CuspAI Launches the AI Materials Foundry to Accelerate Next-Gen Discoveries

In July 2026, Cambridge-based CuspAI launched the AI Materials Foundry, a global network of over 45 organizations collaborating to escalate the discovery of novel materials. The platform aims to advance breakthroughs in sectors such as semiconductors and clean energy, backed by partners including NVIDIA for compute infrastructure and Meta for its Universal Model for Atoms (UMA).

Data-Driven Certainty for AI Predictions

The AI Materials Foundry emphasized discovering novel semiconductors and other materials useful in clean energy and advanced manufacturing. The four key attributes for software-led materials discovery are:

• To make AI predictions reliable with high-quality training data at scale
• To screen at molecular resolution across billions of candidates with computing power
• To move from digital design to physical reality with synthesis infrastructure
• To interpret what the machine finds and know what to do with it with domain expertise

Dr. Chad Edwards, founder and CEO of CuspAI, remarked that the next 50 years of industrial progress could be hindered without new materials to advanced materials discovery. Over 45 leaders in their fields are participating, including companies like Caelux, Oxford PV, Mitsui Chemicals, Applied Materials, 3M, and Fujifilm, all linked to the solar industry.

Oxford PV's CTO, Ed Crossland, emphasized that AI-driven materials discovery can escalate the shift from scientific insight to commercial impact. Caelux CEO Scott Graybeal expressed pride in being a founding member of CuspAI’s AI Materials Foundry, reinforcing their strong focus on improving perovskite solar technology to improve efficiency.

From Algorithms to Atoms: How CuspAI Works

• Dataset and AI Orchestration: CuspAI uses its MIRA AI platform along with massive curated materials datasets to autonomously generate and screen material candidates based on specific attributes shared by partners.
• Private Research and Development Integration: The platform operates as a private as well as deployable instance within a company's existing research and development workflow.
• Atomic Simulation: The generated candidates' physical properties are simulated using kUPS, an open-source toolkit developed with NVIDIA ALCHEMI that utilizes Meta's UMA to simulate atomic interactions.
• Fast Virtual Screening: The system can evaluate hundreds of trillions of molecular structures in a fraction of the time, allowing companies like Kemira to condense years of material screening into just months.

How AI Smooths Renewable Power Integration

AI systems like Google’s GNoME and Microsoft’s MatterGen drastically accelerate the discovery of stable, novel compounds. These models have predicted hundreds of thousands of new structures for developing advanced clean energy technologies like highly efficient lithium-ion batteries to improve battery performance.

Machine learning aids in manufacturing and testing stable perovskite materials for durable solar cells. Beyond material science, AI algorithms analyze the physical properties of deployed solar installations to optimize operations and maintenance practices efficiently.

Impact on the ICT Industry

This launch has a great impact on the ICT industry by mitigating the physical limits of hardware scaling. By compressing materials research and development from decades to months, the initiative directly tackles the foundational materials bottleneck in semiconductors, advanced electronics, and energy storage.

Additionally, it actively near-shores the supply chain, recognizing alternative materials that reduce the demand for rare metals by shielding the ICT industry from critical geopolitical supply-chain vulnerabilities. Instead of isolated, siloed laboratories, the foundry ecosystem creates a shared data advantage, thereby accelerating technological timelines.

As traditional trial-and-error chemistry is constrained by time, CuspAI accelerates materials discovery from years to months, together with high-quality data from the CCDC and ICSD, with its MIRA platform and Meta's UMA to virtually screen trillions of molecules. The technology pairs generative AI with physical synthesis infrastructure to ensure discovered molecular designs are manufacturable.

Impact of the Solar AI Industry

The global solar AI market is surging, with an overall revenue growth expectation of hundreds of millions of dollars during the forecast period from 2026 to 2035. 

According to Precedence Research, this launch accelerates the discovery of new materials by combining computing power and lab access. This software-led approach bypasses traditional research and development challenges across semiconductors and clean energy.

The AI Materials Foundry accelerates materials research and development from decades to months by CuspAI and screening 300 trillion molecular structures in six months. To overcome industry limitations, the Foundry solves data scarcity and computing limitations by leveraging exclusive experimental records and advanced technology to screen billions of molecular candidates. It directly links AI to physical labs and incorporates deep domain expertise to ensure manufacturing feasibility and improve interpretability.

About CuspAI

CuspAI is a UK-based, AI-driven materials discovery startup founded in 2024 by Dr. Chad Edwards and Prof. Max Welling. The company leverages generative AI and deep molecular simulations to design novel materials with specific chemical and physical properties on demand by lowering traditional research and development timelines.

Their main offerings include the MIRA Platform and kUPS. These also accelerate industrial materials discovery by combining its MIRA AI platform with a global AI Materials Foundry network through collaborative co-development and pooled infrastructure for carbon capture and energy-efficient semiconductors.