In recent years, the main focus of the OLED industry was on reducing power consumption. OLED developers introduced several novel technologies, including new backplanes, materials and device architectures, to increase device and system efficiency.
This article looks into the potential pressure that the acceleration of on-device AI (locally running AI engines as opposed cloud-based AI) will have on OLED developers to reduce the power consumption even further, especially within the laptop segment, and adopt novel next-gen technologies.
Every generation of on-device AI silicon asks for more power than the last. Model sizes on flagship devices have gone from roughly 100M parameters in 2020 to 3–7B today, with 13–20B expected on premium hardware by 2028. Each doubling roughly doubles compute and memory-bandwidth demand. NPUs are dramatically more efficient than running the same workload on a CPU or GPU, but they still consume a lot of power.
To read the entire article, sign up for OLED-Info Pro
OLED-Info Membership Benefits:
- Access to premium content
- A comprehensive guide to the OLED industry
- OLED Insights and trends analysis
- AI-powered assistant built specifically for the OLED industry
- A content library of 250+ brochures, catalogs, roadmaps, presentations and more.
- Detailed structured information on OLED devices, fabs, microdisplays, automotive applications and more
- Online access, easy subscription model, cancel at any time
- Only $199.99 per year (with group discounts)
