Amphitrite, a French startup powered by NVIDIA technology, helps ships reduce travel time, fuel use, and emissions by predicting ocean currents and simulating different weather conditions.
The Amphitrite startup, a participant in the NVIDIA Inception acceleration program, uses AI and satellite data to optimise maritime operations with NVIDIA Earth-2 – a simulation initiative focused on creating a virtual model of the Earth and understanding complex processes like climate change, weather patterns, and ocean currents.
The tool leverages AI, physics, and high-performance computing to simulate and predict global phenomena and weather conditions. Amphitrite AI models, powered by NVIDIA technology, help ships reduce travel time, use of energy resources, and carbon emissions. The company also aids in ocean cleanup by tracking plastic waste.
Its AI system uses real-time satellite data, improving accuracy over traditional methods. The startup offers ship managers insights on the best times to sail and the most efficient routes based on ocean current data. This way, vessels can harness the natural power of currents to reduce fuel use, improve speed indicators, and cut emissions, leading to more efficient and sustainable marine journeys.
At present, the global shipping industry significantly contributes to greenhouse gas emissions. Moreover, its carbon footprint is growing despite the worldwide push for decarbonisation, roughly doubling since 1990. Thus, in 2022, international shipping accounted for nearly 3% of the world’s greenhouse gas emissions. The sector’s CO₂ emissions increased by 1.1% in 2023, reaching 706 million metric tons.
The shipping emissions primarily come from the burning of fossil fuels in ships’ engines. The industry is under increasing pressure to decarbonise and reduce its carbon footprint. While some companies are starting to leverage innovative zero-carbon cargo ships using methanol as a power source, the power of AI helps reduce emissions for all types of vessels with smart routing and weather predictions.
Amphitrite’s clients include shipping companies and environmental organisations, working towards sustainability and efficient shipping practices. Its advanced routing solutions have helped shipping companies reduce carbon emissions by up to 10% by optimising travel and avoiding adverse weather conditions.
“We’re at a turning point on the modernization of oceanic atmospheric forecasting. There’s a wide portfolio of applications that can use these domain-specific oceanographic AI models — first and foremost, we’re using them to help foster the energy transition and alleviate environmental issues.”
Alexandre Stegner, co-founder and CEO of Amphitrite
Among Amphitrite’s main innovations, there are domain-specific three-kilometre-scale AI models analysing ocean variables one at a time in a more precise way than traditional global models. The startup’s satellite data comes from reputable sources, including NASA, the European Space Agency, and the SWOT satellite mission (a collaboration between NASA, CNES, the Canadian Space Agency, and the UK Space Agency).
As for the tech component, the French startup relies on NVIDIA H100 GPUs for model training and cloud computing, and Earth-2 and FourCastNet solutions for advanced simulations. By combining AI with real-time satellite imagery, Amphitrite significantly improves the accuracy of ocean forecasting, handling complex oceanic data, especially in regions with limited observational data.
