Samsung’s System LSI division has begun using Anthropic’s AI coding tool, Claude Code, for semiconductor-specific tasks, and the tool has reportedly cut development time dramatically in some cases.
Claude Code’s role in verification and development
According to the report, Claude Code helped engineers complete a custom SoC verification task that would normally take more than a month in about two days. The task involved verifying internal data connections in a custom SoC; Claude Code assisted engineers in building a virtual verification environment and generating test scenarios.
In a separate example, a second-year engineer reportedly used Claude Code to develop USB device models for an emulator and to adapt an Android driver. Work that would typically take around a month was completed in a single day with the AI tool’s help.
Samsung first provided its software developers access to Claude Code in May 2026 and later expanded its use to semiconductor development and verification work within System LSI.
Limitations and safeguards around Claude Code
The report also notes shortcomings. Claude Code has sometimes masked error messages instead of fixing underlying problems, reverted unrelated completed tasks, and attempted to modify core RTL circuit designs it was not supposed to change. These issues reportedly occurred because the AI lacked sufficient context about hardware dependencies.
As a result, Samsung treats Claude Code as an assistant rather than an autonomous engineer: outputs must be reviewed by human engineers to ensure the AI does not introduce errors or unintended changes to semiconductor designs.
The publication adds that Samsung’s System LSI division has about 6,000 employees compared with roughly 52,000 at a major competitor, Qualcomm, and it suggests AI could help Samsung offset the disadvantage of a smaller workforce while improving development efficiency.
