Is Discovered Materials profitable?

No. As of February 2026, Discovered Materials reports $0 revenue and has raised only $500K total via a convertible note dated March 1, 2026. The two-person team is pre-revenue and pre-customer, still cold-emailing chip companies for introductions as of July 2026.

How many employees does Discovered Materials have?

Discovered Materials has 2 employees, the co-founders, based in San Francisco as of July 2026. There are no active job postings on its YC page, suggesting no near-term hiring push despite competing against far larger rivals like CuspAI.

Who is Discovered Materials' biggest competitor?

CuspAI is the most direct threat: it raised a $450M Series B in July 2026 at a $2.6B valuation and runs a 48-plus partner 'AI Materials Foundry' coalition including Nvidia, Meta, and AMD, with 80% of its 2026 research focused on semiconductors, the same niche Discovered Materials targets.

Could AI replace what Discovered Materials is trying to build?

The company's own co-founders say frontier general-purpose AI models score only about 10% on their internal materials-science evaluations as of June 2026, meaning off-the-shelf AI cannot yet do the job; they call it 'the GPT-3.5 moment' for the domain, requiring proprietary domain-specific training.

Does Discovered Materials have any customers among chipmakers?

No confirmed commercial customers as of July 2026. The company states it is actively seeking introductions to packaging and thermal engineers at NVIDIA, AMD, Groq, Cerebras, Google, Intel, Samsung, and TSMC, which implies outreach is still in progress rather than closed deals.

Why did Discovered Materials change its name from Matforge?

The company launched on Y Combinator in April 2026 as 'Matforge' and was referenced under that name in a June 2, 2026 Substack interview, but by July 2026 YC's own directory and its site discoveredmaterials.com listed it as 'Discovered Materials,' the same two-founder entity.

YC Spring 2026 · United States

Discovered Materials

Two-person startup using AI agents and DFT to find new chip interconnect materials.

Protection under test

Assessed 27 July 2026 · Evidence as of 27 July 2026

Something worth protecting exists and is being tested right now.

On the recordSince 27 July 2026 · Checkable from Q3 2026 · Every call on the record Checkpoint: Runway from the $500K note runs out, forcing a seed raise or a named fab design-partner win

The call

Two founders with real chip-industry credibility are racing billion-dollar AI materials rivals to land a fab customer before their funding runs out.

Discovered Materials uses AI and lab experiments to find new materials for computer chips faster than traditional methods. The founders know the field well and their approach makes sense, but bigger competitors like CuspAI (with $2.6 billion in backing) and DeepMind's free discoveries are moving in the same direction with far more money and researchers. The biggest wildcard is whether they can convince a major chip factory to work with them exclusively before their money runs out, which would give them real data and lock-in that competitors cannot easily copy.

Why

AI pressure: CuspAI's AI Materials Foundry and DeepMind's GNoME database. CuspAI's 48-partner coalition and DeepMind's free GNoME candidates already generate materials leads at a scale Discovered Materials cannot match yet. Could AI do the core job? Yes, likely.

Bigger players: Nvidia, Meta, and AMD via CuspAI's partner coalition. Nvidia, Meta, and AMD already back CuspAI's funded materials coalition, giving chipmakers an alternative before Discovered Materials lands one customer. Could a giant copy it? Yes, easily.

What protects it. Co-founder Akash Ramdas's PhD interconnect materials already sit on Intel and TSMC roadmaps, a specific credibility rivals lack.

What would change this call

CuspAI announces an exclusive interconnect or packaging partnership with a named chipmaker within 6 months. Would show CuspAI capturing the fab relationships Discovered Materials needs before it can compete.
Discovered Materials announces a seed round or names a fab design-partner (Intel, TSMC, Samsung). Would show the company converting founder credibility into real commercial traction.
Discovered Materials' current $500K runway runs out without a follow-on raise by early 2027. Would show the company failing before reaching the ~2-year semiconductor qualification milestone.

