IIT Madras alumnus Akash Ramdas raises $9 million for semiconductor AI startup

Discovered Materials, an artificial intelligence startup co-founded by Indian Institute of Technology Madras graduate Akash Ramdas, has secured $9 million in seed funding to speed up the discovery of semiconductor materials that prevent advanced microchips from overheating.

The venture applies AI models, including systems built by Anthropic, to generate candidate compounds capable of handling heat in high-power computing hardware. These potential solutions are then tested through physics-based computational modeling and simulations to pinpoint the most viable options for laboratory production.

Research pace and platform performance

During his doctoral research, Ramdas could evaluate about 20 potential material candidates each day using computational methods. The AI agents developed by Discovered Materials now screen thousands of possibilities in the same timeframe, reducing a process that traditionally took years down to days.

Alongside the funding, the company has publicly released hundreds of AI-discovered materials and established a standard testing benchmark to measure how effectively AI models evaluate potential semiconductor compounds.

From Chennai to Silicon Valley

Ramdas completed his undergraduate degree in Metallurgical and Materials Engineering at IIT Madras in 2019. He moved to Stanford University for his master's degree in Materials Science and Engineering, which he finished in 2021, before earning his PhD in Materials Engineering in June 2025. His doctoral work focused on computational tools to find materials for nanoscale electronic devices.

After working as a postdoctoral researcher at Stanford until April 2026, he co-founded Discovered Materials in March 2026 to turn academic research into commercially available technology. The venture underscores the strength of premier Indian institutions in training world-class engineers who go on to lead global technological innovation.

Physical testing remain the central hurdle

While AI dramatically accelerates candidate selection, turning computer models into functional hardware still requires physical synthesis in laboratories. Researchers must manufacture and test these compounds in physical facilities to prove that they perform reliably and can be produced at scale.

AI-discovered semiconductor materials have not yet achieved widespread commercial use across the technology industry, as physical testing and manufacturing integration remain major operational bottlenecks.