Normal Computing

Normal Computing develops AI software for hardware engineering and electronic design automation (EDA) solutions, building a new paradigm of AI hardware powered by novel device physics. The company offers two core product lines: Normal EDA, an AI-accelerated co-design platform for silicon engineering teams that is in production with the world largest semiconductor companies, and Normal ASICs, custom silicon built on new device physics targeting 10-100x gains in AI inference per dollar, per watt. Normal EDA builds fast AI-native simulation and synthesis engines for rapid experimentation and optimization, with agents that learn from every run and close the loop from architecture to signoff. The platform features structured representations of design through an Ontology that grounds every downstream artifact from test plans to stimulus to RTL in a single source of truth, spec-to-plan-to-simulation workflows that extract design requirements from specs and generate test plans, and debug-and-fix agents that run regressions, root-cause failures against specs, and propose fixes. Normal EDA achieves state-of-the-art performance on industry benchmarks for RTL design and verification, delivering 2x faster verification on the most complex IP and SoCs, and is deployed on-premises with post-training on customer design data. Normal ASICs compute the way physical systems do, using noise as a resource, compute with memory, and asynchronous operation, with thermo processing-with-memory that runs model heaviest operations inside memory itself for long-context decoding, stochastic analog computation, transformer attention computed in memory for up to approximately 500B parameter models at rack scale, and PCIe accelerator cards in standard air-cooled servers. The team includes co-creators of core TensorFlow frameworks, co-founders of Meta Probability team, architects of Google first production-scale AI deployments, pioneers of NISQ at Los Alamos, and silicon designers from NVIDIA, Apple, and Graphcore. Founded in 2022 and headquartered in New York City, Normal Computing has raised approximately $85 million in Series A funding from investors including Samsung Catalyst Fund, Celesta Capital, First Spark Ventures, Micron Ventures, Galvanize Climate Solutions, Brevan Howard, ArcTern Ventures, and Drive Capital.