Key numbers

CuspAI raised $450M, now valued at $2.6B (2026-07-20)
Frontier AI models score ~10% on their materials evals (2026-06-02)
2 employees, $500K raised, still pre-customer (2026-07-21)

Named risks

Only $500K raised with 2 employees against rivals funded at hundreds of millions to billions. Severity: high
No shipped dataset, IP, or customer lock-in; moat rests only on founder tacit expertise. Severity: high
~2-year semiconductor qualification cycle means no revenue inflection likely within current runway. Severity: high
CuspAI's 48-partner coalition and DeepMind's GNoME already operate at scale in the same niche. Severity: high
Brand confusion between 'Matforge' and 'Discovered Materials' across YC and press pages. Severity: low

Rivals, as read in July 2026

Rival	Direction	Evidence
CuspAI	losing to	$450M raise, $2.6B valuation, 48-partner coalition targeting semis, July 2026
DeepMind GNoME	losing to	2.2M AI-discovered structures published free with Berkeley Lab synthesis partnership, July 2026
Periodic Labs	losing to	Founded by ex-OpenAI/DeepMind researchers, reportedly seeking ~$7B valuation, March 2026
Lila Sciences	losing to	~$550.7M raised across seed and Series A by late 2025, broader materials AI scope
83 Sciences	contesting	Similarly small YC S26 peer mining unpublished lab data for materials discovery, July 2026
Radical AI	losing to	$55M seed plus robotic autonomous labs and an Air Force contract, July 2025

Questions people ask

Is Discovered Materials profitable?: No. As of February 2026, Discovered Materials reports $0 revenue and has raised only $500K total via a convertible note dated March 1, 2026. The two-person team is pre-revenue and pre-customer, still cold-emailing chip companies for introductions as of July 2026.
How many employees does Discovered Materials have?: Discovered Materials has 2 employees, the co-founders, based in San Francisco as of July 2026. There are no active job postings on its YC page, suggesting no near-term hiring push despite competing against far larger rivals like CuspAI.
Who is Discovered Materials' biggest competitor?: CuspAI is the most direct threat: it raised a $450M Series B in July 2026 at a $2.6B valuation and runs a 48-plus partner 'AI Materials Foundry' coalition including Nvidia, Meta, and AMD, with 80% of its 2026 research focused on semiconductors, the same niche Discovered Materials targets.
Could AI replace what Discovered Materials is trying to build?: The company's own co-founders say frontier general-purpose AI models score only about 10% on their internal materials-science evaluations as of June 2026, meaning off-the-shelf AI cannot yet do the job; they call it 'the GPT-3.5 moment' for the domain, requiring proprietary domain-specific training.
Does Discovered Materials have any customers among chipmakers?: No confirmed commercial customers as of July 2026. The company states it is actively seeking introductions to packaging and thermal engineers at NVIDIA, AMD, Groq, Cerebras, Google, Intel, Samsung, and TSMC, which implies outreach is still in progress rather than closed deals.
Why did Discovered Materials change its name from Matforge?: The company launched on Y Combinator in April 2026 as 'Matforge' and was referenced under that name in a June 2, 2026 Substack interview, but by July 2026 YC's own directory and its site discoveredmaterials.com listed it as 'Discovered Materials,' the same two-founder entity.

Timeline

2026-07-21Listed in YC's Semiconductors directory with 2 employees, San Francisco HQ, P26 batch
2026-07-20CuspAI raises $450M Series B at $2.6B valuation, 80% research focus on semiconductors
2026-07-06DeepMind GNoME blog reaffirms 2.2M candidate materials, 380K stable, Berkeley Lab partnership
2026-06-02Substack interview details three-stage agent plus DFT plus lab-loop architecture and 10% model score
2026-04-21Company launches on YC under the name Matforge
2026-03-01Company raises $500K via convertible note, its total funding to date
2025-10-31Rival Lila Sciences reaches ~$550.7M total funding after Series A
2025-07-31Rival Radical AI raises $55M seed led by RTX Ventures with Nvidia's NVentures

Revision history

27 July 2026

First assessed. Protection under test. The text above is fixed as of this date and is not edited afterwards.

In the same band

Ploy Standout Playabl.ai