to insurance industry standards. This includes the creation of a pet disease atlas, disease grading, definitions, treatment plans, medication dosages, and fee reference standards. These standards will provide strong support for the design of insurance products, actuarial models, and pricing systems. They will also precipitate platform claims review standards and hospital rating hierarchies, enabling users to experience more standardized, transparent, and certain services during the claims process. *This interview has been published in CBN Reporters: Huang Jingjing, Yuan Yuli. Original title: How is pet insurance moving towards standardized development under the backdrop of high growth prospects? | An exploration Standardized diagnosis and treatment are lacking, and fraud drives up compensation costs. The industry calls for differentiated pricing of pet insurance | A look   Contact number: 021-5407-5836 Contact email: PR@frostchina.com

Insights from Frost & Sullivan Several PCB industry chain companies such as Shanghai Electronics Group Co., Ltd., Pengding Holdings, and Shengyi Technology are expected to see increased performance in the first half of the year. The reasons mentioned in the attribution statements include structural demand driven by emerging scenarios such as high-speed computing servers and AI, an increase in the proportion of high-value-added products, and process improvements. What industry signals are being released behind this? Market reports indicate that the demand for high-end PCBs due to AI computing power is growing rapidly, and there is an obvious supply-demand gap this year. Leading manufacturers are accelerating production expansion. What is the current market supply and demand situation for related PCB products? It is reported that relevant departments have issued notices on implementing "window guidance" for computing power infrastructure and carried out preliminary surveys this year, aiming to avoid redundant construction. Will manufacturers in the PCB industry face challenges such as overcapacity after expanding production?   Xie Shuqin, Executive Director of Frost & Sullivan Greater China, was interviewed by Cailian News to discuss the new structural cycle of the PCB industry chain under the AI wave. *Click at the end of the article Read the original article View the complete report   Q: Several PCB industry chain companies such as Shanghai Electronics Group Co., Ltd., Pengding Holdings, and Shengyi Technology are expected to see increased performance in the first half of the year. The reasons mentioned in the attribution statements include structural demand driven by emerging scenarios such as high-speed computing servers and AI, an increase in the proportion of high-value-added products, and process improvements. What industry signals are being released behind this? Xie Shuqin Executive Director of Frost & Sullivan Greater China The PCB industry has entered a period of profound structural transformation. The core of this phenomenon lies in the reshaping of the industrial value chain by the computing power revolution - the explosive growth in the demand for AI large model training and inference directly drives the surge in demand for high-speed computing server PCBs. Such equipment requires high multi-layer boards with more than 18 layers to support the PCIe 5.0/6.0 high-speed transmission protocol, which is the essence of the "increase in the proportion of high-value-added products" by enterprises. With the acceleration of AI hardware iteration, the technical barriers for high-end PCBs continue to rise, and the industry concentration has significantly increased.   Coating board companies such as Shengyi Technology are accelerating the domestic substitution of high-frequency and high-speed substrates, breaking through the overseas monopoly of ultra-low-loss materials for AI servers and 6G high-frequency boards; Pengding Holdings, etc., are upgrading high-end HDI through micro-hole stacking technology to meet the miniaturization requirements of AIPC/AI mobile phones. Chinese policies are promoting the migration to high-end fields such as computing power, and local enterprises have made breakthroughs in the high-end field. The shortened technical iteration cycle further raises the competition threshold.   