EdgeCortix, a Japan-based company founded in 2019 by AI veteran Sakyasingha Dasgupta, is revolutionizing energy-efficient AI processors for edge computing. Despite the dominance of Nvidia, EdgeCortix pursues its goal of delivering cloud-level performance at the edge with improved energy efficiency and processing speed, starting with a unique AI-specific processor architecture and a software-first approach.
EdgeCortix’s software-first approach, through its patented hardware and software co-exploration system, addresses the increasing demand for AI applications in a sustainable way. Processing data at the edge, where 75% of the world’s data is generated or processed, is crucial.
Dasgupta, who has extensive AI experience, realized the importance of making computing more energy-efficient to cope with the growing data demands of modern AI models.
Dasgupta’s journey into AI began with developing games as a teenager, eventually leading him to work on early neuromorphic systems and later on AI research at institutions like the Max Planck Society and IBM Research.
His understanding of AI’s evolution and the need for software and hardware integration positioned him to address the stalling of Moore’s Law and the inefficiencies of traditional processors.
EdgeCortix’s approach enables seamless transitions to new hardware without extensive software changes, benefiting customers in aerospace, commercial sectors, and more.
The company’s innovations include the MERA Compiler and Software framework, and the Dynamic Neural Accelerator (DNA), which allow for highly efficient and adaptable AI processing at the edge.
The company progressed to designing its chips, becoming a fabless semiconductor company, and creating the Sakura chip solution. The latest advancement, the Sakura-II AI Accelerator, performs 60 trillion operations per second on just eight watts of power, ideal for real-time AI inferencing in edge applications.
The World Economic Forum recognized EdgeCortix as a Technology Pioneer in 2024, acknowledging its significant contributions to edge AI across various domains. Collaborations with industry leaders and government bodies, like Japan’s Ministry of Defense, highlight the company’s impactful technology, including radiation-tested chips for aerospace applications.
EdgeCortix’s market presence is expanding, with 51% of customers in Japan and 29% in the U.S. The company has seen rapid revenue growth, with revenues doubling in 2024. With under 55 employees, EdgeCortix continues to grow and is preparing for its next round of capital raising, having already secured nearly $40 million in equity and debt.
Dasgupta’s leadership and EdgeCortix’s innovative technologies, like the Sakura-II AI Accelerator, exemplify the company’s commitment to efficient, real-time AI processing at the edge, poised to disrupt markets including defense, aerospace, smart cities, autonomous vehicles, and robotics.
