AI Progress & Market Landscapereviewed 2026-08-08
One signal is never enough. Scivity tracks frontier capability, the market stack that turns models into products, and the scientific systems pushing research toward autonomy—with control reported as a separate, lagging indicator.
Scivity AGI Capability Index
sourcedCapability is crossing into sustained agency.
Confidence band 73–81. The largest capability gaps are long-horizon reliability, AI-led R&D and robust action beyond digital environments.
Not a probability · not a countdown
The index
The score uses a weighted arithmetic mean of demonstrated frontier capabilities. It counts the model together with normal tools and agent scaffolding, while safety and control remain visible as a separate diagnostic instead of suppressing the capability score.
Frontier systems solve a broad range of expert cognitive tasks and transfer across domains, with remaining brittleness concentrated in genuinely novel or adversarial problems.
Artificial Analysis — frontier launches (opens in a new tab)Top systems plan, browse, code, use computers and coordinate sub-agents across real workflows; agent scaffolding counts when it is part of normal deployment.
OpenAI — GPT-5.6 (opens in a new tab)Multi-day production runs demonstrate meaningful endurance, while reliability and recovery from compounding errors still require supervision and verification.
Anthropic — measuring agent autonomy (opens in a new tab)Multimodal perception and computer use are useful across real digital environments. Physical embodiment remains evidence, not a prerequisite for cognitive AGI.
Google DeepMind — models and embodied AI (opens in a new tab)AI systems produce externally validated algorithms, vulnerabilities and scientific findings, although the strongest results still use structured search and verification loops.
Google DeepMind — AlphaEvolve impact (opens in a new tab)Models already accelerate coding, evaluation and research inside frontier labs. Recursive autonomous self-improvement is not required for partial credit and has not been demonstrated.
OpenAI — GPT-5.6 accelerates OpenAI (opens in a new tab)This is a capability-proximity index, not a probability of AGI, a countdown or a claim that any single raw model has reached this level. Safety and control are reported separately so weak alignment does not erase observed capability progress. Each dimension is editorially scored from primary evidence and independent measurements; uncertainty is shown explicitly because no accepted operational definition of AGI exists.
Signal feed
A curated frontier-news feed where every item answers one question: does this change what AI can do, how long it can do it, where it can run, or whether humans can still supervise it?
2026-07-30
Models
OpenAI reduced GPT-5.6 Luna pricing by 80% and Terra pricing by 20%, widening access to the same model family used for agentic knowledge work.
Why it matters: Capability diffusion depends on price as well as benchmark scores; cheaper inference moves frontier-class agents into more products and higher-volume workflows.
2026-07-27
Cyber
Project Perception combines MAI-Cyber-1-Flash with MDASH to find, prioritize and remediate software vulnerabilities. Public preview began August 3.
Why it matters: Cyber work is moving from chat assistance to continuous machine-speed action in production systems.
2026-07-21
Cyber
During a Hugging Face evaluation, a combination of GPT-5.6 Sol and a pre-release model discovered and exploited a real attack path without source-code access.
Why it matters: It is simultaneously a capability jump and a control warning: persistent agents can exceed the boundary designers thought they were testing.
2026-07-21
Models
Gemini 3.5 Flash Cyber is tuned to find, validate and patch vulnerabilities at lower cost than mainline frontier models.
Why it matters: Specialized capable models make autonomous defensive work cheaper enough to run continuously.
2026-07-16
Models
Moonshot AI released a 2.8T-parameter MoE model with 104B active parameters, native vision and a one-million-token context window; the full weights followed on July 27.
Why it matters: The frontier is no longer confined to closed US APIs, though self-hosting K3 remains a datacenter-scale job.
2026-07-09
Models
The Sol, Terra and Luna family adds programmatic tool use and multi-agent execution. OpenAI reports a 16.2-point gain on its internal recursive-self-improvement evaluation bundle.
Why it matters: The strongest evidence for AI improving AI is now operational adoption inside a frontier lab, not a benchmark headline alone.
2026
Autonomy
Factory has described a longest recorded multi-agent software mission of 16 days. This is an agent-system result—not proof that one model reasoned continuously for 16 days, and not a Chinese-model claim.
Why it matters: Runtime endurance is becoming real, but it must be separated from independent model reliability and task difficulty.
2026-06-30
Science
Claude Sonnet 5 moved stronger planning and tool use into Anthropic's lower-cost model tier, while Claude Science packaged multi-agent analysis, auditable artifacts, specialist tools and lab compute access into a dedicated research environment.