Q: Market reports indicate that the demand for high-end PCBs due to AI computing power is growing rapidly, and there is an obvious supply-demand gap this year. Leading manufacturers are accelerating production expansion. What is the current market supply and demand situation for related PCB products? Xie Shuqin Executive Director of Frost & Sullivan Greater China The hardware upgrade of AI servers drives a surge in demand for high-end PCBs. Taking NVIDIA's AI servers as an example, their core modules (such as UBB, OAM acceleration boards) require high multi-layer boards, and the value of a single PCBA is significantly higher than that of traditional servers. The global shipments of AI servers increased from 500,000 units in 2020 to 2 million units in 2024, with an annual compound growth rate of 45.2%, directly driving up the demand for high-layer boards and HDI boards. At the same time, the PCIe 6.0 protocol requires PCBs to support high-speed transmission rates and use ultra-low-loss materials, further raising the technical threshold and cost. On the supply side, there is a capacity bottleneck. First, technical barriers restrict production. High-end products require processes such as high-precision lamination and laser drilling, and only a few manufacturers globally can provide them. There is a short-term supply-demand gap for high-end products.   Currently, leading manufacturers are tilting their new production capacity towards products with more than 18 layers. In the medium and long term, with the release of production capacity in Southeast Asia and the commissioning of new high-frequency coating board production lines, the supply-demand contradiction may be alleviated. However, continuous technological iteration creates new demands. For example, Intel's next-generation Birch Stream platform requires more advanced PCB products, and the R & D reserve and capacity elasticity of manufacturers will become the key to competition.   Q: It is reported that relevant departments have issued notices on implementing "window guidance" for computing power infrastructure and carried out preliminary surveys this year, aiming to avoid redundant construction. Will manufacturers in the PCB industry face challenges such as overcapacity after expanding production? Xie Shuqin Executive Director of Frost & Sullivan Greater China Currently, AI computing power drives an explosive growth in demand for high-end PCBs, but there are significant structural contradictions between supply and demand, forming a dual pattern of "high-end shortage and low-end pressure". The hardware upgrade of AI servers and high-speed switches drives an annual growth rate of 16.7% in the demand for high multi-layer boards with more than 18 layers. The market scale increased from $1.3 billion in 2020 to $2.5 billion in 2024, but only a few manufacturers globally have stable mass production capabilities. Technical barriers make it difficult to fill the demand gap in the short term. Policy regulation is guiding production capacity to tilt towards high-quality fields, strictly controlling the expansion of inefficient production capacity, and strengthening technical thresholds. New projects are required to support the PCIe 6.0 protocol and advanced heat dissipation design, and the energy consumption per unit output value is included in the audit to promote green manufacturing. This precise regulation makes leading manufacturers' production expansion more rational.   The policy aims to shift the industry from "scale expansion" to "technology anchoring", building a new ecosystem with AI computing power as the engine and high-end manufacturing as the cornerstone. Only enterprises that accurately position themselves at the technological high ground can reap long-term dividends during the production expansion cycle.   Q: From the perspective of the PCB application market, which sectors currently have the greatest growth potential and why? Which PCB sub-categories and manufacturers are expected to benefit? Xie Shuqin Executive Director of Frost & Sullivan Greater China In the PCB application market, computing power scenarios (such as servers/data centers) and industrial scenarios (such as automotive electronics/communication equipment) have become the most growth potential areas due to technological explosion and demand expansion. The growth rate of computing power scenarios leads the way. Its PCB market scale is expected to increase from $12.5 billion in 2024 to $21 billion in 2029, mainly driven by the popularization of AI and big data upgrades. The training of large models drives up the demand for AI servers, and the growth of data center cabinets drives a surge in PCB usage.   At the manufacturer level, leading companies in computing power server PCBs such as King Kong Electronics, Kingtech, and Guanghe Technology are significantly benefiting from their technical advantages in high multi-layer boards; in the industrial scenario, Shanghai Electronics Group Co., Ltd. and Shenzhen Nan Circuit are leading in the fields of automotive electronics high-frequency PCBs and communication base station boards, while coating board supplier Shengyi Technology captures the dividends from high-frequency demand from the material side. *This interview has been published in Cailian News The reporter is Lu Tingting, and the original title was: AI Hardware Upgrade Spurs High-End PCB Demand, Industry Says Supply Chain Already Shows a Gap   Contact phone: 021-5407-5836 Contact email: PR@frostchina.com

