A startup founded by alumni of the Indian Institute of Technology Madras has raised $9 million to tackle a problem that sits underneath the AI boom but rarely gets talked about: the physical materials that chips are made from.
According to a Times of India report, every new generation of artificial intelligence demands more powerful chips — but there is a problem hidden beneath the algorithms, the data centres and the billion-dollar AI models. The problem is the chips themselves.
The question the company is chasing, as the Times of India frames it, is whether AI can be turned around and pointed at its own foundations: can AI discover the materials of the future?
It is worth being clear about what is and isn't known here. The available reporting establishes the funding amount, the founders' IIT Madras roots, and the broad problem area — chip materials — but does not detail the company's technology, its investors, or its timeline. Those specifics aren't in the source material.
Still, the framing is telling. Most of the money and attention in AI flows to the software layer: the models, the training runs, the applications. Progress at that layer, though, quietly assumes that chipmakers will keep delivering faster, denser, more efficient hardware every couple of years. That assumption rests on materials science — the substances that carry current, dissipate heat and hold circuits together at scales measured in atoms. Discovering new ones has traditionally been slow, expensive laboratory work.
Using AI to search for those materials would close a loop: the technology that is straining today's chips helping to design the chips of tomorrow.
It also places an Indian research institution at an interesting point in the global semiconductor conversation, which has been dominated by manufacturing capacity and export controls rather than the underlying science.
This matters because the pace of AI over the next decade may depend less on bigger models than on whether the physical materials beneath them can keep up.