83 Sciences
Early-stage startup mining labs' unpublished experimental data to train AI for materials discovery.
On the record
Checkpoint: YC Demo Day funding announcement and whether the co-authored manuscript case study gets published independently
The call
83 Sciences has a real idea in a crowded field, but its only proof point is an unverified case study while far richer rivals chase the same data.
83 Sciences looks for old, unused experimental data from drug labs and uses AI to find patterns that could lead to new discoveries. The problem is that bigger companies like Periodic Labs and Lila Sciences could do the exact same thing, and they have way more money to spend. The only thing that could make 83 Sciences survive is if they lock labs into their workflow so deeply that switching costs too much, but no one has proven that works yet. If they don't publish real proof of their idea working, or raise enough money in the next six months, they will likely get squeezed out or bought for parts by larger competitors.
Why
01
AI pressure: Periodic Labs and Lila Sciences. Both run full-stack AI-scientist platforms with $7B+ valuations that can bolt on dark-data mining without needing 83 Sciences' wedge. Could AI do the core job? Only partly.
02
Bigger players: Benchling and other ELN/LIMS vendors. These vendors already sit inside labs holding the raw data, they just lack the analysis layer 83 Sciences is betting on. Could a giant copy it? Yes, easily.
03
What protects it. None apparent yet; the plan is exclusive access to unpublished lab data, but no signed multi-year exclusivity deals are confirmed.
What would change this call
83 Sciences announces YC S26 demo day funding within 6 months. Shows whether investors value its dark-data moat over rivals like Periodic Labs.
Co-authored manuscript from its case study is independently published. Would verify its one proof point is real, not vendor claim.
Periodic Labs or Lila Sciences ships a dark-data mining feature. Would show its wedge is easily copied by better-funded rivals.
Co-authored manuscript from its case study is independently published. Would verify its one proof point is real, not vendor claim.
Periodic Labs or Lila Sciences ships a dark-data mining feature. Would show its wedge is easily copied by better-funded rivals.
Key numbers
- Team of 3 founders as of mid-2026
- Periodic Labs valuation talks near $7B
- Citrine Informatics has run this motion 13 years with no dominant share
Named risks
- No signed exclusivity contracts with labs; data moat is aspirational, not locked in.
- Better-capitalized rivals (Periodic Labs, Lila Sciences) can bolt on same dark-data mining motion.
- Single case study is vendor-authored and unverified, with no named lab or independent publication.
- On-site, white-glove go-to-market is services-heavy and may not scale economically.
- Frontier labs (e.g. Google DeepMind) already ship native materials-discovery AI, reducing differentiation.
Rivals, as read in July 2026
| Rival | Direction | Evidence |
|---|---|---|
| Periodic Labs | losing to | Raised $300M seed Sept 2025, deal talks at ~$7B valuation Mar 2026, full-stack autonomous lab platform |
| Lila Sciences | losing to | Raised $200M+$350M, in talks for $8.5-10.5B valuation by Jun 2026, backed by CalPERS and NVentures |
| Citrine Informatics | contesting | 13-year incumbent since 2013 with on-site lab data capture motion, no dominant share yet |
| Radical AI | losing to | Raised $55M seed Jul 2025 and won AFWERX Air Force Phase II contract same month |
| CuspAI | contesting | Markets Autonomous Materials Search Engine with vendor-claimed experimentation time cuts, as of Dec 2025 |
| Benchling and ELN/LIMS vendors | contesting | Already own raw lab data as system of record, lacking only the AI analysis layer 83 Sciences offers |
Questions people ask
- Is 83 Sciences profitable?
- No. 83 Sciences is a pre-revenue, 3-person Y Combinator Summer 2026 batch startup founded in 2026. No revenue figures, customer counts, or enterprise logos have been disclosed as of July 2026, so profitability is not applicable at this stage.
- Could AI or big rivals replace 83 Sciences?
- Very possibly. 83 Sciences' entire product is an AI application mining unpublished lab data, with no non-AI workflow behind it. Better-funded rivals like Periodic Labs (~$7B valuation, March 2026) and Lila Sciences (~$8.5-10.5B talks, June 2026) could add the same capability.
- Who owns 83 Sciences?
- 83 Sciences is privately founder-owned, currently in Y Combinator's Summer 2026 (S26) batch. Founders listed are Ian Naccarella, Eric Riesel, and Yankang Yang, per YC's jobs page as of July 2026. No outside lead investor or valuation has been disclosed.
- Who is 83 Sciences' biggest competitor?
- Periodic Labs and Lila Sciences are the largest threats: Periodic Labs raised a $300M seed in September 2025 and was in deal talks near a $7B valuation by March 2026, while Lila Sciences sought a ~$2B Series B at $8.5-10.5B by June 2026, dwarfing 83 Sciences' scale.
- What is 83 Sciences' actual product moat?
- 83 Sciences claims a data moat from on-site capture of labs' unpublished and discarded experimental data, plus workflow lock-in. As of July 2026, no signed exclusivity contracts or named customers have been disclosed, so the moat remains unproven.
- Has 83 Sciences shown any real results?
- The company cites one case study: analysis of an unnamed lab's discarded characterization data reportedly led to a co-authored manuscript in under two months. This is a vendor-authored claim with no named lab or independent verification as of the July 2026 profile.
Timeline
- 2026-07-21Third-party directory lists 83 Sciences founders from MIT, Stanford, Harvard, Columbia
- 2026-07-0983 Sciences posts Founding AI Engineer job listing, $120K-$250K plus equity
- 2026-0783 Sciences confirmed as part of Y Combinator Summer 2026 (S26) batch, team of 3
- 2026-06Rival Lila Sciences reportedly in talks for ~$2B Series B at $8.5-10.5B valuation
- 2026-03-25Rival Periodic Labs reportedly in deal talks at about $7B valuation
- 2025-10Lila Sciences raises $350M Series A at over $1.3B valuation
- 2025-09Rival Periodic Labs raises record $300M seed round
- 2025-07Rival Radical AI raises $55M seed, wins AFWERX Air Force contract
- 2025-03Lila Sciences raises $200M seed round
- 2023Lila Sciences founded as a Flagship Pioneering spinout
Revision history
27 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources
- 83 Sciences: AI-native materials discovery powered by unpublished experimental data
- Jobs at 83 Sciences
- 83 Sciences, API Provider, Schemas (2026-07-21)
- Founding AI Engineer at 83 Sciences (2026-07-09)
- 83sciences.ai
- Top Citrine Informatics Alternatives, Competitors
- Top Kebotix Alternatives, Competitors
- Report: Periodic Labs Business Breakdown & Founding Story
- AI Science Startup Periodic Labs Is in Deal Talks at About $7 Billion Valuation (2026-03-25)
- Lila Sciences valuation, funding & news
- Discovering new materials with AI has a winding road to VC returns
- Periodic Labs (2026-07-23)