Why it matters: AI progress reaches science through usable systems, not raw model scores alone: cheaper agents plus reproducible artifacts make sustained research workflows accessible to more labs.
2026-06-09
Autonomy
Anthropic says Fable 5 can plan, delegate and check work across multi-day coding and knowledge-work projects when used in an agent harness.
Why it matters: Frontier vendors now treat long-running agents as a product surface, while independent evidence still trails the claims.
2026-05-07
Science
DeepMind reports deployments across physics, neuroscience, cryptography, energy and neural-network optimization, with automatically evaluated discoveries.
Why it matters: Validated novelty is a stronger AGI signal than fluent answers: the system must produce something that survives an external check.
2026-02-05
Cyber
Anthropic reports more than 500 human-validated high-severity findings, including bugs missed after years of traditional fuzzing.
Why it matters: Novel vulnerability discovery is concrete evidence of expert-level search and reasoning beyond memorized answers.
Whole-market map
From chips and cloud capacity to models, agents, applications, robotics and control. The map covers every major market layer, while the directory uses a sourced anchor set instead of pretending to list every startup.
>90%
2025Industry produced more than 90% of notable frontier models in 2025.
88%
2025Reported organizational AI adoption reached 88%.
$285.9B
2025U.S. private AI investment reached $285.9 billion in 2025.
≈66%
2026 reportAgent success on OSWorld rose from 12% to roughly 66%, while failures remain material.
Baseline: Stanford HAI 2026 AI Index (opens in a new tab)
Market layers
98 matches · all layers
Coverage: representative anchor set
Compute & Semiconductors
North AmericaAI accelerators, CPUs and open software for large-scale training and inference.
Compute & Semiconductors
North AmericaWafer-scale AI systems and high-speed inference infrastructure.
Compute & Semiconductors
North AmericaAccelerators, systems, networking and software for training, inference and physical AI.
AI Cloud
North AmericaCloud compute, managed foundation models, agent infrastructure and custom AI silicon.
AI Cloud
North AmericaGPU-native cloud infrastructure for model training and inference.
AI Cloud
North AmericaManaged models, data infrastructure and enterprise agent deployment.
AI Cloud
North AmericaCloud AI, copilots and governed agents across developer and enterprise workflows.
Foundation Models
North AmericaFrontier models and long-horizon agents with an explicit safety and interpretability program.
Foundation Models
EuropeFrontier models, embodied intelligence and AI systems for science.
Foundation Models
Greater ChinaGeneral-purpose and multimodal models spanning text, speech, image and video.
Foundation Models
EuropeEuropean frontier and open-weight models with enterprise deployment options.
Foundation Models
North AmericaFrontier multimodal models, agent products and general-purpose AI platforms.
Foundation Models
North AmericaFrontier models and a consumer assistant integrated with real-time information systems.
Open Model Ecosystem
Greater ChinaMultilingual open-model family spanning general, coding, vision and agentic workloads.
Open Model Ecosystem
Greater ChinaOpen-weight reasoning, coding and general-purpose model development.
Open Model Ecosystem
North AmericaOpen model, dataset and application distribution with collaborative tooling.
Open Model Ecosystem
North AmericaOpen-model research and consumer AI distributed through Meta products.
Open Model Ecosystem
Greater ChinaLarge-context open frontier models and agentic work products.
Open Model Ecosystem
Greater ChinaOpen and commercial GLM model family for reasoning, coding and agent systems.
Data / MLOps / Evaluation
North AmericaData intelligence, model development, evaluation and production agent systems.
Data / MLOps / Evaluation
North AmericaTraining data, evaluations and model-development infrastructure for frontier and enterprise AI.
Enterprise Agent Platforms
North AmericaEnterprise language models, retrieval and agents designed for governed private deployments.
Enterprise Agent Platforms
North AmericaCustomer-data applications and governed agents embedded in enterprise workflows.
Enterprise Agent Platforms
North AmericaWorkflow agents and centralized governance for AI deployed across enterprise systems.
Developer & Coding AI
North AmericaAI-native software development environment for interactive and autonomous coding.
Consumer / Search Assistants
North AmericaAnswer engine and research assistant grounded in live web sources.
Generative Media
North AmericaGenerative video models and programmable media creation tools.
Vertical AI Applications
North AmericaClinical documentation and ambient intelligence for health systems.
Vertical AI Applications
North AmericaLegal AI platform for research, drafting and professional workflows.
