Companies

83 Sciences

83sciences.ai

83 Sciences uses AI to mine discarded experimental data and accelerate materials discovery for scientific research labs.

HQSan Francisco, California, United States
Employees1-50
3 active roles
Jobs checked 16h ago
AI / MLAI Application

About

83 Sciences builds Dalton, an AI platform that mines discarded or unpublished experimental data to help research labs discover materials and improve experiments. It appears to sell through partnerships with laboratories, differentiating itself through access to otherwise discarded data and rapid conversion of partner data into co-authored scientific manuscripts.

Market

83 Sciences competes in AI-native scientific R&D and materials-discovery software, adjacent to the self-driving-laboratory market. Its data-first positioning combines capture of unpublished experiments and historical failed batches with a queryable scientific brain and Dalton AI, then connects discoveries to papers, patents, scale-up, and commercialization. Compared with platforms centered more heavily on robotics, LIMS/ELN integration, or closed-loop autonomous labs, its clearest differentiation is extracting value from existing experimental data through discovery contracts and partnerships.

Target Customers

83 Sciences targets industrial materials and chemistry R&D organizations seeking faster commercialization of new materials, especially teams with large stores of unpublished or discarded experimental data. It also serves academic labs and industry-academia partnership teams; likely users and buyers include materials scientists, lab leaders, and heads of R&D.

At a Glance

Problem

Materials and industrial research labs generate large volumes of experimental data, but 83 Sciences says more than 83% of it is discarded, representing over $100 billion in lost R&D value each year. The pain is both economic and scientific: researchers repeat work, miss insights in failed experiments, and struggle to turn scattered notes and instrument outputs into commercially useful knowledge. The core use case is recovering value from unpublished or unsuccessful experiments to identify new materials and optimized process conditions, then convert those findings into papers, patents, scale-up opportunities, and faster, cheaper commercialization.

Product / Service

83 Sciences combines AI-native research software with a partnership-led discovery service. Its system captures and structures information ranging from voice notes to instrument output, creating a searchable, agentic record of experiments that can interpret failures and propose improved process conditions. For industrial customers, the company offers discovery contracts; for academic labs, it offers a way to uncover additional papers and commercialization opportunities in existing data. The stated benefit is to move labs from raw experimental history to concrete outputs such as manuscripts, patents, optimized processes, and industry partnerships, with the company claiming it can turn a partner’s discarded data into a co-authored manuscript in less than two months.

Market

83 Sciences operates at the intersection of AI-native materials discovery, scientific research software, and laboratory data integration, serving materials scientists in industrial R&D and academia. Its approach is differentiated by mining unpublished experimental data rather than relying primarily on simulation or new experiments. Adjacent competitors include Schrödinger’s molecular-modeling platform for materials prediction, CuspAI’s AI materials-discovery system, and Benchling’s cloud platform for organizing scientific data and experiments; these overlap with parts of 83 Sciences’ workflow but are not shown in the available evidence to be direct like-for-like competitors.

The company appears to be at a very early commercial stage. Y Combinator lists it as founded in 2026, part of the Summer 2026 batch, active, and a three-person team, while its public profile shows launch activity, recruiting, and early visibility rather than disclosed customer or revenue metrics. The website says it is backed by Y Combinator and invites industry partners to scope discovery contracts, but no public customer names, revenue, or funding amount appear in the available evidence; accordingly, it is best characterized as pre-revenue or early commercialization rather than a company with demonstrated market traction.

Founders & Leadership

Ian NaccarellaFounder
Co-Founder & CEO
Eric RieselFounder
Co-Founder & CTO
Yankang YangFounder
Co-Founder & COO

Funding History

2026-06
Seed$125k

Y Combinator

Recent News

2026-07-24
83 Sciences listed as an active S2026 hard-tech startup by Y Combinator

Y Combinator’s hard-tech directory listed 83 Sciences as an active Summer 2026 startup with three employees in San Francisco. The listing describes its use of AI to mine discarded experimental data for scientific breakthroughs.

2026-07-19
Y Combinator S26 Batch Companies

An independent roundup of Y Combinator’s Summer 2026 companies described 83 Sciences as an intelligence engine and workspace for scientific research, with an emphasis on materials scientists.

2026-07-09product
Founding AI Engineer at 83 Sciences

A Y Combinator job listing identified 83 Sciences as a YC S26 company building an intelligence engine for research and materials discovery. The role description says the company structures raw laboratory signals and experimental data to surface new discoveries.

2026-07-06
83 Sciences: AI-native materials discovery powered by unpublished experimental data

Y Combinator’s company profile introduced 83 Sciences as a 2026-founded startup using AI to mine discarded experimental data and working alongside laboratories to discover new materials.

Active Roles

3
San Francisco, CA, US / New York, NY, US / Remote (US)/Engineering/32d ago
New York, NY, US / San Francisco, CA, US / Remote (US)/Sales/32d ago
New York, NY, US / Remote (US)/Engineering/32d ago

Business Model

83 Sciences appears to use a B2B partnership model, working with laboratories and research organizations to analyze unused experimental data through its AI platform. Its public website promotes partnership inquiries and does not disclose specific pricing or subscription terms.

Products

83 Sciences PlatformDalton AI lab-brain interfaceScientific data capture and structuring layerDiscovery contracts and materials-discovery partnershipsResearch-to-commercialization support for papers, patents, scale-up, and industry-academia partnerships

Customers

No named enterprise customers publicly disclosed; 83 Sciences references an unnamed battery materials lab partner (name withheld).

Tech Stack

Dalton AI conversational interfaceAgentic AI and scientific modelsStructured, queryable experimental-data corpusMultimodal experimental-data ingestion and structuring for voice notes, instrument outputs, notebooks, and historical runs

Competitors

Atinary
Citrine Informatics
Kebotix
Benchling
Radical AI
IBM RoboRXN