'Frost & Sullivan' Insights On June 26th, the postpartum care group —— SAINT BELLA was listed on the Hong Kong Stock Exchange, becoming the first family quality care stock in Hong Kong. With star clients such as Qi Wei, Tang Yixin, and Gina staying at the center, and prices almost twice as expensive as ordinary maternity packages on the market, Saint Bella is hailed as a "Haute Couture" maternity center. Taking Shanghai as an example, the cheapest package is priced at 188,000 yuan. What services are included in this package, and why is this price more expensive compared to a maternity nanny? What are the reasons for consumers' choice? Overall, what is the acceptance level of maternity care among mothers in the past two years? Has consumption degradation and a decrease in childbearing affected it? What are the core competitive advantages? What expectations do mothers have been met with?   Li Junyi, consulting manager for Frost & Sullivan Greater China, was interviewed by Blue Whale News to discuss the changing needs of new-generation mothers . Blue Whale News Q: Taking Shanghai as an example, the cheapest package is priced at 188,000 yuan. What services are included in this package, and why is this price more expensive compared to a maternity nanny? What are the reasons for consumers' choice? What are the core competitive advantages? What expectations do mothers have been met with? Li Junyi Consulting Manager for Frost & Sullivan Greater China Maternity centers provide services mainly during the critical postpartum period to new mothers and newborns, offering postpartum care and support, including basic maternal and infant care, nutritional dietary adjustments, postpartum recovery, psychological counseling, and more. Although maternity centers and maternity nannies serve the same target group of new mothers and newborns, there are still differences in professionalism and service matrix provided. Compared to maternity nannies, maternity centers can equip clients with professional service personnel with diverse skills and a more comfortable and spacious living and recuperation environment, thus providing a more specialized, standardized, and comprehensive service matrix; in addition, compared to the limited service hours of a single maternity nanny per day, maternity centers have multiple professionals rotating shifts to ensure 24-hour uninterrupted health monitoring and service guarantees for mothers and infants. These are all core advantages of maternity centers that can better meet the health care needs of new mothers and newborns.   Q: Overall, what is the acceptance level of maternity care among mothers in the past two years? Has consumption degradation and a decrease in childbearing affected it? Li Junyi Consulting Manager for Frost & Sullivan Greater China In recent years, the penetration rate of maternity centers in China has steadily increased. Maternity centers are no longer exclusive to the wealthy or middle-class; many women from low- and middle-income families also choose maternity centers as their first choice for postpartum recuperation. Moreover, more and more young women have stated in surveys that although their income is not high, they consider using dowry to stay in a maternity center. And as women born in the 85s, 90s, and 00s become the main consumer group, they place more emphasis on professional care, scientific parenting, and a "pleasing oneself" experience, driving the transformation of maternity centers from "luxuries" to "necessities." *This interview has been published in Blue Whale News, with reporter Tu Jun, and the original title was: Maternity Center 'Haute Couture' Saint Bella Listed: Minimum Package 188,000 Yuan, 3 Years of Losses of 12 Billion

