Talk to China ▏CES 2024

LAS VEGAS, January 15 (TMTPost) -- In today's automotive industry, a transformation driven by technological advances is unfolding.

The advances in large models, artificial intelligence (AI), and simulation technology underpin the transformation of the smart travel industry. With widespread impacts not only on the convenience of transportation but also on various aspects of society, the industry harbors vast market opportunities.

As one of the highly anticipated events in the global technology community, the International Consumer Electronics Show (CES) kicked off last Tuesday in Las Vegas, Nevada.

As the only Chinese media hosting the MediaStage series of events at CES, TMTPost's CES Talk to China Stage invited Vice President Qiu Mingquan of XPeng Aeroht, Cluster Manager Shen Tao of Amazon Web Services in China, and Chief Marketing Officer Yang Yuxin of Black Sesame Technologies to join the "Infinite Possibilities of Smart Travel" in the Talk to China series at CES 2024 in Las Vegas.

From automakers to tech companies

This CES exhibition displays the technological trends and innovation of the automotive industry. Faced with the transition from automakers to technology companies, the technological opportunities within the industry are gradually emerging.

TMTPost founder & CEO Zhao Hejuan pointed out the close integration of the automotive and technology industries at this year's CES exhibition, emphasizing that traditional automotive companies are striving to become comprehensive technology companies. This viewpoint was echoed by Shen. According to him, automobiles will be the largest intelligent platform and comprehensive improvements in vehicle innovation and digital experiences are becoming a new trend.

Shen said that innovation in the automotive industry is focused on the intelligent reform of vehicles, especially in areas such as vehicle-to-everything (V2X) and autonomous driving technology. The combination of digital transformation and generative AI is expanding the innovation of the entire industry, enhancing the digital experience for car users. He also mentioned data processing and storage, citing BMW's autonomous driving platform built on Amazon Web Services (AWS), providing continuous computing power and AI services. This infrastructure support allows vehicles to iterate and upgrade continuously, meeting user demands and improving the overall experience through such changes.

With the integration of V2X and autonomous driving technology, ensuring the security of vehicle data has become a significant challenge for the entire industry. Shen also emphasized the importance of data security and compliance in the development of smart cars and shared examples of how AWS leverages its global user experience to provide over 300 security and compliance services and products. These services and products offer necessary support for automakers in balancing innovation and user data privacy.

Regarding the technological development of autonomous driving, Shen believed that the combination of algorithms and hardware is the key to advancing the technology quickly and steadily. He sees the multi-sensor data fusion (MSDF) and machine learning technology as the core of automotive intelligence and they drive the innovation of sensors, chips and internet connection technology. This cross-domain fusion of innovation is driving the widespread application of V2X, which in turn promotes advancements in autonomous driving technology.

From the perspective of flying car development, Qiu discussed the application of AI in flight control and design, as well as the advantages of simulation technology significantly shortening product development cycles and reducing trial-and-error costs.

Qiu said that AI technology has played a substantial role in the development of autonomous driving. He discussed the role of AI in the emerging flying cars field, including its application in the development of flight control systems and autonomous driving modes. He talked about achieving one-key vertical takeoff and landing and flying along a set route in the design of the autonomous driving system for flying cars, further illustrating the contribution of AI technology in reducing the difficulty of product usage.

In addition, Qiu mentioned that the combination of AI and human designers' creativity has maintained the diversity and innovation regarding the appearance of flying cars. He also emphasized the significant impact of simulation technology on reducing the research and development costs of traditional aviation products and shortening the development cycle, reflecting the trend of enhancing the R&D process through digital innovation.

Yang focused on the instrumental role of AI in the field of autonomous driving and discussed the challenges of mass production and engineering of L2 and L3 autonomous driving. Yang himself is optimistic about the application of Generative AI (AIGC) in the field of autonomous driving. He believed that through limited data collection and scene recognition, AIGC can achieve extensive coverage of corner cases, matching or surpassing human scene recognition capabilities.

