Jens Hinrichsen, Executive Vice President and General Manager, Analog and Automotive Embedded Systems at NXP Semiconductors, spoke at COMPUTEX 2025 on May 20, 2025, about the company’s advancements in bringing AI to edge devices.
A world with intelligent machines that reason, process information, and complete tasks on their own “is no longer a sci-fi novel, it is on the cusp of becoming a reality,” Hinrichsen told the crowd at Taipei Nangang Exhibition Center.
“An autonomous future will make us more productive, more sustainable, safer, because of AI,” Hinrichsen said.
The vision he laid out depends on edge devices being locally able to process, compute, and analyze their surroundings without depending on the cloud.
Hinrichsen pointed out that despite the focus of AI in cloud computing, most of the information originates on the edge, a half trillion GB of data every day across the world’s edge devices.
Hinrichsen highlighted several of the benefits of AI at the edge, such as reducing bandwidth, allowing for local processing in remote areas with limited connectivity and coverage, real-time computing, energy efficiency, and data privacy.
“That is the edge, where AI meets the physical world,” he said.
He cited examples of edge AI applications like smart wearables alerting emergency responders of a medical emergency or autonomous vehicles eliminating traffic fatalities.
Hinrichsen demonstrated how the pace of progress from cloud technology moving to the edge has become faster, from the decade it took perception AI to make the transition, to five years for generative AI, to only one year for agentic AI.
With agentic AI, the edge “becomes reactive rather than proactive,” and allows for devices to respond independently, safely and securely.
Comparing agentic AI to a human brain, he pointed out how agentic AI models are not limited to the datasets they are originally trained on but can learn from their own failures and successes.
“Agentic AI is a paradigm shift from something we use to something that works alongside us,” Hinrichsen said.
Looking ahead, Hinrichsen called attention to the next challenges in the industry – the need for complex integration in edge AI systems, models specifically designed for edge devices, and the need for greater safety and security.
He then emphasized NXP’s products’ ability to respond to each of these challenges: integrating hardware building blocks and software foundation, providing tools to partners to help them develop their own models, and leveraging industry expertise to ensure systems are secure.
“Together we make the autonomy real, scalable, and trusted,” Hinrichsen said about the need for not just NXP, but all participants in the edge AI ecosystem to work together and enable this transformation.
Although he did not make a prediction for what AI technology would be the next evolution after agentic AI, Hinrichsen ended by emphasizing that “NXP and our partners will be right there to bring it to reality.”
