# 83 Sciences: materials discovery from unpublished experimental data

Canonical: https://mudpie.ai/companies/83-sciences/
Breadcrumb: [Home](https://mudpie.ai/) / [Companies](https://mudpie.ai/companies/) / [83 Sciences: materials discovery from unpublished experimental data](https://mudpie.ai/companies/83-sciences/)
Author: Ali Abouelatta (https://mudpie.ai/authors/ali-abouelatta/)
Published: 2026-09-19
Updated: 2026-09-19
Research type: Company profile
Method: Company and accelerator sources checked 2026-09-19. Product claims are attributed to their sources; this is research, not a hands-on product trial.

83 Sciences uses AI to turn unpublished experimental data into materials-discovery work for industrial R&D and academic labs. The decision is whether a team wants a partner that can structure its discarded experiments and move toward a material, paper, patent, or scale-up question—not just another model trained on published literature.

## What it does

83 Sciences says it captures and structures data from voice notes through instrument output, builds a queryable record of experiments, and uses agents to propose materials and process conditions. Its public priority areas include critical minerals, energy storage, catalysis and chemicals, life-science solid forms, and semiconductors ([83 Sciences homepage](https://83sciences.ai/); [83 Sciences team](https://83sciences.ai/team)).

| Fact | What the public sources say |
| --- | --- |
| Buyer | Industrial R&D teams and academic labs |
| Input | Unpublished experimental data, from lab notes to instrument output |
| Output | Discovery candidates, process conditions, papers, patents, and commercialization paths |
| Public timing claim | The homepage says it can reach a first new material from partner data in under six weeks |
| Founders | Ian Naccarella, Eric Riesel, and Yankang Yang |

## Why it fits

The wedge is the data most frontier models do not see: failed experiments, tacit lab context, and process details that never became a paper. An industrial team can care less about a benchmark score than whether the system learns its actual materials history and produces a candidate that survives a real experiment. The academic path adds a different buyer job: turn years of unstructured work into papers, citations, and commercialization opportunities.

The public results are company claims. The YC launch says the team has discovered a novel material and drafted a manuscript; the homepage says more than 85% of experimental data is discarded, cites $100B+ in lost R&D value, and advertises the under-six-week path ([83 Sciences YC launch](https://www.ycombinator.com/launches/SIv-83-sciences-materials-discovery-from-the-90-of-experimental-data-that-never-gets-published); [83 Sciences homepage](https://83sciences.ai/)). A buyer should resolve data ownership, reproducibility, lab validation, IP terms, scale-up support, and how negative results are preserved rather than silently filtered.

The founder mix matches the commercial-science problem. YC describes Naccarella as having worked on battery materials at Sila, Riesel as an MIT inorganic-chemistry PhD and ML researcher, and Yang as a former BCG principal who led an AI program ([YC company profile](https://www.ycombinator.com/companies/83-sciences)). Pricing was not published; the site directs industry teams to scope a discovery contract.

Short version: 83 Sciences is a strong fit when valuable experiments are trapped in lab files and the buyer can fund real validation. It is not a shortcut around chemistry, ownership, or manufacturing constraints.

## Sources checked — 2026-09-19

- [YC company profile](https://www.ycombinator.com/companies/83-sciences)
- [83 Sciences homepage](https://83sciences.ai/)
- [83 Sciences team](https://83sciences.ai/team)
- [83 Sciences YC launch](https://www.ycombinator.com/launches/SIv-83-sciences-materials-discovery-from-the-90-of-experimental-data-that-never-gets-published)

## Cohort context

83 Sciences is listed in Summer 2026. In our 2026-09-18 directory snapshot, 56 of 232 listed companies in that cohort have YC’s primary industry label Industrials (24.1%). This is a current-directory comparison, not an original intake count or a performance ranking. [Nine-cohort dataset](https://mudpie.ai/research/yc-cohorts-2026-09-19.json).

## Public website snapshot

Observed 2026-09-19T16:18:31.038Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.

| Signal | Homepage observation |
| --- | --- |
| Product description metadata | Observed |
| Canonical link | Not observed in this response |
| H1 or H2 heading | Observed |
| Typed structured data | Not observed in this response |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |

[Public observations](https://mudpie.ai/research/yc-homepage-links-2026-09-19.json) · [Collection method](https://mudpie.ai/research/yc-homepage-methods/README.md). Missing links here do not establish that a capability or file is absent elsewhere.


## Author disclosure

I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.