EN / 中文

NXP Yuan Wenbo: Robotics Semiconductor Market to Hit $100B by 2040 via Neural Axis Architecture

by zhongguodianzibao·November 17, 2026

On November 12, the 23rd China International Semiconductor Expo (IC China 2026), co-hosted by the China Center for Information Industry Development and the China Semiconductor Industry Association, opened at the National Convention Center in Beijing. The 8th IC Entrepreneurs Conference (hereinafter referred to as "the Conference") was held concurrently. During his keynote speech at the Conference, NXP Semiconductors Vice President Yuan Wenbo stated that for the robotics industry to truly achieve scale, it requires building trustworthy Physical AI through a neural axis architecture.

Yuan Wenbo predicts that the robotics semiconductor market will grow rapidly over the next decade. Citing UBS forecasts, he noted that the market size is expected to reach approximately $6 billion around 2030, potentially expand to between $20 billion and $50 billion by 2035, and approach $100 billion by 2040, possibly surpassing automotive semiconductors to become one of the most important edge-side scenarios.

However, market prospects do not equate to industrial maturity; the robotics industry is currently in a critical phase of scaling up mass production. While the industry has previously focused on feature stacking, demo showcases, and motion capability validation, challenges in the upcoming industrialization stage will concentrate on foundational capabilities such as system latency, functional safety, information security, and energy efficiency management. Additionally, details regarding the maturity of human-machine interaction, complementary capabilities of cameras with millimeter-wave radar, UWB, and wireless communication, as well as the flexibility and durability of mechanical structures, still require further improvement.

Consequently, NXP drew inspiration from the human nervous system to propose the Neural Axis Architecture, where the brain handles reasoning and decision-making, the cerebellum manages coordination, and the spinal cord is responsible for local rapid reflexes. Yuan Wenbo pointed out that robots are not solved by "a stronger brain chip" alone; rather, they resemble complete living systems requiring hierarchical collaboration among the brain, cerebellum, and spinal cord. He particularly emphasized the importance of the spinal cord: "In daily life, when humans are touched, the brain or cerebellum does not need to perform operations. It is crucial that neural reflexes provide a rapid response first, with the spinal cord enabling immediate action through local processing capabilities."

Under this layered architecture, NXP is focusing on three key pillars. The first pillar is Time-Sensitive Networking (TSN). Yuan Wenbo stated that robot internals currently suffer from fragmented network standards and technologies. As cameras, sensors, and multimodal information increase, the value of TSN lies in upgrading standard Ethernet to a deterministic real-time network, unifying bandwidth design and scheduling capabilities, thereby transforming internal robot communication from "as fast as possible" to "predictably fast."

The second pillar is the functional restructuring of the cerebellum. Yuan Wenbo emphasized that while the brain must handle millisecond-level reasoning and decision-making under strict energy efficiency constraints, the cerebellum needs to upgrade from traditional motion control functions to a real-time coordination hub for robots, integrating motion control, real-time networking, and safety monitoring to undertake deterministic real-time control.

The third pillar is the spinal cord, whose significance lies in sinking intelligence to the edge, enabling robots to possess local rapid reaction capabilities. Using dexterous hands as an example, Yuan Wenbo explained that today's dexterous hands primarily perform data collection and action response—sending detection data back to the brain, which then commands subsequent actions. Next-generation dexterous hands require more edge computing power, combining visual-tactile sensors, torque, and other multimodal information with models, sensors, and chips to achieve local closed-loop functions such as slip detection, grip adjustment, and collision response. This optimizes energy management and enables real-time feedback in dangerous or special situations. These three pillars collectively build a modular and scalable architecture that serves not only humanoid robots but also extends to Autonomous Mobile Robots (AMRs) and can be reused in other Physical AI forms like automobiles.

In Yuan Wenbo's view, the common foundation supporting the Neural Axis Architecture is safety performance. He noted that as Physical AI expands from industrial robots to home applications and enters mass production, functional safety and information security must permeate the entire chain—from perception and networking to computing—beyond just cost-effectiveness and manageability. Functional safety ensures robots do not lose control or harm people; information security ensures devices are not attacked and data is not tampered with. It is understood that NXP protects deployed devices through its EdgeLock 2GO information security system and advances post-quantum security certification to address potential risks from quantum computing, meeting stricter global standards.

"China plays an irreplaceable core role in the global robotics supply chain," Yuan Wenbo stated. NXP is upgrading China from a sales center to a source for product definition, local R&D, and supply chain collaboration, consistently implementing the collaborative innovation philosophy of "In China, For China; In China, For the World." In 2026, NXP established a Robotics Innovation Center, forming a local closed loop for R&D, marketing, and product definition around its China Business Unit, and will continue to expand its ecosystem partnerships to jointly promote the prosperity and development of the entire industry.

Author | Wang Xinhao

Editor | Wu Lilin

Layout | Ma Liya Supervisor | Lian Xiaodong