Lightspeed leads $9 million seed in Discovered Materials
The deep tech startup is building thermal materials for AI chip packaging. Y Combinator, Peak XV and Paul Graham joined the round.
- Discovered Materials raised $9 million, about Rs 85 crore, in a seed round led by Lightspeed India Partners.
- Y Combinator, Peak XV Partners and angels including Paul Graham, Gokul Rajaram and Thariq Shihipar also participated.
- The startup targets thermal dissipation in AI chips that generate over 140W per square centimetre.
- The company says its AI systems developed new thermal materials in three months, matching products that took years to build.
Discovered Materials has raised $9 million, about Rs 85 crore, in a seed round led by Lightspeed India Partners. Y Combinator and Peak XV Partners also took part, along with angels including Paul Graham, Gokul Rajaram and Thariq Shihipar.
The company was founded by Advaith Sridhar and Akash Ramdas. Its problem is heat. Modern AI silicon can generate more than 140W per square centimetre, and moving that heat out of a 3D stacked package is a hard materials question. Discovered Materials is building thermally conductive dielectric materials for 3D chip packaging.
The method is the pitch. The company runs cloud based autonomous AI agents that test thousands of virtual material hypotheses a day. Custom model harnesses built around frontier AI models generate candidate materials. Physics simulations then screen those candidates for stability, dielectric constants and thermal properties. The company says its AI systems produced new thermal materials in three months with performance comparable to products that took years to develop. The funding goes to expanding the team and laboratory, and to scaling the research agents.
A $9 million seed is large by Indian standards, and the shape of the syndicate says more than the size. Lightspeed India leading, with Y Combinator and Peak XV alongside and Paul Graham writing a cheque, is a round assembled across two markets at once. That structure has become ordinary for Indian deep tech at seed stage. The read for founders is that capital is available for pre revenue science, provided the team and the problem are legible to a global buyer rather than only a domestic one.
The caution is timeline. Materials companies do not ship on software schedules. Three months to a promising candidate in simulation is not three months to a qualified part sitting in a customer package. Semiconductor qualification runs in years, and the buyer list is short and conservative. So the honest read on this round is narrow but real. Indian teams can now raise seed money sized for long research cycles. Whether the follow on capital turns up at Series A and B, when the lab bills are larger and revenue is still distant, remains the open question.
Zane’s analysis draws on original reporting by Entrackr. Read the original report.