Source: Zhitong Finance
Well-known startup Chemix said its artificial intelligence platform can speed up the electric vehicle battery development and manufacturing process to ensure that electric vehicles are more efficient and run longer.
The next revolutionary target of artificial intelligence technology will most likely be in the field of electric vehicle battery research and development.
Recently, Chemix, an electric vehicle battery developer driven by artificial intelligence technology, completed Series A financing of up to 20 million US dollars, led by Ibex Investors. The AI startup, which has only been in business for three years, will use this funding to expand the scale of its technology, which uses artificial intelligence algorithms to design and optimize the development and manufacture of new batteries for electric vehicle manufacturers.
As global business interest in artificial intelligence technology surges, the energy intensity of the technology and its corresponding capacity burden on data centers and power grid systems are also increasing. But in addition to these challenges, artificial intelligence also has the opportunity to help accelerate the decarbonization process by speeding up the development process of sustainable materials and the electric vehicle battery development cycle.
The development of electric vehicle batteries is usually a slow and resource-intensive process requiring extensive testing and retesting. For example, measuring the usage time of an electric vehicle battery may require thousands of charges and discharges, which is time consuming and costly. Using artificial intelligence, developers can reduce battery development design time and development costs through faster and more efficient technology iterations.
According to a 2020 “Nature” study, the booster effect of an early prediction model reduced the time required to fully certify fast charging technology from more than 500 days to 16 days, partly due to using valid data from previous R&D cycles to predict the final cycle life.
Alp Kucukelbir, co-founder and chief scientist of artificial intelligence company Fero Labs, said: “Generally speaking, R&D and testing costs in the pharmaceutical industry are very expensive, and this approach has been used to save costs for years.” Kucukelbir recently wrote a report on the potential of artificial intelligence technology to facilitate climate solutions. He said there are now financial and social incentives to apply the same technology to the battery and electric vehicle industries.
Chemix, based in Sunnyvale, California, said the technology company's artificial intelligence algorithms can help customers make customized electric vehicle batteries with specific properties, such as being able to withstand high temperatures or eliminate certain materials. Chemix designs batteries and collaborates with manufacturing partners to produce batteries, which are then sold at a premium to select customers.
According to Kaixiang Lin, the company's co-founder and CEO, in a typical case, Chemix helped a customer develop a novel electrolyte — a substance that transfers ions between the battery cathode and anode, which can extend the battery cycle life by 400% within two months. Without artificial intelligence, this process usually takes almost twice as long and is expensive.
Kaixiang Lin declined to reveal Chemix's current customer base, but he said it includes high-end electric vehicle manufacturers that want to surpass non-specially designed batteries. Currently, the startup is focusing on electrolyte design work for electric vehicles, but it plans to switch to cathodes, anodes, and slurries — all of which are critical battery components, as well as household batteries.
The electric vehicle market is growing rapidly, and this is a challenge for battery manufacturers, who want to keep up with the pace of development.
“Chemix is not inventing new chemicals; it is trying to make existing chemicals better,” said Yann Lagalaye, managing partner of BNP Paribas Solar Impulse Venture Fund (BNP Paribas Solar Impulse Venture Fund), which participated in this round of financing. “Simply improving existing (chemical work) is huge, because by doing so, they can solve underserved markets”. These markets currently rely on popular market batteries, and these batteries are not customized to their needs, such as electric sports cars and motorcycles.
Lin said Chemix is developing a proprietary data set collected in the company's physics lab to help train big models of artificial intelligence.
“To actually apply artificial intelligence and unlock the full potential of artificial intelligence technology, what the industry lacks is the availability of data,” he said. “If you look at GenAI, and apps like ChatGPT, they can sift through large amounts of data from the internet, but when it comes to electric vehicle battery or hardware development, there's often not enough data.”
Established automotive giant GM has bet on “AI+ electric vehicle battery development”
In the US, “AI+ electric vehicle battery development” has become the most important emerging technology in the field of vehicle construction. This technology uses AI to drastically reduce the time and cost of developing electric vehicle batteries. A traditional giant in automobile manufacturing$General Motors (GM.US)$Mitra Chem led Mitra Chem's $60 million Series B funding project in the second half of last year. Mitra Chem was described by some big players in the battery technology industry as a battery materials innovator driven by artificial intelligence technology in Silicon Valley. Mitra Chem's artificial intelligence platform and advanced R&D equipment in Mountain View, California are expected to help greatly accelerate the commercialization process of GM's various affordable batteries for electric vehicles.
Under the partnership agreement, General Motors (General Motors) and Mitra Chem will develop advanced iron-based cathode active materials, such as lithium ferromanganese phosphate, to drive affordable electric vehicle batteries compatible with GM's Ultium electric vehicle architecture platform. GM's financial support is designed to help Mitra Chem expand its current business scale and accelerate the full introduction of its novel battery material formulations to market.
Mitra Chem is a new startup with few public information. According to some sources, Mitra Chem's original battery R&D facility is famous for being able to simulate, synthesize, and test thousands of cathode design solutions every month, ranging in size from grams to kilograms. These AI-driven processes have greatly shortened the development cycle and the time required to launch new battery formulations.
The “atom-to-ton acceleration platform” provides the core power for Mitra Chem's laboratory, using simulation and physical-information machine learning models to accelerate recipe discovery, cathode synthesis optimization, battery life evaluation, and process amplification. The “internal cloud platform” is specifically built for battery cathode development, automating data ingestion across various syntheses, material characterization, battery prototyping, standardized analysis, and visualization.