NTU Singapore Develops AI to Forecast Food Spoilage, Targeting Waste Reduction

Nanyang Technological University, Singapore (NTU Singapore) has engineered a sophisticated artificial intelligence model designed to forecast the progression of bacterial growth within various food products. This development is intended to provide retailers in Singapore with precise, data-driven guidance for shelf life determinations and inventory management, with the objective of substantially mitigating the nation's food waste footprint. The research initiative forms a key component of NTU's broader efforts in food security, aligning with national priorities to enhance local resilience in the food supply chain, particularly as Singapore currently imports 90 per cent of its food supply.

The core mechanism of this AI involves rigorous training based on observing microbial behavior under a spectrum of controlled food conditions, including variations in temperature and acidity levels. Researchers at the Future Ready Food Safety Hub (FRESH@NTU), a joint venture with the Agency for Science, Technology and Research (A*STAR) and the Singapore Food Agency (SFA), are meticulously tracking pathogens like Salmonella to train the model for greater real-world accuracy. This predictive framework allows for the estimation of contamination levels of harmful pathogens under actual storage conditions across the distribution chain.

Beyond waste reduction, the implementation of this predictive modeling carries significant operational and economic implications for the industry. Optimized refrigeration use, guided by the AI's forecasts, presents an opportunity to realize considerable energy cost savings without compromising established food quality or safety standards. Professor William Chen, Executive Director of FRESH@NTU, noted that adjusting refrigeration temperatures, even by a single degree, can yield substantial reductions in electricity expenditures for food businesses.

This technology is transitioning from the laboratory to commercial application, with concrete plans for real-world validation. Supermarket chain Sheng Siong has expressed interest in collaborating and is scheduled to commence a trial of the technology during the latter half of 2026. Sheng Siong has previously refocused efforts on food waste reduction and diversion in 2025, and this trial represents a further step in digitalizing their operations.

The research at FRESH@NTU is designed to offer food businesses a more granular method for monitoring freshness, thereby supporting critical day-to-day decisions regarding handling, storage, and stock turnover. The alignment between the AI's projections and actual bacterial activity validates its potential to complement existing food safety guidelines by proactively managing spoilage risks. This integrated approach aims to enhance food security by minimizing waste across the entire supply chain while simultaneously lowering energy consumption.

This NTU innovation reflects a wider trend in the agri-food sector where artificial intelligence is being leveraged to modernize perishable goods supply chains globally. For instance, a separate 16-month pilot in the United Kingdom, backed by an Innovate UK grant, utilized AI technology to map food waste on production lines in real time, successfully leading to the redistribution of the equivalent of nearly 500,000 meals to charities. Such international efforts underscore the potential for AI to improve the speed, accuracy, and predictability of food waste analysis compared to conventional manual assessments.