Normal Computing Raises $50M for AI Chip Design

Normal Computing has successfully closed a substantial $50 million funding round, propelling its mission to revolutionize semiconductor design and address the escalating energy demands of artificial intelligence. The investment, spearheaded by Samsung Catalyst Fund, brings the company's total capital raised to over $85 million. This infusion of capital is earmarked to accelerate the development of Normal's proprietary AI-driven Electronic Design Automation (EDA) platform and its novel physics-based computing architectures.

The strategic financing saw participation from a robust cohort of new backers, including Galvanize, Brevan Howard Macro Venture Fund, and ArcTern Ventures. They join a strong base of existing investors, such as Celesta Capital, Drive Capital, First Spark Ventures (backed by Eric Schmidt), and Micron Ventures, underscoring significant confidence in Normal's disruptive potential. The company is tackling a critical juncture where the complexity of chip design is outpacing traditional methods, while AI's insatiable appetite for power threatens to create an energy bottleneck for data centers, projected by some to occur around 2030.

Normal Computing's core innovation lies in its AI-native approach to silicon design. Its flagship product, Normal EDA, is engineered to dramatically shorten the time-to-market for custom silicon, promising to cut design cycles by up to half. Beyond accelerating current design workflows, the platform is designed to enable future generations of chips that could achieve orders of magnitude improvements in computational efficiency, a crucial factor for scaling AI capabilities sustainably. The company reports partnerships with over half of the top ten semiconductor firms by revenue, indicating strong industry validation for its production-ready solutions.

Faris Sbahi, CEO of Normal Computing, emphasized the urgency of the situation, stating, "Meeting the growing demands for 'intelligence per dollar per Watt' requires a fundamentally novel architecture and a unified design methodology." He further elaborated on the company's vision: "Normal EDA exists to accelerate custom silicon to market by 2x today, and, over time, to enable 1000x gains in efficiency with our platform." This ambition is supported by the development of its Carnot hardware program, which leverages Normal EDA to create its own silicon IP. Notably, Normal successfully taped out CN101, the first thermodynamic computing chip, in August 2025, targeting enhanced performance for generative AI models.

The semiconductor industry, while a powerhouse of innovation, has seen its design tools evolve minimally compared to other knowledge work sectors. This has led to challenges such as talent scarcity, costly design iterations, and the pressure to meet the performance and energy benchmarks set by AI scaling laws. Normal's platform aims to bridge this gap by integrating advanced AI techniques, including auto-formalization, to assist human engineers in solving complex design challenges. This collaborative AI-human approach promises to compress design timelines from years to mere months, without compromising the rigorous standards of hardware development.

Dede Goldschmidt, SVP & Managing Director at Samsung Catalyst Fund, highlighted the transformative impact of AI on the semiconductor sector and Normal's unique position: "Normal has assembled a strong team with broad experience across AI and semiconductors. The platform they have built has the potential to offer faster time-to-market for the most demanding chip design customers in the ecosystem." The company's commitment extends to industry collaboration, evidenced by its founding membership in the Silicon Integration Initiative (Si2) LLM Benchmarking Coalition, advocating for open EDA standards.