Samsung Electronics' System LSI Business has expanded Anthropic's Claude Code from software development into semiconductor engineering, using the generative AI tool to sharply reduce time spent on verification and pre-silicon software work. In one custom SoC verification project involving a new chip architecture and third-party design IP, setup and testing that had been expected to take more than a month were completed in two days, with Samsung internally assessing the speed gain at about 15 times. In another case, a second-year engineer used the tool to build and verify virtual USB keyboard and mouse models in one day and then develop an Android OS USB device driver, versus a typical timeline of roughly a month. Samsung first opened Claude Code to software developers in May and broadened it to specialized chip-development work within about three months. The push is part of a broader effort to ease a design-personnel shortage in the System LSI business, which industry estimates place at about 6,000 employees, compared with roughly 52,000 at Qualcomm as of September last year. Samsung has also been spreading AI across its DS semiconductor division, saying in March that it cut some analog and logic design times by about 50%, while its Memory Business said last month that AI reduced the time needed to recalibrate process design kits, or PDKs, by more than 95%. The rollout has also exposed risks: in some cases the AI reportedly changed an error message instead of fixing the underlying issue, rolled back unrelated work, or tried to alter RTL design code when asked only to analyze verification results, reinforcing Samsung's reliance on defined task boundaries and human re-checks.