2025 World Artificial Intelligence Conference   On the afternoon of July 28, the closing ceremony of the 2025 World Artificial Intelligence Conference and the High-Level Meeting on Global Artificial Intelligence Governance was successfully held. Chen Jie, member of the Standing Committee of the Shanghai Municipal Party Committee and Vice Mayor, summarized the achievements of the conference. Zhang Ying, Deputy Secretary-General of the Municipal Government and Director of the Municipal Economic Informatization Commission, attended the event. Executives from Frost & Sullivan were invited to participate in this activity.   This year's conference brought together over 1,500 top experts from more than 70 countries and regions. The exhibition area exceeded 70,000 square meters for the first time, attracting over 800 enterprises to participate, showcasing more than 3,000 cutting-edge technologies, including over 100 "world premiere" and "China debut" products. The total number of exhibits and the number of premieres and debuts achieved a "double increase."     Dr. Wang Xin, Global Partner and Chairman of Frost & Sullivan Greater China Region   As a globally leading investment banking, corporate finance and growth consulting firm, Frost & Sullivan is committed to providing professional consulting services for industries with high growth potential, strategic importance, and leadership. Over the past 30 years, Frost & Sullivan has helped nearly 3,000 companies successfully list on global capital markets through its professional industry analysis, market insights, and investment and financing consulting services. Among them are many outstanding enterprises in the fields of artificial intelligence, semiconductors, robotics, and related areas.   According to LiveReport's big data statistics, as of July 24, 2025, there are 214 companies that have submitted listing applications to the Hong Kong Stock Exchange but have not yet gone public (including one company that has passed the listing hearing), among which 48 companies have submitted their prospectuses by industry classification for AI. From the perspective of industry consultants for AI companies planning to go public, Frost & Sullivan provides industry advisory services to the vast majority of AI companies, participating in 33 companies with an outstanding participation rate, accounting for 68.8% of the market share.   The World Artificial Intelligence Conference has been successfully held for its seventh edition in Shanghai. We are delighted to see that as a global hub for artificial intelligence, Shanghai is nurturing a cluster of world-leading AI enterprises with an open ecosystem and innovative vitality. Frost & Sullivan looks forward to walking alongside the pioneers on this hotbed - through in-depth industry insights, global resource connectivity, and digital strategy empowerment - to help enterprises achieve a leap from technological breakthroughs to business success.   Meanwhile, Frost & Sullivan itself will also actively embrace AI, integrating artificial intelligence technology with a more open attitude, upgrading its think tank research and consulting services capabilities, capturing innovation and technological opportunities in the global industrial transformation, and exploring the infinite possibilities of the intelligent computing economy with partners. The future has arrived. Frost & Sullivan is ready to join hands with AI pioneers to empower new growth with intelligence and define a new future with innovation!     Recommended Reading   Recent whitepapers related to the AI field 1. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the 'White Paper on the AI Smart Glasses Industry 2025'. 2. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the 'China Financial Large Model Market Tracking Report for the Whole Year 2024'. 3. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the '2025 AI Cloud Storage Solutions Market Report'. 4. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released 'Best Application Practices for Advanced Storage in China 2025'. 5. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released 'Insights into the Content Marketing Industry in the New AI Era 2025'. 6. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the '2025 China Large Model Annual Benchmarking Report'. 7. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released 'China Enterprise Outbound Cloud Services Market Report 2024'. 8. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the '2024 China GenAI Technology Stack Market Report'. 9. Frost & Sullivan, in collaboration with LeadLeo Research Institute, released the '2024 China AI Agent Annual List'. 10. Frost & Sullivan, in collaboration with LeadLeo Research, released the '2024 China Industry Cloud Hybrid and Private Cloud Market Report'. AI Industry Advisory Cases Frost & Sullivan assists Lens Technology in successfully listing on the Hong Kong Stock Exchange (6613.HK) Frost & Sullivan assisted Julong in successfully going public in the US (NASDAQ: JLHL) Frost & Sullivan assists Xiangjiang Electric Appliance in successfully listing on the Hong Kong Stock Exchange (2619.HK) Frost & Sullivan assisted Lianzhang Technology Holdings Limited in successfully going public in the US (LZMH:NASDAQ). Frost & Sullivan assisted Yipin Weike in successfully going public in the US (NASDAQ: EPWK) Frost & Sullivan assists Homeaway