Skip to content
ContentLora

    Tip: press / anywhere to search.

    concept

    Self-driving labs (autonomous laboratories)

    Also known as autonomous laboratory, autonomous lab, AI scientist, closed-loop discovery, A-Lab

    Self-driving labs combine AI planning with robotic experiments so that a system can propose, make and test new molecules or materials with little human input.[1] The best-known example, the A-Lab, drew a high-profile critique, and since 2025 well-funded start-ups and the US Genesis Mission have bet on the approach.[2][3][4]

    Editor reviewedUpdated AI for scienceArtificial intelligenceScience
    Key facts

    A self-driving lab closes the loop between AI and experiment. The AI proposes what to make, robots make and measure it, and the results train the next round of proposals.[1] The idea matters because prediction has outpaced testing: GNoME alone proposed 2.2 million candidate crystals.[5]

    The A-Lab

    The A-Lab, described in Nature in November 2023, synthesises inorganic powders. It uses computations, literature data, machine learning and active learning to plan and interpret experiments carried out by robots.[1] Its targets came from large-scale stability data from the Materials Project and Google DeepMind.[6] The current version of the paper reports 36 compounds realised from 57 targets over 17 days of continuous operation.[7]

    The controversy

    In March 2024, chemists writing in PRX Energy re-examined the A-Lab’s products and concluded that no new materials had been discovered. They argued that about two-thirds of the claimed successes were likely known, compositionally disordered versions of the predicted ordered compounds.[2] They also said automated Rietveld analysis of X-ray diffraction data was not yet reliable.[8] MIT Technology Review later summarised the dispute as questions over novelty and whether automated analysis met experimental standards.[9]

    In January 2026 the authors published a correction. It says the original novelty claims were meant to indicate materials new to the prediction platform, not necessarily new to science. A manual re-analysis of the diffraction patterns confirmed 36 of the 40 reported successes, with four inconclusive, and one compound was removed because it had been in the training data.[10] Nature says this re-analysis was peer-reviewed after publication.[11] The original 41-from-58 figure still appears in Nature’s one-line summary of the article, which explains why older coverage quotes it.[12][10]

    Language-model lab agents

    A second strand uses general-purpose language models as the planner. Coscientist, described in Nature in December 2023, used OpenAI‘s GPT-4 to design, plan and carry out chemistry experiments, including optimising palladium-catalysed cross-coupling reactions.[13]

    The 2025–26 wave

    Investors have since backed autonomous discovery at scale:

    • Lila Sciences, unveiled by Flagship Pioneering in March 2025 with $200 million in committed seed capital, aims to build a “scientific superintelligence” platform and fully autonomous labs for life, chemical and materials sciences.[14] It closed a $350 million Series A in October 2025, with NVIDIA‘s venture arm among the investors, and won three Phase I Genesis Mission awards in July 2026.[15][16] In September 2026 it reported that its AI-directed lab screened 2,942 green-hydrogen catalysts in three months; the result is company-reported.[17]
    • Periodic Labs emerged in September 2025 with a $300 million seed round; its first goal is better superconductors.[3]
    • The US Genesis Mission, launched in November 2025, asks the Department of Energy to build a closed-loop AI platform that automates experiment design.[4] The order also told DOE to review national laboratory capacity for robotic labs able to run AI-directed experiments.[18] See genesis-mission.

    AI agents are also driving lab work through outside services. In Anthropic’s August 2026 protein-design campaign, Claude produced designs that two external companies, Adaptyv Bio and Twist Bioscience, made and tested.[19]

    What to watch

    The test for self-driving labs is the same as for any discovery claim: independently verified new results. As of December 2025, MIT Technology Review reported that AI materials discovery had not yet produced a convincing big win.[20] The debate is laid out in this course’s materials analysis page.

    Questions readers ask

    What does a self-driving lab actually do?

    It uses computation, literature data and machine learning to plan experiments, runs them with robots, interprets the results and chooses the next experiment.[1]

    Did the A-Lab discover new materials?

    Disputed. A 2024 PRX Energy critique concluded no new materials were discovered. The authors' January 2026 correction said "novel" meant new to the prediction platform, not necessarily new to science, and the current paper reports 36 compounds realised from 57 targets.[7][2][10]

    Who is building self-driving labs?

    Examples include start-ups Periodic Labs and Lila Sciences, and the US Department of Energy under the Genesis Mission.[3][14][4]

    Sources

    Each numbered claim is a statement we checked against the sources listed with it. Status shows how well established it is.

