Discovered Materials
Discovered Materials (Matforge Inc.) builds AI agents that autonomously discover new semiconductor materials, targeting chipmakers needing thermally conductive dielectrics for 3D packaging. Pre-revenue, backed by Lightspeed-led $9M seed, with open-source Material Discovery Bench co-built with IBM, IMEC, Stanford, and Cambridge.
- Company typePrivate
- Founded2026
- HeadquartersSan Francisco, United States
- Headcount51–100
- GTM typeB2B
- OfferingSoftware
What Discovered Materials does
Discovered Materials (legally Matforge Inc., a Delaware corporation headquartered at 128 King Street, San Francisco) builds AI agents that autonomously discover new materials for semiconductor chips, compressing months of interdisciplinary scientific work into days. The agents interact with computational materials science tooling, including web search, a coding sandbox built on pymatgen, mp_api, and ASE, and Machine-Learning Interatomic Potential (MLIP) property calculators based on the PET-MAD foundation model, to screen candidate crystalline materials against multi-objective constraints such as thermal conductivity, dielectric constant, and mechanical moduli, and to propose synthesis recipes. In parallel, the company releases and maintains the Material Discovery Bench, an open-source long-horizon benchmark built in collaboration with researchers from IBM, IMEC, Stanford, and Cambridge, which has been used to evaluate frontier models including GPT-5.6 Sol, Claude Opus 5, Claude Fable 5, and Kimi K3 and has surfaced over 500 computationally novel materials publicly.
The business is pre-revenue and at an early research-to-commercialization stage. The company does not publicly disclose pricing or product tiers; the stated revenue model is a recurring subscription arrangement for AI-powered materials discovery sold to semiconductor manufacturers (logic and memory chipmakers and data center hardware vendors) facing thermal management constraints, particularly for 3D chip packaging and BEOL-compatible thermally conductive dielectrics. Distribution is primarily through the company's website, open-source research releases, and earned media (TechCrunch, CNBC, The Economic Times, Business Standard) used to recruit talent and attract enterprise partners. Target customer acquisition is expected to flow from direct enterprise engagement rather than a self-serve or channel model.
The company is led by co-founders Advaith (AI background, Carnegie Mellon; ex-Persona AI, Luma Labs) and Akash (Stanford PhD in Material Science, 11 years of semiconductor materials research), who met at IIT-Madras. It is backed by a $9M seed round led by Lightspeed, with participation from Y Combinator, Peak XV, and angels Paul Graham, Gokul Rajaram, and Thariq Shihipar. The team has grown to 51-100 employees and is actively recruiting AI/ML engineers and materials scientists via the Y Combinator job portal. Near-term execution risk centers on converting computationally discovered materials into fab-qualified recipes, since only one of the 500+ candidates is currently described as having a plausible synthesis pathway.
Discovered Materials firmographics
Firmographics- Name
- Discovered Materials
- Legal name
- Matforge Inc.
- Website
- https://discoveredmaterials.com
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Headcount range
- 51–100 employees
- Short description
- Discovered Materials (Matforge Inc.) builds AI agents that autonomously discover new semiconductor materials, targeting chipmakers needing thermally conductive dielectrics for 3D packaging. Pre-revenue, backed by Lightspeed-led $9M seed, with open-source Material Discovery Bench co-built with IBM, IMEC, Stanford, and Cambridge.
- Ownership category
- akta.pro rank
Where Discovered Materials is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
Discovered Materials business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Operations
Revenue model
- AI-Powered Materials Discovery as a Service: The company develops AI agents that discover new materials for semiconductor manufacturers. Revenue model is not explicitly stated but the company is pre-revenue, focused on product development and market validation. Target customers are semiconductor chip manufacturers and materials suppliers.
Go-to-market motion2 records
Distribution channels1 record
Marketing channels8 records
Discovered Materials product offering
Product offeringCore offering
Discovered Materials develops autonomous AI agents that perform semiconductor-materials discovery. The agents use web search, Python coding, computational materials tools, ML interatomic potentials, and property calculators to screen candidate materials against thermal, dielectric, mechanical, and stability constraints, then propose synthesis recipes.
Product overview
Discovered Materials operates as a two-product company: it builds AI agents that discover new materials for semiconductor chips (core product), and simultaneously releases the Material Discovery Bench — an open-source long-horizon benchmark that measures frontier LLM progress on this same materials discovery task. The AI agents and the benchmark are interdependent: the agents perform the discovery work, while the benchmark evaluates how well frontier models (including the company's own agents) perform at generating novel, stable materials that meet multi-objective semiconductor property constraints and include plausible synthesis recipes.
Differentiator
Problem solved
Functional benefit
Products and services
- AI Agents for Semiconductor Materials Discovery Autonomous AI agents for semiconductor manufacturers and materials suppliers that run materials simulations, generate candidate materials, and propose synthesis recipes while screening thermal, dielectric, mechanical, and stability requirements. Companies interested in licensing discoveries or co-developing materials engage directly with the company.