Regarding chip technology, Yang illustrated the different purposes of training chips and inference chips. The former is used to generate large models, while the later for real-time computing on the edge.

In the future, specific-scene vertical large models are expected to emerge, requiring efficient computing at the edge by inference chips. Therefore, the challenge for chip companies is to design cost-effective solutions that can support these large models. He mentioned that Black Sesame Technologies is focusing on the development of inference computing chips, aiming to achieve efficient model inference on the edge with higher computing power.

Overall, the automotive industry has many opportunities in the development of technological integration, new AI applications, autonomous driving systems, and intelligent platforms. They are expected to significantly impact the future of smart travel, including enhancing safety, improving user experiences, and reducing costs. With the maturity of technology, these changes are expected to be realized in the near future and greatly drive the progress of smart travel.

For entrepreneurs, this landscape presents both challenges and opportunities

Zhao and the guests had in-depth discussions about the challenges faced in driving the smart travel industry.

Shen emphasized the impact of key areas such as computing power, data processing capability, global deployment, platform consistency, and data compliance and security management. He pointed out the significant demand for computing power in autonomous driving technology. This is because autonomous vehicles must process and analyze a large amount of real-time data from sensors, including but not limited to cameras, radar, and LiDAR.

Furthermore, to improve the efficiency of algorithms, it is necessary to store and analyze massive amounts of data to train machine learning models, enabling them to effectively identify patterns and objects in the environment. Shen believes that this demand for high-performance computing resources is a key challenge for the successful implementation of autonomous driving technology.

Shen also discussed the issues that need to be considered when deploying smart travel solutions globally and how to ensure the consistency and uniformity of global services to adapt to the needs of different markets. He mentioned that when introducing products and services in different countries and regions, companies will encounter various complex differences in regulations, culture, and infrastructure conditions. Shen believes that providing a consistent service experience globally while complying with local regulations and market characteristics is one of the major challenges facing existing smart travel companies.

Shen then proposed solutions from the perspective of cloud computing, which provides services ranging from computing power to storage and big data analysis. It supports the needs of automotive and technology companies with tremendous flexibility, especially in the fields of V2X and AI. Thanks to the global data center network, manufacturers can efficiently and rapidly deploy solutions worldwide. In terms of security and compliance, Shen pointed out that AWS sets the industry benchmark with its highest level of service responsibility model and helps companies quickly deploy solutions to target markets by meeting compliance requirements in different regions.

For entrepreneurs, Shen sees challenges as opportunities. He mentioned that cloud computing platforms can provide technical support for entrepreneurs and help them bring their business solutions to market. Technology stacks such as cloud infrastructure, data platforms, artificial intelligence, and machine learning enable startup companies to quickly build their workloads. At the application level, startups can rapidly create their solutions at the foundational, intermediate levels, and application, which is crucial for innovation and technological empowerment.

Qiu focused his speech on the challenges in the field of flying cars. He believed that although autonomous driving technology in areas like fixed-wing drones has developed quickly, there are still many technical challenges before achieving fully autonomous commercial flying cars. Especially the takeoff and landing phases are the most challenging parts for the autonomous driving technology of aircraft and often the frequent stages of aviation accidents. Solving these technical challenges requires strict adherence to aviation safety rules, precise control algorithms, and complex system redundancy designs.

Qiu particularly emphasized the importance of redundancy design in ensuring flight safety. This approach allows the aircraft to continue operating even when some systems fail, ensuring the safety of the flight. He mentioned that the aircraft that their developing possesses overall multi-parachute rescue system, aiming to provide a safe means of rescue in emergency by rapidly applying parachutes to reduce the risk of crashes.

Regarding the specific development status of flying cars, Qiu mentioned that the company's developed flying car currently allows switching between manual and autonomous driving, with higher authority given to human drivers. He predicted that products to be launched in the next few years will still have the option of manual driving, but the subsequent development will focus on autonomous driving. This means that flying cars will no longer require manual control systems and can automatically fly after setting the destination. He emphasized that the main driving force is not just for fad but based on practical needs and application scenarios.

