WillItSurvive
United States · United States

Periodic Labs

Startup pairing AI models with robotic labs to discover new materials.

Losing its protection
Assessed 29 July 2026 · Evidence as of 29 July 2026
The structural protection is eroding.
On the recordSince 29 July 2026 · Checkable from end of 2027 · Every call on the record Checkpoint: Confirmed named customer contract or independently verified material discovery
The call

Periodic Labs has famous founders and a huge valuation but has not yet shown a discovery, a paying customer, or proof its data edge beats DeepMind to the finish.

Periodic Labs uses robots and AI to discover new materials faster than traditional methods. The main risk is that this is still mostly a bet on future success: the company has raised enormous amounts of money but has not yet discovered a material good enough to sell to a customer, which is the only real proof the idea works. Google DeepMind and Microsoft have similar AI tools and massive resources, so Periodic's advantage depends entirely on whether its robot labs and data can stay ahead of bigger competitors in the race to find useful new materials first.

Why
01
AI pressure: Google DeepMind's GNoME-style discovery models. DeepMind already predicts new materials at scale; adding its own robotic synthesis loop would erase Periodic's data edge. Could AI do the core job? Only partly.
02
Bigger players: Google DeepMind, Microsoft Research, and materials giants like BASF or TSMC R&D. Deep-pocketed labs and manufacturers could build their own closed-loop robotic synthesis, copying the approach without needing Periodic. Could a giant copy it? Yes, easily.
03
What protects it. Proprietary experimental data from its own robotic labs, but only if it piles up before rivals close their own loop.
What would change this call
Periodic Labs announces a named customer contract with disclosed revenue by end of 2027. Would show the data moat is converting to real commercial traction, not just narrative.
Periodic Labs publishes an independently verified novel material discovery within 18-24 months. Would validate the proprietary experimental data moat as real rather than theoretical.
DeepMind, Lila Sciences, or another rival announces a working closed-loop autonomous synthesis program. Would show the incumbent or peer has closed Periodic's head-start window.
Key numbers
  • $300M seed closed at ~$1B valuation (2025-10)
  • $500-750M raise reported at $7.5B valuation (2026-05)
  • No confirmed revenue; roughly 30 employees (2026-06)
Named risks
  • Zero confirmed material discovery or named customer despite reported $7.5B valuation as of mid-2026. Severity: high
  • DeepMind could close the prediction-to-synthesis loop it already started with GNoME, erasing Periodic's head start. Severity: high
  • Multiple well-funded rivals, Lila Sciences and Radical AI, pursue the identical strategy in the same 12 to 18 month window. Severity: medium
  • Latest growth round terms reported mainly via secondary-market data provider, not fully confirmed by tier-1 press as of mid-2026. Severity: medium
Rivals, as read in July 2026
RivalDirectionEvidence
Google DeepMindcontestingDeepMind's 2023 GNoME predicted 2.2M candidates but never closed the synthesis loop; still capable of doing so
Lila SciencescontestingRaised $350M Series A at over $1.3B valuation in October 2025, same closed-loop AI-plus-robotics strategy
Radical AIlosing toWon a confirmed AFWERX Direct-to-Phase II Air Force contract for hypersonic-alloy discovery in August 2025; Periodic has no confirmed named contract
Questions people ask
Is Periodic Labs profitable in 2026?
No. A May 2026 industry commentary explicitly states Periodic Labs has no revenue and no commercial products as of that date, despite the company reportedly seeking a raise at a $7 to $7.5 billion valuation the same month.
What is Periodic Labs worth in 2026?
Periodic Labs raised a $300 million seed at roughly $1 billion pre-money valuation around September to October 2025, then was in talks for a growth round near $7 billion by March 2026, reportedly reaching $7.5 billion valuation discussions by May 2026.
Who founded Periodic Labs?
Periodic Labs was founded in May 2025 by Liam Fedus, a ChatGPT co-creator and former OpenAI VP of Research for post-training, and Ekin Dogus Cubuk, a former Google DeepMind scientist who led DeepMind's GNoME materials-discovery project.
Who is Periodic Labs' biggest competitor?
Google DeepMind is the closest incumbent threat since co-founder Cubuk led DeepMind's 2023 GNoME project predicting 2.2 million candidate crystal structures there; direct startup rivals Lila Sciences and Radical AI are racing the identical closed-loop AI-plus-robotics strategy.
Has Periodic Labs discovered a new material or superconductor?
No. As of the most recent reporting in July 2026, no public product launch, published material discovery, or superconductor breakthrough has been confirmed for Periodic Labs, despite superconductors being its stated near-term north star since September 2025.
Did OpenAI invest in Periodic Labs?
No. OpenAI was reportedly expected to participate in Periodic Labs' $300 million seed round around September to October 2025 but ultimately did not invest, per TechCrunch reporting dated 2025-10-20.
Timeline
  • 2026-06Reported growth round reaches $7.5B valuation discussions per press coverage
  • 2026-05Report cites talks for roughly $500M raise at $7.5B valuation
  • 2026-03Bloomberg reports deal talks at about $7 billion valuation
  • 2026-01Headcount estimated around 30 to 33 employees per directory sources
  • 2025-10$300M seed round publicized, led by Andreessen Horowitz
  • 2025-09a16z announces investment; company states superconductor research focus
  • 2025-08LA Times reports staffers set for $1 billion valuation in a16z-led round
  • 2025-05Periodic Labs founded by Liam Fedus and Ekin Dogus Cubuk
Revision history
29 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