Orbital Industries, a startup that uses AI to design advanced materials, has raised $50 million in a Series B funding round led by venture capital firm Plural. Nventures, the venture arm of AI chip company Nvidia, which previously invested in Orbital, participated in the round along with other previous investors Radical Ventures, Compound, and Fly Ventures.
The company, which has offices in London and San Francisco, recently rebranded from Orbital Materials. The new capital will expand the commercial rollout of its first two products, both for the data center industry, and expand its team of 50 people, the company said. We also plan to develop our AI platform for industrial applications beyond the data center.
Over the past two years, a number of startups have launched with the hope of using AI models to discover new materials for everything from longer-lasting batteries to biodegradable plastics to non-stick coatings that don’t rely on toxic “permanent chemicals.” Some of Orbital Industries’ competitors include CuspAI, which raised $100 million in a Series A round in September, and Periodic Labs, which raised $300 million in a seed round around the same time.
But Orbital Industries is pursuing a different business model than other startups, said Jonathan Godwin, a former Google DeepMind researcher and Orbital’s co-founder and CEO. luck. Rather than simply using AI models to find new compounds and licensing that intellectual property to large chemical and coating companies, such as Germany’s BASF or Pittsburgh-based PPG, Orbital intends to directly sell the materials discovered by its AI models.
The company is also using AI models to optimize the manufacturing of these compounds, including building new hardware, Godwin said. In this respect, the company is similar to Prometheus, a startup backed by Amazon founder Jeff Bezos and others. Prometheus has raised $6.2 billion and plans to use AI to optimize manufacturing and engineering processes in many high-tech industries, including aviation, defense, and pharmaceuticals.
“We are pushing the frontiers of AI in science and engineering, but we are what you would call vertically integrated,” Godwin said in an interview. luck. “We’re not selling that software. We have the hardware, we have the manufacturing team, we have the advanced materials team, we have the research labs, and we use that software internally to develop new advanced materials and hardware devices, and we sell those hardware devices.”
Orbital’s core AI model is called Orb, and it can predict and simulate the quantum mechanical behavior of atoms. The company says Orb is the only model that can simulate 100,000 atoms on a single GPU, runs about 10 times faster than alternative models, and outperforms models released by Meta and Microsoft. Through a multi-year partnership with Amazon Web Services announced in December 2024, Orbital made a version trained on public data freely available, while maintaining a separate version trained on proprietary data for its own product work.
Orbital used Orb to screen hundreds of thousands of candidates to find a new class of liquid coolant that can be used to alleviate the extreme heat generated by GPU-equipped server racks. This liquid refrigerant is free of PFAS “forever chemicals,” which are increasingly subject to environmental regulations in the United States and Europe.
The challenge, Godwin said, is that operating a modern GPU rack is “basically like putting a supermarket’s worth of energy into a filing cabinet.” The company says the entire family of molecules has now been synthesized and certified by major chip providers.
Orbital also found a contract manufacturer to scale production. Godwin said developing a new coolant traditionally takes “10 years and $100 million.” Orbital completed it in a matter of months at a much lower cost, he said.
The coolant, combined with a refrigeration system that Orbital is also building, is designed to ship with next-generation GPUs in 2027, Godwin said. If this schedule holds, it would be the first time in any industry that an AI-designed molecule would be brought to the commercial market, he said. He points out that even in the field of drug discovery, where startups have been using AI for years to search for new small molecules that can be used as medicines, new drug candidates discovered by AI have yet to make it through clinical trials and onto the market. However, other types of new materials often have a faster path from the lab to the market because they do not require lengthy clinical trials.
The company’s second product is a modular data center system that is manufactured off-site and delivered as a ready-to-deploy unit. Orbital says this allows high-density computing power to be brought online in as little as six months, compared to up to three years for traditional builds. Both products are sold through the commercial brand Orbital IT.
Orbital was co-founded in 2022 by Godwin, who spent five years developing AI for science and advanced materials at Google DeepMind, along with chief technology officer James Zinn Pollock (a regular AI founder who previously sold his company to Shutterstock) and chief operating officer Daniel Miodovnik. Godwin said he left DeepMind because the Google-owned lab itself had no intention of building an advanced materials and hardware company.
Plural partner Ian Hogarth said in a statement that advances in AI are currently constrained by “energy, heat, and infrastructure,” and Orbital is “addressing those constraints directly.” Godwin said part of Plural’s appeal is the company’s involvement in Proxima Fusion, a German fusion energy startup that has raised about $200 million in public and private capital. “There aren’t many funds that have this level of ambition for what we can do here,” he says. “Plural would like to see that happen in Europe and I’m British so I feel passionate about that too,” he added.
Godwin said he ultimately wants to build an AI-native company to rival the “largest industrial conglomerate” in Europe, the century-old chemical giant. “The reason all these companies are old is because they have really deep moats,” he said. “To break through this solid moat, we need a fairly fundamental technological innovation, and that is AI.”
