Rust Belt Industrials
Systems & Software / Materials Discovery AI

matforge

San Francisco, CAStartupdiscoveredmaterials.com
Total Funding
~$500K (YC pre-seed, estimated)
Founded
2026
Employees
2
Headquarters
San Francisco, CA 94105

Discovered Materials (operating as Matforge) is a San Francisco-based AI startup building autonomous "AI scientists" that dramatically accelerate the discovery of new materials for the semiconductor industry. The company's core thesis is that the semiconductor industry's continued progress - Moore's Law - is increasingly gated not by chip design or manufacturing process, but by the availability of better underlying materials. Finding a commercially viable new material today typically requires 10+ years of iterative laboratory work across thousands of candidate experiments. Matforge's platform deploys a swarm of AI agents to compress that timeline to months, autonomously generating new material candidates, coordinating their synthesis, and testing them in physical labs - covering the full end-to-end materials discovery pipeline.

The company's initial focus is on materials for datacenters and fabs, two of the most energy- and resource-intensive nodes of the semiconductor supply chain. With AI chip power consumption doubling approximately every year, the thermal and electrical properties of packaging and interconnect materials are becoming critical bottlenecks. Akash Ramdas, co-founder and materials science expert, conducted his PhD and postdoc at Stanford specifically on novel interconnect materials for nanoscale electronics - work that was adopted into the official technology roadmaps of both Intel and TSMC. This gives Matforge an unusually direct line from foundational research to industrial adoption. Advaith Sridhar, the AI/ML co-founder, previously built long-horizon autonomous agent systems at Luma Labs and Persona AI, bringing frontier model expertise directly to the scientific discovery workflow.

For the US energy supply chain, Matforge represents a high-leverage intervention point: by unlocking materials with better thermal conductivity, lower resistivity, or superior dielectric properties, the company could meaningfully reduce the energy demands of AI training and inference infrastructure, lower the cost and improve the reliability of domestic semiconductor fabrication, and reduce dependence on materials currently sourced from geopolitically sensitive supply chains. As the US works to re-shore semiconductor manufacturing through the CHIPS Act and related initiatives, companies like Matforge that can accelerate materials innovation are strategically critical to making next-generation domestic fabs viable and competitive.

Funding

Backed by Y Combinator (Spring 2026 batch). Standard YC pre-seed funding (~$500K) assumed; no additional disclosed rounds as of early 2026.

Investors

  • Y Combinator

Key People

  • Akash Ramdas, Co-Founder (PhD & PostDoc, Stanford - Materials Science & Engineering; interconnect materials research adopted by Intel and TSMC)
  • Advaith Sridhar, Co-Founder (MS AI, Carnegie Mellon; ex-Research Engineer at Luma Labs and Persona AI)

News