January 13, 2026
news
Phison Electronics has unveiled enhancements to its aiDAPTIV+ technology that extend its advanced AI acceleration to an integrated GPU framework to accelerate inference, increase memory capacity, simplify deployment, and unlock AI capabilities for larger models on laptops, desktops, and mini PCs.
aiDAPTIV+ addresses the challenges of growing data volumes, complex AI training, and inference workloads by leveraging NAND flash as memory to remove compute constraints while enabling on-premises inference and optimizing large-scale models.
“As AI models grow to tens or even hundreds of billions of parameters, the industry continues to run into the same wall of GPU memory limitations,” said Michael Wu, president and general manager of Phison US. “By expanding GPU memory with aiDAPTIV+’s high-capacity flash-based architecture, we give everyone from consumers to small businesses to large enterprises the ability to train and run large models on affordable hardware. In effect, we are turning everyday devices into supercomputers.”
For more information, please visit phison.com/en/.
Chad Cox. The Embedded Computing Design Production Editor’s responsibilities include handling the news cycle, newsletters, social media, and advertising. Chad graduated from the University of Cincinnati with a BA in Cultural and Analytical Literature.
Other works by Chad
