TodayToday's edition
Samsung Foundry's claims for 8LPU+ also include an 8% smaller logic area, on existing DUV lines, per Automotive World on Oct. 21. No customer or ramp date has been named.
1 min read
Illustrative graphic from Samsung's foundry process-technology page; it does not show an 8LPU+ product.
Image: Samsung Semiconductor
Source · Press image from the Samsung Semiconductor website, used for news reporting.
Samsung Foundry has launched 8LPU+, a second-generation extension of its 8nm process, according to Automotive World on Oct. 21. Samsung says the node delivers 12% more performance, 15% lower power consumption and an 8% smaller logic area than the previous generation. Those are vendor claims, and the report cites no independent measurements.
The process is aimed at automotive, robotics and edge AI chips that do not need the most advanced nodes. It runs on existing deep ultraviolet (DUV) lines, so customers avoid the cost of extreme ultraviolet (EUV) lithography, Automotive World reported.
Samsung's claimed gains for 8LPU+ over the previous generation, as reported by Automotive World; not independently measured.
Illustration (AI-generated)
Carmakers are shifting to more centralized electronic architectures that need more processing power to control vehicle functions, the outlet said. It argued that much of this work does not call for 2nm-class silicon. Samsung's 8nm node already has an automotive record, since Nvidia's Drive Orin is built on it, according to the same report.
Voltgrid News reported earlier that Samsung is moving a modem to 2nm, a shift that frees 4nm capacity for HBM base dies.
The report names no customer or product on 8LPU+ and gives no volume-production date. The percentage gains come from Samsung, and the sources read do not say which earlier process is the baseline beyond calling it the previous generation. Automotive World also said Samsung is struggling to keep pace with TSMC at the leading edge. DigiTimes also covered the launch, but its full report is behind a paywall and was not used for figures here.
