NVIDIA’s CEO, Jensen Huang, is looking toward the future with an optimistic view on GAA (Gate-All-Around) technologies. These cutting-edge processes are expected to deliver substantial performance improvements, with Huang estimating a potential 20% boost in a given generation. Although the specifics remain somewhat general, the implication is clear: GAA could redefine performance standards, particularly for large-scale AI operations.
NVIDIA is poised to feature GAA technologies within TSMC’s impending N2 process, which will likely play a crucial role in the company’s upcoming Feynman AI architecture, set for release in 2028. Despite the skepticism surrounding Moore’s Law, NVIDIA consistently strives to challenge conventional performance benchmarks, focusing instead on key architectural advancements to achieve significant gains.
The adoption of GAA technology has been met with some uncertainty within the chip industry. While Samsung was the pioneer in announcing a complete GAA-based process with their 3nm technology, they have struggled to meet industry-standard yield rates. This leaves the field open for TSMC, which plans to incorporate GAA in its N2 (2nm) line.
Unlike some of its competitors, NVIDIA tends to wait before embracing the latest chip processes, allowing the technology to mature and stabilize in the marketplace. This strategy aligns with what Huang calls “Huang’s Law,” emphasizing that true performance advancements come through innovative architectural designs. Over the last decade, this approach has allowed NVIDIA to amplify chip efficiency by more than a thousandfold.
In sum, while GAA promises to enhance performance significantly, NVIDIA seems committed to leveraging its architectural prowess to remain at the forefront of technology, continually pushing the boundaries of what is possible in AI and beyond.
