Aizip, Inc., in partnership with SoftBank Corp., has announced the successful demonstration of AI solutions tailored for fish-farming applications, earning them the CES® 2025 Innovation Award for their breakthrough technology. This collaboration reflects Aizip’s mission to bring high-performance machine-learning solutions to global industries, particularly in environments where traditional cloud-based AI systems face major limitations. The partnership has resulted in a successful demonstration of AI-powered fish counting and fish-net inspection for aquaculture.
While AI has been widely adopted across many industries, its use in aquaculture has been slower to take off. Monitoring fish populations and detecting net damage are essential tasks for efficient fish farming, traditionally performed by divers, which incurs substantial costs and risks. The aquaculture industry faces unique challenges for digital transformation using AI, including the need for frequent data collection and real-time analysis to monitor fish populations and detect damage. However, ocean environments often lack the reliable power sources and network connectivity required for traditional cloud-based AI systems to function or transmit large volumes of real-time visual data.
In their collaboration, SoftBank’s computer graphics (CG) simulations of fish schooling behavior automatically generated training data, while Aizip provided its compact AI technology. Together, they developed a highly efficient and accurate AI model optimized for real-time edge applications. This partnership reflects SoftBank’s ongoing effort to apply advanced technologies for societal benefit and further cements Aizip as a leader in the TinyML market.
The resulting AI model, which can run on compact devices like smartphones, is now available to fish farmers. The model achieved an impressive 95% recognition rate in underwater field tests, showcasing its commercial potential. The software suite, called Watatumi, received the CES® 2025 Innovation Award in the Food & AgTech category, underscoring the power of edge AI to provide server-level computational accuracy in challenging environments, enabling real-time analysis without relying on external power sources or network connectivity.
“This demonstration represents a significant milestone in applying AI to aquaculture, enhancing operational efficiency and reducing risks for farmers,” said Yuko Ishiwaka, Principal Investigator at SoftBank’s Advanced Technology Promotion Office. “Our partnership with Aizip has been highly productive, and we look forward to expanding AI technology into more applications,” added Kazuto Suda, Director of the same office. “This is just the start of AI’s potential in aquaculture and other agricultural sectors,” remarked Yen-Chi Xu, an AI scientist at Aizip who participated in the model design and field tests. “We’re excited to continue collaborating with SoftBank to deliver efficient on-device AI solutions to our customers.”
Founded in 2020, Aizip has pioneered efficient, robust AI solutions for a global clientele. Its diverse portfolio includes vision, audio, time-series, and language models for both consumer and enterprise applications. The aquaculture solution is part of Aizip’s broader industrial AI offerings for enterprise customers. As industries increasingly recognize the value of locally processed data and the operational advantages of edge computing, Aizip is poised to deliver versatile, production-ready solutions across sectors that demand real-time, efficient analysis in constrained environments, enabling businesses to leverage AI where it’s most needed.
Aizip remains committed to providing powerful, low-cost AI solutions on edge devices, aiming to benefit society through cutting-edge technology.
