Normal Computing Corporation

Normal Computing is an applied AI company building physics-based accelerator chips for AI inference (Normal ASICs) and an AI-native electronic design automation stack (Normal EDA). The company states that its ASICs use stochastic, stateful and asynchronous computation, and that its first thermodynamic compute chip, CN101, was taped out in June 2025.

The website describes Normal EDA as AI-accelerated RTL design and verification software, deployed on-premises and trained on customers' design data. It says the company works with the world's largest semiconductor companies and with advanced institutions, and that pilot deployments of its hardware are open. Pricing and the detailed revenue model are not described.

The About page lists offices in New York City, San Francisco, London, Copenhagen and Pangyo. The website reports more than $85M in funding, including a strategic round led by Samsung Catalyst. Open job listings cover silicon and hardware engineering, AI research, software and recruiting, with locations including New York, London, Mountain View, Copenhagen and Zurich. The founding year is not stated.

Product: Normal ASICs: physics-based thermodynamic accelerator cards for AI inference; Normal EDA: AI-accelerated RTL design and verification software deployed on-premises.