Nissan partnered with Microsoft to create a model predicting EV battery performance over time, improving it by about 80%.

Judge as of 19 November 2024. True: Nissan and Microsoft collaborated on a model, for example a machine-learning or AI model on Azure, that predicts EV battery performance or degradation over time, and a reported improvement, such as in prediction accuracy or speed, was about 80% (roughly 70–90%, given the 'something like' hedge). Mostly true: the partnership and model exist, but the improvement figure is somewhat outside that range or measures a slightly different metric. Mostly false: there was a partnership, but the model concerned something else, or the improvement was far from 80%. False: no such Nissan–Microsoft collaboration existed.

Set before the research began.

1. E1 AI model boosts EV battery prediction accuracy Microsoft Research (Microsoft Research blog) · 16 July 2024 Own record

   “Compared with popular state-of-the-art battery prediction methods, this data-driven model improves accuracy by approximately 80% with Nissan’s simulation data and by over 30% with real-world experimental data.”

2. E2 AI model boosts EV battery prediction accuracy Microsoft Research (Microsoft Research blog) · 16 July 2024 Own record

   “To address this, Microsoft Research collaborated with Nissan Motor Corporation to develop a new machine learning method that predicts battery degradation with an average error rate of just 0.94%, significantly bolstering Nissan’s battery recycling efforts.”

3. E3 Full Keynote: Satya Nadella at Microsoft Ignite 2024 - Summary videohighlight.com · 19 November 2024

   “Nissan has collaborated to develop a model that predicts electric vehicle (EV) battery performance over time, achieving an improvement of approximately 80%.”

A Microsoft Research blog post from July 2024 says Microsoft Research worked with Nissan Motor Corporation on a machine-learning method that predicts EV battery degradation (E2). It reports that the model improved accuracy by about 80% over state-of-the-art methods on Nissan's simulation data, and by over 30% on real-world experimental data (E1). That meets the criteria: there was a partnership, the model predicts battery degradation, and a reported accuracy gain is about 80%. One caveat is that the 80% figure applies only to simulation data, and the only source for it is Microsoft's own published research record; E3 just summarises the speech, so it does not independently confirm the figure.