CuspAI, a Cambridge-based artificial intelligence startup focused on accelerating materials discovery, has closed a $450 million Series B funding round that values the two-year-old company at $2.6 billion. The deal marks a significant institutional bet on AI-driven solutions to one of manufacturing‘s most consequential bottlenecks: the shortage of advanced materials needed for semiconductor production, energy infrastructure, and industrial innovation.
The funding was led by venture firms Kleiner Perkins and NEA, with a “significant contribution” from Bezos Expeditions, Amazon founder Jeff Bezos‘s personal investment vehicle. The government of the United Kingdom also participated through its sovereign AI fund, underscoring both private and state-level confidence in frontier AI applications that extend beyond consumer-facing models. The round follows a $100 million Series A investment last year and represents one of the year’s largest fintech and deep-tech funding commitments across Europe.
CuspAI’s timing reflects a broader capital shift toward AI infrastructure and scientific computing. The startup has developed MIRA, an AI platform that compresses materials research timelines from years to months by simulating molecular and atomic structures to identify promising compound candidates. The technology combines Meta’s open-source UMA models, which simulate materials at the atomic level, with CuspAI’s proprietary kUPS toolkit for molecular-level analysis and coding automation. The company licensed exclusive AI training rights to academic datasets and published scientific journals to train these models.
Rare Materials and Supply Chain Pressure
The business case for CuspAI rests on a material reality: chipmakers, battery manufacturers, and energy companies depend on rare and expensive elements like iridium and ruthenium that face supply constraints and geopolitical risk. CuspAI aims to discover substitutes or entirely new compounds that perform equivalent functions without relying on scarce materials. The founders, Dr. Chad Edwards and Professor Max Welling, framed the problem starkly: “If we don’t make progress fast, the next 50 years of industrial progress will be constrained by a single challenge: the world needs materials that don’t yet exist.”
To operationalize this mission, CuspAI launched the AI Materials Foundry, a coalition of more than 48 research institutions, chipmakers, and industrial companies including Nvidia, Samsung Electronics, Meta Platforms, and Hyundai. Members can submit requests to MIRA for specific material properties-for example, a chipmaker seeking a new interconnect material-and the platform identifies candidates and routes them to partner laboratories for physical fabrication and testing. The consortium model allows CuspAI to validate outputs against real manufacturing constraints while expanding its customer base across semiconductors, automotive, and clean energy sectors.
European Fintech and AI Capital Are Recovering Momentum
CuspAI’s raise arrives as fintech and AI funding across Europe shows signs of renewed vigor after a period of contraction. Germany’s fintech sector reached EUR 1.2 billion in the first half of 2026, a 196% increase from the second half of 2025 and a 9% year-over-year gain. German fintech became the second-most funded startup category in the country, rising from fifth place six months earlier. Across Europe, Germany captured 26% of all fintech funding in the first half of 2026, trailing only the UK at EUR 1.9 billion.
Other major German fintech rounds in the same period included Cloover, a green fintech platform, which closed EUR 1.4 billion in January through a combination of equity and debt financing to expand clean-energy payment infrastructure across European markets. Upvest, an investment infrastructure provider, raised $125 million in March. These rounds reflect investor appetite for both financial infrastructure solutions and applications addressing energy transition and industrial decarbonization-sectors where materials discovery plays a critical enabling role.
Government Backing and the Sovereign AI Strategy
CuspAI is the fourth company to receive funding from the UK’s sovereign AI fund, a program designed to scale early-stage British AI ventures to global competitiveness. The fund typically makes equity investments between £1 million and £10 million and requires portfolio companies to maintain a primary UK office and have British founders. By backing CuspAI alongside private venture capital, the UK government has signaled that materials science and industrial AI are strategic priorities for national economic resilience.
Liz Kendall, the UK’s science and technology secretary, positioned the investment in terms of solving existential supply-chain problems: “CuspAI is putting the power of AI to work to help us tackle some of the biggest challenges of our time.” The comment reflects a policy shift across advanced economies to view frontier AI not as a consumer entertainment play but as infrastructure enabling manufacturing, energy, and resource constraints.
The company plans to expand internationally with new offices in Singapore and additional hiring across Cambridge, Amsterdam, Berlin, Tokyo, and the United States. This geographic footprint positions CuspAI to serve regional semiconductor and industrial clusters while distributing compute resources and research talent across continents.
What Remains Unproven
CuspAI’s valuation and funding scale rest on the assumption that AI-accelerated discovery can translate laboratory findings into reproducible, economically viable materials at scale. MIRA has demonstrated speed-completing research that normally takes years in six months-but the startup has not yet disclosed how many of its AI-identified compounds have successfully transitioned from simulation to pilot manufacturing or commercial deployment. The AI Materials Foundry consortium adds credibility through multinational partners, but consortium membership does not guarantee adoption of CuspAI’s discoveries into production systems.
Materials science timelines remain stubbornly long even with AI acceleration. Testing, regulatory approval, and manufacturing integration can take years regardless of discovery speed. The real value of MIRA will emerge only when its candidate materials demonstrate superior performance or cost at scale and meet the rigorous qualification standards of semiconductor fabs and energy companies.
For now, CuspAI’s funding round signals that institutional capital-from Bezos, venture firms, and governments-believes the frontier AI model trained on materials data can compress a fundamental industrial bottleneck. Whether that belief translates to chipmakers actually replacing rare earth elements with AI-discovered alternatives remains the unresolved test ahead.
