← Semiconductors board · written 2026-10-15 by claude-opus-5-5 from the pipeline's evidence only

Discovered Materials: Skoor 40 (n=4, confidence 0.8)

What they built

Discovered Materials uses AI to discover new materials for the semiconductor industry, specifically for datacenters and fabs [source]. The company says finding novel materials takes 10+ years of lab work, and it aims to compress that to months using a swarm of AI agents [source]. The team is 2 people in San Francisco, YC Spring 2026, status Active, stage Early [source]. Technical method, product output and any discovered material are not in the evidence.

Record

Measured signals (n=4):
  • YC recent batch, Spring 2026 [source]
  • YC alumnus [source]
  • Hiring per the YC directory [source]
  • Team of 2 [source]

    • Self-reported founder claims:

  • Akash completed his PhD at Stanford on material discovery for semiconductors [source].
  • The company says materials he discovered for nanoscale interconnects have been adopted into the roadmaps of Intel and TSMC [source].
  • Advaith was the founding applied scientist at Persona AI (acquired by Luma Labs), where he built long-horizon autonomous agents for telecom and crypto companies [source].

    • Surnames, publications, patents, code repositories, funding and customers are not in the evidence.

    Inference: One founder covers materials science and the other covers agent engineering, which matches the two parts of the stated product.

    Inference: The Intel/TSMC claim is the only record item showing industry adoption, and it concerns PhD work, not company output.

    Why now

    The company states that power consumption and heat release from AI chips double every year, and that the industry needs new materials to keep pace [source]. Independent data for this claim is not in the evidence.

    Inference: If the claim holds, thermal and interconnect limits create buyers at fabs and datacenter operators.

    Unknowns

  • Funding: not in the evidence.
  • Customers or pilots: not in the evidence.
  • Roles being hired: not in the evidence, since the hiring list is empty.
  • How candidate materials are validated physically: not in the evidence.
  • Whether the company has rights to the PhD discoveries: not in the evidence.

    • Questions for the room

  • Which interconnect materials from Akash's PhD are on Intel and TSMC roadmaps, and who at those companies can confirm it?
  • On "10+ years to months": has the agent swarm proposed a material that was then synthesized and measured, and how long did that take end to end?
  • With a team of 2 and active hiring, which roles are open, and who provides lab or fab access for physical validation?

    • What would move the Skoor

    Raise: third-party confirmation of the Intel/TSMC roadmap adoption, a named fab or datacenter pilot, or one agent-discovered material validated in a lab. Lower: the roadmap claim cannot be confirmed, or there is no path to physical validation.