Quantifiable outcome
- Over 500 previously unknown thermally conductive dielectric materials computationally discovered and publicly released for further study
- +4 more outcomes
Companies that use Discovered Materials
Customer profileSegments1 record
Ideal customer profiles2 records
Discovered Materials technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability3 records
Feature5 records
Discovered Materials partnerships and signals
Strategic signalPartnerships
Four partnerships are on record, tiered major.
- IBMmajorIBM researchers (Dr. Daniel Edelstein) collaborated in building the Material Discovery Bench and reviewed synthesis rubrics. IBM contributes expertise in semiconductor materials and BEOL-compatible processes.
- IMECmajorIMEC researchers (Dr. Zsolt Tokei, Dr. Geoffrey Pourtois) collaborated in building the Material Discovery Bench and reviewed synthesis rubrics. IMEC is a leading European research organization in nanoelectronics and semiconductor technology.
- Stanford UniversitymajorStanford researchers (Prof. Krishna Saraswat) collaborated in building the Material Discovery Bench. Prof. Girolami (UIUC) and multiple Cambridge researchers also contributed as reviewers. Co-founder Akash holds a PhD from Stanford.
- University of CambridgemajorCambridge researchers (Prof. Judith MacManus-Driscoll, Dr. Sebastian Dixon, Dr. Manisha Bansal) collaborated as reviewers of the Material Discovery Bench benchmark and synthesis rubrics.
Scale indicators4 records
Recent moves5 records
Expansion highlights5 records
Discovered Materials competitors and assessment
Company assessmentDirect peers
- Citrine Informatics: Citrine Informatics is an AI platform purpose-built for materials and chemicals R&D, serving enterprise customers with data infrastructure and ML-driven candidate screening. It is the closest direct peer to Discovered Materials in applying AI to industrial materials discovery.
- Kebotix: Kebotix combines AI and autonomous robotics to discover new chemicals and materials, with a similar mission of compressing lab-to-market timelines. It is directly comparable as an early-stage AI-for-materials peer.
- Orbital Materials: Orbital Materials is an AI-for-materials company founded by ex-DeepMind researchers, focused on designing novel materials using foundation models. It overlaps closely with Discovered Materials in applying frontier LLMs to industrial materials discovery.
- Cusp AI: Cusp AI applies generative AI to the design of new materials, with similar early-stage positioning and a comparable AI-first discovery thesis. It is a direct peer in the AI-driven materials design category.
Broad incumbents
- Schrödinger: Schrödinger is a publicly traded computational chemistry and materials simulation platform serving pharmaceuticals and industrial materials R&D. It is a much broader incumbent whose materials platform competes in the same end-market as Discovered Materials.
- Synopsys: Synopsys is a leading EDA and semiconductor IP vendor whose tools increasingly incorporate materials-aware simulation. It is a broad incumbent in the semiconductor design-and-manufacturing stack that touches the same end-customer relationships Discovered Materials targets.
- Cadence Design Systems: Cadence is a major EDA and computational software provider for semiconductor and system design, with overlap in materials and physics simulation used in chip development. It is a broad incumbent in the same semiconductor value chain.
- Ansys: Ansys provides multiphysics simulation software widely used for materials, thermal, and semiconductor design. It is a broad incumbent whose simulation capabilities overlap with the computational materials workflow Discovered Materials automates.
Emerging players
- Uncountable: Uncountable provides an R&D data and informatics platform for materials and chemicals teams, sitting adjacent to AI-driven discovery. It is comparable as a software layer helping industrial R&D organizations structure and accelerate materials development.
Others
- Google DeepMind (GNoME): DeepMind's GNoME project used AI to discover millions of new crystalline materials and is the most prominent AI-for-materials research effort to date. It is not a commercial product but is thematically adjacent and pressures the open research positioning of Discovered Materials.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key risks6 records
Key highlights7 records
Customer concentration
Discovered Materials social profiles
Digital presenceDiscovered Materials financial estimates
Financial estimateRevenue estimate
Valuation estimate
Discovered Materials leadership team
Management profileNumber of profiles
Profiles2 records
Discovered Materials funding detail
Funding detailFunding overview
Funding rounds1 record
Investors3 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Discovered Materials M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Discovered Materials
What does Discovered Materials do?
Discovered Materials develops autonomous AI agents that perform semiconductor-materials discovery. The agents use web search, Python coding, computational materials tools, ML interatomic potentials, and property calculators to screen candidate materials against thermal, dielectric, mechanical, and stability constraints, then propose synthesis recipes.
Is Discovered Materials a public or private company?
Discovered Materials is a private company. It is classified as venture growth investor backed and is currently operating.
When was Discovered Materials founded?
Discovered Materials was founded in 2026. It employs 51 to 100 people.
Where is Discovered Materials based?
Discovered Materials is headquartered in San Francisco, United States, in the North America region.
How does Discovered Materials make money?
One revenue line is on record: AI-Powered Materials Discovery as a Service.
Who are Discovered Materials's main competitors?
Direct peers on record are Citrine Informatics, Kebotix, Orbital Materials and Cusp AI. Broad incumbents are Schrödinger, Synopsys, Cadence Design Systems and Ansys. Uncountable is listed as an emerging player. Google DeepMind (GNoME) is listed as an others.
Does Discovered Materials have an API?
No public API is recorded for Discovered Materials.