A German startup plans to jumpstart European EVs with an AI-powered brain.
Sphere Energy built the system to simulate battery behaviour. The company then predicts a power source’s lifetime in numerous scenarios, from driving styles to temperatures on the road.
According to Sphere, the insights shrink the battery testing cycle by at least a year. Developing a car, meanwhile, could be completed “at least” twice as quickly.
Sphere envisions endless benefits: manufacturers will save millions, car prices will plummet, and innovations will increase at exponential rates.
The startup’s co-founder, Lukas Lutz, said the plans are unprecedented.
“Nobody right now — not even Tesla — can accurately estimate the lifetime of their battery,” Lutz told TNW. “This is something that will be really groundbreaking.”
A lifeline for European EVs?
Sphere unveiled the project last month at the IBM Research Lab in Switzerland.
In a futuristic facility overlooking Lake Zurich, the startup introduced an AI brain called Batty.
Batty was initially trained on years of testing data from over 1,000 batteries. Car manufacturers also mix in their own information. The system then simulates a specific battery’s life under various conditions.
Customers can test the effects of speeding down motorways and crawling around mountains, applying fast and slow chargers, driving in searing summers and freezing winters. Every aspect will impact the battery’s degradation.
The system’s power derives from the transformer architecture — the founding stone of today’s large language models (LLMs). But Sphere’s approach doesn’t rely solely on text. The startup extends the model’s scope by integrating time-series data. As a result, the system can simulate a battery’s behaviour over years.
The approach adds a new twist to the LLM paradigm. While a chatbot predicts the next best word, Batty will predict the next best data point.
Car companies have been impressed by the results. According to Sphere, the majority of European manufacturers have already used the tech.
Batty could provide a vital boost to the continent’s EV makers, which are rapidly losing market share to their Chinese rivals.
“Battery development is a huge pain for them — and it shouldn’t be,” Lutz said. “We really want to take away the burden.”
But batteries are just the start of Sphere’s ambitions. The company envisions simulating endless energy applications, from electric boats to grid storage.
Alongside IBM, the startup is also exploring new levels of simulating batteries.
“With these foundation AI models, we understand atomic level behaviour intrinsically,” Lutz said. “But we want to go sub-atomic — with quantum.”