Discovered Materials
Category: AI Chips / Semiconductors
AI agents ("AI scientists") that discover novel materials for semiconductor chips, targeting heat reduction and power efficiency in datacenters and fabs. Discovered Materials was founded in 2026. The company is led by Advaith Sridhar. Based in San Francisco, California, US. Team size: 1-10. Total funding raised: $9M. Latest round: Seed. Key investors include Lightspeed (Lightspeed India Partners), Y Combinator, Peak XV Partners, Paul Graham (angel), Gokul Rajaram (angel), Thariq Shihipar (angel).
- Founded
- 2026
- Headquarters
- San Francisco, California, US
- Team size
- 1-10
- Total funding
- $9M
Value proposition
Compresses the traditional 10+ year materials discovery timeline for semiconductor chips down to months using a swarm of AI agents that simulate, synthesize, and test new materials in physical labs.
Products and solutions
AI agent platform for end-to-end materials discovery (candidate generation, synthesis, physical lab testing), Material Discovery Bench (open-source benchmark for AI-driven materials discovery built with IBM, IMEC, Stanford, Cambridge).
Unique value
Combines a materials-science PhD founder with an AI/agents research engineer to build autonomous 'AI scientist' agents spanning the full discovery pipeline — from candidate generation to physical synthesis and lab testing — rather than pure computational/simulation-only approaches.
Target customer
Semiconductor and chip manufacturers (e.g., NVIDIA, AMD, Intel, TSMC, Samsung, Groq, Cerebras) and datacenter/fab operators needing better thermal and interconnect materials.
Industries served
Semiconductors, AI chip manufacturing, datacenter infrastructure
Technology advantage
Founder's prior Stanford PhD research on nanoscale interconnect materials was adopted into Intel and TSMC roadmaps; combines physical lab synthesis/testing with AI agent orchestration (not simulation-only); released open-source Material Discovery Bench for credibility/community validation.
How they differentiate
Focused specifically on semiconductor/chip thermal and interconnect materials (vs. broader materials science scope of Lila Sciences/Periodic Labs); combines physical synthesis and lab testing with AI agents rather than pure simulation.
Main competitors
Periodic Labs, Lila Sciences, Orbital Materials (Orbital Industries), CuspAI
Key partnerships
Material Discovery Bench built in collaboration with IBM, IMEC, Stanford, and Cambridge, prior research materials adopted into Intel and TSMC roadmaps (research collaboration, not commercial partnership)
Major milestones
Y Combinator Spring 2026 batch, during the 3-month YC batch, synthesized and tested thermal interface materials matching performance of products major chemical companies have guarded as trade secrets for 20+ years, released open-source Material Discovery Bench, closed $9M seed round led by Lightspeed (Aug 2026), rebranded from initial YC launch name 'Matforge' to 'Discovered Materials'.
Market positioning
Early-stage (YC S26, seed) specialist entrant in the emerging 'AI for materials discovery' category, narrowly targeting the semiconductor heat/power efficiency problem versus broader-scope rivals with much larger war chests (Periodic Labs $300M seed, CuspAI $450M Series B).
Geographic focus
US-centric with global chip industry customers (targeting NVIDIA, AMD, Intel, TSMC, Samsung); competitor CuspAI is UK-based.
About Advaith Sridhar
MS AI, Carnegie Mellon; founding applied scientist at Persona AI (acquired by Luma Labs) building long-horizon autonomous agents; research engineer at Luma Labs (agent harnesses, video models, frontier models); IIT Madras
Latest news about Discovered Materials
More AI Chips / Semiconductors companies
Official website: https://discoveredmaterials.com
