NVIDIA CEO Jensen Huang Predicts 20% Performance Boost from GAA Technology, Emphasizes Architectural Innovations

NVIDIA CEO Predicts Modest Gains from GAA Technology

NVIDIA CEO Jensen Huang has shared his thoughts on the future of semiconductor technology, particularly focusing on Gate-All-Around (GAA) transistors. During a Q&A session at the GPU Technology Conference (GTC), Huang suggested that GAA-based technologies could bring about a 20% performance boost for NVIDIA's processors 1

1

Emphasis on Architectural and Software Innovations

While acknowledging the potential benefits of GAA technology, Huang emphasized that the most significant performance improvements for NVIDIA's GPUs come from architectural advancements and software innovations. He noted that as AI systems scale, the efficiency of managing vast numbers of processors is becoming more critical than the raw performance of individual processors 1

1

NVIDIA's Approach to Process Node Adoption

Unlike some competitors, NVIDIA typically does not adopt the latest process technologies immediately. The company has been using tailored versions of TSMC's 4nm-class process technologies for its current generation of GPUs. Looking ahead, NVIDIA is expected to use TSMC's 3nm-class fabrication process for its next-generation AI GPUs, codenamed Rubin, in 2024 1

1

Future GPU Architectures and GAA Implementation

Speculation surrounds NVIDIA's plans for implementing GAA technology in future GPU architectures. The Feynman architecture, expected in 2028, might be the first to utilize GAA-based processes, potentially adopting TSMC's N2P or A16 nodes. These advanced nodes are projected to offer significant performance improvements over current technologies 2

2

Beyond Semiconductor Manufacturing

Huang stressed that NVIDIA has evolved beyond being simply a semiconductor company. He described NVIDIA as a provider of large-scale AI infrastructure and a leader in algorithm development for various fields, including computer graphics, robotics, and computational lithography 1

1

Related Stories

Nvidia Set to Pioneer TSMC's Advanced A16 Node, Reshaping AI Chip Landscape

Nvidia CEO Claims AI Chips Outpacing Moore's Law, Promising Faster AI Progress

Jensen Huang teases AI chip that will 'surprise the world' at Nvidia's GTC Conference in March

Nvidia CEO Jensen Huang Unveils New AI Technologies at Computex 2025

Industry Perspective on GAA Technology

The chip industry has shown some uncertainty regarding GAA adoption. Samsung, the first to announce a fully developed 3nm GAA process, has faced challenges in achieving industry-standard yield rates. TSMC is expected to introduce GAA with its N2 (2nm) lineup of processes 2

2

NVIDIA's Performance Scaling Strategy

NVIDIA's approach to performance scaling, dubbed "Huang's Law" by some, focuses on driving performance through architectural advancements rather than relying solely on process node improvements. This strategy has allowed the company to scale up chip efficiency by over 1,000x in a decade 2

2

Implications for the AI and GPU Markets

As NVIDIA continues to push the boundaries of GPU and AI technology, the company's stance on process node advancements and its focus on architectural innovations could have significant implications for the broader AI and high-performance computing markets. The balance between leveraging new manufacturing technologies and optimizing existing architectures will likely remain a key factor in NVIDIA's strategy moving forward.