Development times reduced from weeks to days

Samsung Accelerates Chip Design with Claude Code

Samsung
Facebook
X
LinkedIn
Reddit
WhatsApp
Source: Arcansel/Shutterstock.com

Samsung uses Claude Code in the development of semiconductors. Despite time savings, unauthorized errors make human oversight indispensable.

The South Korean technology giant Samsung is increasingly relying on Anthropic’s AI tool Claude Code in its System LSI semiconductor division. As reported by the trade portal TechSpot, the company was able to drastically reduce the processing time for complex engineering tasks in the development and verification of microchips.

Ad

In one specific case, engineers used the AI system to check internal data connections of a custom system-on-chip (SoC). Despite incomplete documentation and missing design blocks, Claude Code created a virtual test environment. The task, which is normally scheduled for more than a month, was completed in about two days. When modeling USB interfaces and adapting Android drivers, working time was shortened from a month to just one day.

Serious errors and unauthorized code changes

Despite the massive leaps in efficiency, critical errors keep occurring in test operations, making constant double-checking by human developers imperative. In use, Claude Code exhibited the following misbehaviors, among others:

Masking error messages: Instead of fixing underlying programming errors, the AI independently downgraded error messages to harmless informational notices.

Ad

Unwanted undoing: When discarding a single feature, the tool accidentally deleted already completed, unrelated work results.

Unauthorized interventions: The AI repeatedly attempted to modify sensitive register-transfer-level (RTL) circuit diagrams without authorization.

Samsung: Offsetting personnel disadvantages compared to competitors

For Samsung, the use of generative AI tools in R&D is gaining strategic importance. The System LSI division employs around 6,000 people—significantly fewer than competitor Qualcomm with approximately 52,000 employees. Generative AI is intended to help compensate for this workforce difference in the development of new chip generations through higher productivity. In addition to Claude Code, Samsung also uses other AI systems such as Google Gemini and ChatGPT in research, manufacturing, and marketing.

(Editorial Team)

Ad

Weitere Artikel