Yang highlighted that the ultimate goal of smart travel is to achieve autonomous driving. However, this requires not only technological development and breakthroughs but also coordination with policies, regulations, and traffic systems. He pointed out that the direction of smart travel is not just about making vehicles smarter but also making the entire road network intelligent. In addition, the intelligent traffic management system in the cloud is a key to accelerating the realization of fully autonomous driving.

As a startup company, the goal of Black Sesame Technologies is to advance the evolution of the entire vehicle's electronic and electrical architecture and drive technological innovation with chips, combining functions and offering more economical vehicles.

Regarding entrepreneurial opportunities, Yang suggested that entrepreneurs focus on specific technological innovations and application scenarios, such as the application of generative AI in intelligent driving cabins and autonomous driving, or the progress in chips and electronic and electrical architecture. He also emphasized that startup companies need to balance technological innovation and cost-effectiveness, making their products and services efficient and attractive to customers.

The development of smart travel faces significant technological challenges but also presents unprecedented opportunities, requiring comprehensive strategies and decisions from the industry, government, and enterprises. The discussions of the guests provide a multi-angle, multi-dimensional perspective to understand this rapidly changing field, offering guidance for industry participants to transform challenges into opportunities.

Widespread adoption of autonomous driving technology still needs to overcome the cost barrier

Discussing the cost of the widespread adoption of autonomous driving technology, Yang considered the battery cost as the current largest part of costs for vehicles. Therefore, increasing energy density is a key challenge in the field of materials science. As electronic systems gradually centralize, hardware costs may also decrease. He explained that although the internal electronic systems of vehicles are becoming more complex, theoretically increasing costs, a centralized architecture can reduce electronic hardware costs.

As for software costs, Yang believed that there is also the possibility of a decrease, especially as basic programming work is gradually replaced by AI, and only innovative architectures and algebraic coding still require human effort. Yang sees this as an opportunity for structural cost reduction achieved through technological innovation rather than simply compressing the supply chain. He emphasized that innovation, especially from non-traditional automotive industry perspectives, has the potential to bring about structural cost improvements during the industry's rapid evolution.

Qiu, in discussing the cost of flying cars, emphasized the significance of shortening the R&D cycle. He mentioned the use of AI and simulation technology to accelerate the development process, thereby reducing the costs of the R&D phase, avoiding operational difficulties caused by excessively high labor and time costs.

In his view, the biggest difference between XPeng Aeroht and the traditional aviation industry is that they leverage a mature supply chain for intelligent electric vehicles to reduce costs and independent R&D can be used as a means to reduce dependence on expensive external procurement. In terms of scaling, he proposed achieving a virtuous cycle by reducing costs and prices, expanding sales scale, and building a comprehensive transportation network. This allows flying cars to serve a broader audience, not limited to specific customers or usage scenarios.

Zhao pointed out that autonomous driving is not only a technological and cost issue but also a challenge for policy makers and regulators. This also creates new opportunities for entrepreneurs who need to find new business models and service paths in the continuously evolving policy environment.

Yang argued that future computing demands will increase with the improvement of intelligent management, not only limited to automobiles but also including all mobile intelligent terminal devices such as flying cars, ships, and robots. In other words, any "central brain" capable of processing these massive sensor data is a powerful support for cost reduction.

Qiu also emphasized the importance of digital airspace management and air traffic management systems, as these are the foundations for ensuring the safe operation of future diversified and high-density airspace traffic. He cited the example of Shenzhen's low-altitude safety traffic system research, demonstrating the city's conceptual consumption of future airspace while playing a demonstrative role.

In summary, reducing the cost of autonomous driving and flying cars requires multiple perspectives, including advances in materials science, integration of hardware systems, the application of AI in software programming, shortening the R&D cycle, leveraging the supply chain, promoting independent R&D, and promoting mass production, as well as digital management. At the same time, this process is accompanied by the gradual maturity of policies and regulations and the continuous innovation of entrepreneurs. (Author | Chang Xiao, Editor | Zhang Min)