Swiss startupJuaRecently announced that it has successfully raised $16 million to build a natural foundation artificial intelligence modelArtificial Intelligence ModelsThe first application area will be weather forecasting. This move indicates that large-scale AI models are gradually becoming the core of generating language, vision and audio data for AI services, which has an important impact on the development of AI, similar to the role of operating systems in the development of smartphones.
Source Note: The image is generated by AI.
Jua CEO Andreas Brunner said that climate change and geopolitical instability have led to a demand for more accurate modeling and forecasting for organizations working in industrial fields such as energy or agriculture. Jua plans to meet this demand by building large-scale "physical" models, thus opening up new prospects for the application of artificial intelligence in the physical world.
Although there are already some similar projects, such as Google's GraphCast, Nvidia's FourCastNet, and Huawei's Pangu, Jua claims that its model is superior and has a larger data scale, 20 times that of GraphCast. Additionally, Jua emphasizes that weather prediction is only the starting point of its model, and it will involve a wider range of physical problems and challenges in the future.
While addressing current climate fluctuations, Jua aims to build the first natural foundation modelFirstlyThe company has not yet launched its first product that lays the foundation for artificial general intelligence by learning physics.the firstbut the funding from investors suggests that their confidence in Jua is more than just a fervent pursuit of the artificial intelligence craze.
Marvin Gabler, one of Jua’s co-founders and chief technology officer, previously worked as head of research at Q.met, a weather forecasting company, and worked on deep learning technology for the German government. This shows that Jua’s team has a strong technical background and a keen understanding of industry problems and potential solutions.Firstlyunderstanding of industry problems and potential solutions
Jua uses a new approach to build forecast models, not only using traditional weather station data, but also integrating more complex and real-time data such as satellite images and terrain. The company claims that its system has a lower computing cost compared to traditional systems and uses only one ten-thousandth of the computing resources of traditional systems.
It is worth noting that Jua's financing comes at a time when basic models are becoming the cornerstone of the development of next-generation artificial intelligence applications. Companies that master and control basic models will have huge value and potential power. In this field, key global companies such as OpenAI, Google, and Microsoft are actively laying the groundwork, and Jua is one of them, and it is also attracting much attention in Europe.
Jua aims to provide accurate insights to multiple industries, including insurance, chemical and energy providers, disaster planning teams, agriculture, airlines, and relief charities, through more comprehensive and larger data models. This effort is not only to address the challenges of climate change today, but also to address broader problems in the physical world in the future. However, similar to other artificial intelligence models, Jua also needs to address challenges in security and reliability. Although these issues are currently in the basic stage, Jua has already begun to consider them.