Sandbox AQ, a Google spinout, is using AI-driven research to accelerate the development of safer, more affordable solid-state batteries, aiming to overcome America’s reliance on China for EV battery manufacturing. By rapidly screening materials and focusing on innovative electrolytes, the company seeks to shorten battery R&D timelines and strengthen the US position in the growing global battery market.
Sandbox AQ, a Google spinout founded in 2022, is leveraging artificial intelligence to address America’s electric vehicle (EV) battery challenges amid China’s dominance in battery manufacturing. China has gained a competitive edge by producing large volumes of batteries quickly and cheaply, enabling automakers like BYD and Geely to offer EVs at significantly lower prices than US and European models. Instead of competing on volume, Sandbox AQ aims to innovate better battery designs using its AI-powered research platform to accelerate the development of safer, more affordable solid-state batteries for EVs, military equipment, and data centers.
The company’s AI platform, AQ Vault 26, focuses on compressing the early, uncertain stages of battery research and development by rapidly screening and evaluating candidate materials. This approach allows scientists to discard ineffective ideas quickly and concentrate on promising ones, potentially reducing the discovery phase of battery R&D by 90 to 95%. While the entire battery development process currently takes 10 to 15 years, Sandbox AQ’s technology aims to significantly shorten this timeline, accelerating the overall pipeline for new battery chemistries.
Sandbox AQ has attracted $950 million in funding from prominent backers such as Alphabet, Nvidia, and AI expert Yann LeCun. The company is already generating revenue through partnerships with clients like battery developer Novonix and the US Army, as well as other unnamed battery and automotive firms. Its business model includes charging fees for access to its research platform, licensing its technology, conducting research on behalf of other companies, and developing proprietary battery materials. With the battery market projected to grow to $500 billion this decade and potentially reach $1 trillion, Sandbox AQ is targeting the high-value segment of materials discovery and performance optimization.
Unlike AI models focused on language generation, Sandbox AQ employs large quantitative models (LQMs) based on physics and scientific data to generate synthetic materials data. The company is developing solid-state batteries that ideally avoid using costly and potentially hazardous materials like lithium and cobalt. Instead, it is exploring halide-based electrolytes—minerals found in rock salt, fluorine, and iodine—as a safer, more stable, and cost-effective alternative to traditional lithium salts, which can overheat and cause fires in EV batteries.
Although commercial products using Sandbox AQ’s technology are still at least five years away, the innovation could bolster the US position in the global battery market, which is currently dominated by China. By advancing battery materials discovery and creating more resilient supply chains, Sandbox AQ hopes to facilitate wider EV adoption and strengthen America’s competitiveness in this critical industry. The company’s efforts reflect a broader trend of applying AI to physical sciences and materials research, aiming to transform energy storage and related sectors.