in successfully going public on the NASDAQ (NASDAQ:GMHS) Frost & Sullivan congratulates SinaVision Technology on successfully listing on the Sci-tech Innovation Board (688583.SH). Frost & Sullivan assisted INLF in successfully going public in the US (INLF.NASDAQ) Frost & Sullivan assists Innoventec in successfully listing on the Hong Kong Stock Exchange (2577.HK) Frost & Sullivan assisted Midea Group in successfully listing on the Hong Kong Stock Exchange (00300.HK) Frost & Sullivan helps Tianju Dihé successfully go public in Hong Kong (2479.HK) Frost & Sullivan helps YunGongTang successfully go public in Hong Kong (2512.HK) Frost & Sullivan assists UOB Holdings in successfully listing on the Hong Kong Stock Exchange (8529.HK) Frost & Sullivan assists MFS Group in successfully listing on the Hong Kong Stock Exchange (2556.HK) Frost & Sullivan assisted ZBAO.US, a leading provider of cloud-based insurance technology, in successfully going public in the US. Frost & Sullivan assisted Logica in successfully going public in the US (LGCL.NASDAQ) Frost & Sullivan assisted Ebuy Group in successfully listing on the Hong Kong Stock Exchange (9880.HK) Frost & Sullivan helps Beck Microelectronics successfully go public in Hong Kong (2149.HK) Frost & Sullivan assisted WILL's successful issuance of GDRs on the Hong Kong Stock Exchange Frost & Sullivan assists Conglian Technology in successfully going public in the US (ICG.NASDAQ) Frost & Sullivan assisted iRobot in successfully going public in the US (AIXI.US) Frost & Sullivan assists Kingsoft Cloud Holdings in successfully listing on the Hong Kong Stock Exchange (3896.HK) Frost & Sullivan assists Haojing Technology in successfully listing on the Hong Kong Stock Exchange (2440.HK) Frost & Sullivan assists Xuanwu Cloud Technology in successfully listing on the Hong Kong Stock Exchange (2392.HK) Frost & Sullivan assisted Huitongda in successfully listing on the Hong Kong Stock Exchange (9878.HK) Frost & Sullivan helps innovative startups successfully list on the Hong Kong Stock Exchange (2121.HK) Frost & Sullivan helps SenseTime Technology successfully go public in Hong Kong (0020.HK) Frost & Sullivan assists Qinhuai Data in successfully going public in the US (CD.NASDAQ) Frost & Sullivan assists Mingyuan Cloud in successfully listing on the Hong Kong Stock Exchange (0909.HK) Frost & Sullivan assists Century Group in its successful listing on the Hong Kong Stock Exchange (1849.HK) Frost & Sullivan assisted WeCom Group in successfully listing on the Hong Kong Stock Exchange (2013.HK) Frost & Sullivan assists Wan Ka Yilian in successfully listing on the Hong Kong Stock Exchange (1762.HK) Frost & Sullivan helps AsiaInfo Technology successfully go public in Hong Kong (1675.HK) Frost & Sullivan assists Hong Kong Asia Holdings in successfully listing on the Hong Kong Stock Exchange (1723.HK) Frost & Sullivan assisted Aurora Mobile in successfully going public in the US (JG.NASDAQ) Frost & Sullivan assists Crown Holdings on its successful listing on the Hong Kong Stock Exchange (8606.HK) Frost & Sullivan helps Qiyi Technology successfully go public in Hong Kong (1739.HK) Frost & Sullivan assists WeBank Jinko in successfully listing on the Hong Kong Stock Exchange (2003.HK) Frost & Sullivan helps HuiFu Tianxia successfully go public in Hong Kong (1806.HK) Frost & Sullivan helps Atlinks successfully list on the Hong Kong Stock Exchange (8043.HK) Frost & Sullivan assists Zioncom in successfully listing on the Hong Kong Stock Exchange (8287.HK) Frost & Sullivan assists ISP Global in successfully listing on the Hong Kong Stock Exchange (8487.HK) Frost & Sullivan assists Vobile in successfully listing on the Hong Kong Stock Exchange (3738.HK) Frost & Sullivan assists Abbot Technology in successfully listing on the Hong Kong Stock Exchange (2708.HK) Frost & Sullivan assisted iClick in successfully going public in the US (ICLK.NASDAQ) Frost & Sullivan assists Shengye Capital in successfully listing on the Hong Kong Stock Exchange (6069.HK) Frost & Sullivan assists Anling International in successfully listing on the Hong Kong Stock Exchange (8410.HK) Frost & Sullivan assists Ankox Systems in successfully listing on the Hong Kong Stock Exchange (8353.HK) Frost & Sullivan assists Junmeng International in successfully going public in Hong Kong (8062.HK) Frost & Sullivan assists Feida in successfully listing on the Hong Kong Stock Exchange (8342.HK) Frost & Sullivan helps Future Data successfully go public in Hong Kong (8229.HK) Frost & Sullivan assists Asia-Pacific Backup Systems Group Limited (8290.HK) in successfully going public in Hong Kong *The above order is not sequential; it is arranged in reverse chronological order based on listing time.