On June 2, 2026, COMPUTEX 2026 officially opened at the Taipei Nangang Exhibition Center under the theme “AI Together.”On June 1, NVIDIA CEO Jensen Huang delivered a nearly two-hour-and-twenty-minute GTC Taipei keynote at the Taipei Music Center. With his bold vision of “reinventing the PC,” he captivated the audience and ignited widespread excitement across the industry. Many observers have even compared the significance of this moment to the historic impact of Steve Jobs unveiling the iPhone.

Jensen Huang Declares the Full Arrival of the Agentic AI Era 

Dressed in his signature black leather jacket, Jensen Huang opened his keynote with the words, “It’s great to be home,” before making a bold statement: the personal computer is being reinvented.

He proclaimed that the era of Agentic AI has arrived, with tokens becoming a new unit of value creation. As he put it, “Compute is revenue. Compute is profit.”

To illustrate the explosive growth of AI-driven development, Huang cited GitHub statistics showing that code commits surged from 500 million to 1.4 billion in the first few months of 2026—nearly three times the total recorded for all of last year.

He further emphasized that AI agents operate in a “nanosecond-scale world,” while traditional CPUs were originally designed for humans living in a “second-scale world.” To meet the demands of this new computing paradigm, NVIDIA introduced its purpose-built Vera Rubin Platform, delivering up to 1.8× the performance of x86 processors and offering memory bandwidth of up to 1.2 TB/s.

Jensen Huang Says NVIDIA and Microsoft Will Fully Reinvent the PC

One of the keynote’s biggest highlights was the unveiling of the RTX Spark chip, designed specifically for next-generation Windows AI laptops. Featuring a 20-core Grace ARM CPU, 6,144 Blackwell CUDA cores, and up to 128GB of unified memory, RTX Spark delivers AI performance of up to 1 PFLOPS (one quadrillion operations per second). The breakthrough platform effectively brings supercomputer-class AI computing power into a single chip, opening new possibilities for on-device AI inference, content creation, and intelligent agent applications.

During the exhibition, he further noted that AI is evolving from generative AI toward Agentic AI, where systems are capable of understanding goals, perceiving their environment, planning tasks, and executing actions autonomously—shifting from “tool-based AI” to “action-based AI.”

The Vera Rubin platform is designed to enhance inference and agent efficiency, enable near real-time processing, and improve energy efficiency to address the growing challenges of data center computing power and power consumption.

Shared Love for Chicken Nuggets with the Legend

Meanwhile, Rene Haas, CEO of Arm, stated that as AI applications continue to scale, global demand for computing power is growing rapidly. Future computing platforms will need not only stronger intelligent processing capabilities, but also improved energy efficiency and sustainability. This trend is accelerating the upgrade of the entire computing ecosystem, from the cloud to the edge.

For the industry as a whole, the rise of Agentic AI not only represents more powerful model capabilities, but also marks AI’s expansion from the cloud into the edge and the physical world. Robotics, industrial equipment, smart terminals, and various edge computing devices are increasingly becoming key carriers for AI deployment.

In parallel with the grand narrative of AI giants, Rocktech , a representative enterprise in edge AI hardware, also showcased its major new products at this year’s exhibition.

During this year’s COMPUTEX, Rocktech exhibited its next-generation core boards built on the MediaTek Genio 720 and Qualcomm QCS6490 platforms.

These two platforms are designed for edge AI, robotic control, machine vision, smart retail, industrial automation, and intelligent terminal applications, further expanding Rocktech’s product portfolio in high-performance edge computing.

The launch of the new platforms reflects Rocktech’s continued commitment to the evolution of AI computing architectures, while strengthening its presence in the Arm ecosystem, edge intelligence, and industry applications.

As demand grows for on-device inference, multimodal processing, and real-time decision-making at the edge, the company is accelerating the development of a computing foundation for next-generation intelligent terminals.

In the future, Rocktech will continue to focus its technological development on edge intelligence and robotic systems, with an emphasis on edge AI platforms, multi-sensor fusion, real-time decision-making systems, robotic control platforms, and the deployment of Agentic AI at the device level. These efforts aim to further accelerate the transition of artificial intelligence from the digital domain into the physical world.