    1. [1]

      The A-Lab combines computations, literature data, machine learning and active learning to plan and interpret experiments carried out by robots. confirmedas of 2023-11-29

    2. [2]

      A March 2024 PRX Energy perspective concluded that no new materials had been discovered in the A-Lab work and that about two-thirds of claimed successes were likely known disordered versions of predicted compounds. disputedas of 2024-03-07

    3. [3]

      Periodic Labs emerged from stealth in September 2025 with a $300 million seed round, with a first goal of inventing better superconductors. confirmedas of 2025-09-30

    4. [4]

      The Genesis Mission tasks the Department of Energy with integrating its supercomputers and datasets into a closed-loop AI platform intended to automate experiment design, accelerate simulations and generate predictive models. confirmedas of 2025-11-24

    5. [5]

      Google DeepMind's GNoME work, published in Nature in November 2023, used graph networks to identify 2.2 million crystal structures, of which it counted 381,000 as newly discovered stable materials. confirmedas of 2023-11-29

    6. [6]

      The A-Lab's synthesis targets were identified using large-scale ab initio phase-stability data from the Materials Project and Google DeepMind. confirmedas of 2026-01-19

    7. [7]

      The A-Lab, an autonomous lab for solid-state synthesis described in Nature in November 2023, realized 36 compounds from 57 targets over 17 days of continuous operation, according to the current version of the paper. confirmedas of 2026-10-10

    8. [8]

      The PRX Energy critics argued that automated Rietveld analysis of powder X-ray diffraction data is not yet reliable. confirmedas of 2024-03-07

    9. [9]

      MIT Technology Review reported that critics questioned the novelty of A-Lab's products and the quality of its automated analysis. reportedas of 2025-12-15

    10. [10]

      A January 2026 author correction to the A-Lab paper said its novelty claims meant materials new to the prediction platform, not necessarily new to science; a manual re-analysis confirmed 36 of the 40 reported successes, with four inconclusive, and one compound was removed because it had been in the training data. confirmedas of 2026-01-19

    11. [11]

      Nature says the A-Lab diffraction re-analysis behind the 2026 correction was peer-reviewed after publication. confirmedas of 2026-01-19

    12. [12]

      The A-Lab article's one-line summary on Nature's site describes 41 novel compounds discovered and synthesised from 58 targets in 17 days, more than its current abstract reports. confirmedas of 2026-10-10

    13. [13]

      Coscientist, described in Nature in December 2023, is a GPT-4-driven system that autonomously designs, plans and performs complex chemistry experiments, including successful optimisation of palladium-catalysed cross-coupling reactions. confirmedas of 2023-12-20

    14. [14]

      Flagship Pioneering unveiled Lila Sciences in March 2025 with $200 million in committed seed capital to build a scientific superintelligence platform and fully autonomous labs for life, chemical and materials sciences. confirmedas of 2025-03-10

    15. [15]

      In October 2025 Lila Sciences closed a $350 million Series A, bringing its total funding to $550 million, with NVIDIA's venture arm NVentures among the investors. confirmedas of 2025-10-10

    16. [16]

      Lila Sciences said in July 2026 that three of its proposals won Phase I Genesis Mission awards, each with a national laboratory or university partner. confirmedas of 2026-07-22

    17. [17]

      Lila Sciences reported in September 2026 that its AI-directed lab proposed, synthesised and screened 2,942 catalysts for green hydrogen in three months and identified six high-performing material families; the result is company-reported. reportedas of 2026-09-25

    18. [18]

      The order required DOE, within 240 days, to review national laboratory capabilities for robotic laboratories able to carry out AI-directed experimentation and manufacturing. confirmedas of 2025-11-24

    19. [19]

      In August 2026 Anthropic reported that Claude models, using open-source protein design tools, designed binders against 14 of 15 targets, with designs produced and tested by Adaptyv Bio and Twist Bioscience. reportedas of 2026-08-18

    20. [20]

      MIT Technology Review reported in December 2025 that AI materials discovery had not yet produced a convincing big win or a new miracle material. reportedas of 2025-12-15

    Revision history (2)
    1. Page created.
    2. Added the January 2026 A-Lab author correction, Coscientist, Lila Sciences' Series A, catalyst screen and Genesis awards, and the Genesis order's robotic-lab review.

    Created Oct 10, 2026. Last reviewed by an editor on Oct 10, 2026. Next scheduled review: Jan 10, 2027.

    Cite this page

    "Self-driving labs (autonomous laboratories)." ContentLora, updated Oct 10, 2026. https://contentlora.com/wiki/self-driving-labs

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.

    Keep exploring

    Connections
    GNoME Uses predictions fromThe Materials Proje… Targets drawn from data inOpenAI Example system built on a…Genesis Mission Part of the agenda of
    ☰ Contents