(18 June): CuspAI, a British startup developing algorithms to design new materials, has raised US$30 million (RM141.45 million) in its first funding round and welcomed artificial intelligence pioneer Geoffrey Hinton to its advisory board.
Cusp AI said in a statement Tuesday that the company is using generative AI and molecular simulations as a materials “search engine.”Hinton, a computer scientist often called one of the “godfathers of AI,” has been handpicking his advisers since leaving Alphabet Inc.’s Google last year.
Cambridge-based CuspAI is one of several startups aiming to fight climate change with AI-designed materials. The company will initially focus on developing software that can improve the porous structures used in carbon capture and storage, a market that has faced setbacks due to cost concerns.
Chatbots are emblematic of the generative AI boom, but scientific labs are also harnessing the emerging technology: Startups and global corporations, from Google to Toyota, are developing algorithms that predict how molecules will behave to make things easier, like inventing new medicines or car batteries.
CuspAI's seed round was led by London-based Hoxton Ventures, with participation from Lightspeed Venture Partners and others.
CuspAI sees opportunities in areas such as green hydrogen, synthetic fuels and semiconductor manufacturing, according to co-founder and CEO Chad Edwards.
The company is developing its own database for its AI through a partnership with Meta Platforms, which Edwards said is also providing the computing resources.
The company's other co-founder is Dutch AI scientist Max Welling, who previously ran a Microsoft lab in Amsterdam. Welling said CuspAI's long-term goal is to create “highly capable” AI-based models customized to generate new materials, like those being built by OpenAI and Google.
AI software can drastically reduce the cost and time of materials development, but it has yet to prove it can solve the complexities of working with molecules in the lab, says Alex Ganouz, a chemistry lecturer at Imperial College London. “It's easy to predict materials,” he says. “It's incredibly hard to move from prediction to real, practical application.”
Welling says recent research advances showing how AI can generate specific crystals and polymers are evidence of progress in the field. “I'm very confident this will be successful,” he says. “Of course, finding a market is always a long journey.”