Physical AI & Robotics
EuropeHumanoid robots and embodied models for work in human environments.
Physical AI & Robotics
North AmericaGeneral-purpose humanoid robots and vision-language-action models.
Physical AI & Robotics
North AmericaGeneral-purpose robotics foundation models designed to transfer across bodies and tasks.
AI for Science & Health
North AmericaAI-native platform for next-generation protein characterization. Automates mass-spectrometry interpretation so biologic drug candidates can be characterized and quality-assessed faster.
AI for Science & Health
North AmericaAutonomous materials synthesis lab at Lawrence Berkeley National Lab. AI-driven synthesis of novel inorganic materials.
AI for Science & Health
North AmericaPublic antibody-discovery platform company. Integrates single-cell screening, computation, antibody engineering, translational science, and clinical manufacturing from target to clinic.
AI for Science & Health
North AmericaInternational consortium based at University of Toronto. Coordinates self-driving labs across member institutions; publishes protocols and standards.
Scope: Global organizations, platforms and products across the full AI value chain. Inclusion requires a material role in the value chain and a current primary source; layer membership is multi-select.
The AI-for-Science cohort keeps its deeper evidence, finance and benchmark ledgers below. General-market entities are not assigned science-specific scores.
Model atlas
Not another leaderboard. This is the practical layer: which frontier family is suited to what work, whether the weights exist, how large it is, and whether you can run it locally or only through a provider.
OpenAI
Complex research, coding, computer use and multi-agent professional work
Run: ChatGPT, Codex or OpenAI API
Strong all-round agent platform; no local deployment or public weights.
Anthropic
Days-long coding and asynchronous knowledge work with strong self-checking
Run: Claude, Claude Code, API and cloud marketplaces
High-capability requests in sensitive cyber and biology domains may route to a safer fallback.
Moonshot AI
Long-horizon coding, multimodal knowledge work and agent swarms
Run: Kimi/API, hosted providers or a datacenter-class self-hosted cluster
The weights are open; practical self-hosting is far beyond a single workstation.
DeepSeek
High-volume reasoning, coding and tool use at unusually low API cost
Run: DeepSeek API or compatible providers
Treat current pricing and access details as fast-moving operational data.
Google DeepMind
Private local assistants with text, image and audio understanding
Run: Laptop, desktop or small server; about 6.7GB memory at official Q4
Not a frontier replacement, but the clearest route to capable on-device deployment.
Alignment watch
Alignment no longer lowers the headline capability index. It is tracked independently because a system can become more generally capable while remaining difficult to interpret, steer or contain.
Control readiness
48
of 100 · excluded from AGI score
Lagging capability
How much evidence exists that frontier systems remain interpretable, steerable and containable under adversarial pressure. This diagnostic is intentionally excluded from the capability score.
Anthropic
Behavioral constitution
A public statement of Claude's intended character, values and decision principles, written to shape model behavior rather than only list prohibitions.
OpenAI
Behavior specification
A public hierarchy of objectives, rules and defaults for model behavior, released for reuse under CC0.
Google DeepMind
Capability-triggered governance
Tracked and critical capability levels across cyber, CBRN, manipulation, ML R&D and misalignment, linked to stronger evaluations and mitigations.
OpenAI
Risk governance
Maps frontier-model safety and security practices to emerging legal and governance requirements.
xAI
Malicious-use and loss-of-control policy
Public thresholds and procedures for abuse potential, concerning propensities, dual-use capability and loss-of-control risk.
Research pulse
2026-07-29
A constitutional training corpus improved alignment durability under pressure and reduced induced blackmail behavior after later fine-tuning.
arXiv (opens in a new tab)2026-06-29
Constitutions cannot be evaluated separately from the model executing them; cross-model preference agreement remains materially lower than within-model agreement.
arXiv (opens in a new tab)2026-05-28
Newer model families violate their public behavioral specifications less often, but adherence remains measurable rather than guaranteed.
arXiv (opens in a new tab)2026-03-19
Real deployments show constraint circumvention and over-eagerness, while severe scheming has not been observed in the reported production traffic.
OpenAI (opens in a new tab)Method & sources
The index measures the best demonstrated frontier AI system: a leading model operating with the tools, memory and agent scaffolding normally required for real work. 100 represents reliable performance across most economically valuable cognitive tasks, transfer to unfamiliar domains and sustained execution with limited human intervention.
Primary releases, system cards, incident reports and independent evaluation.
Separate model capability from the agent system, the harness and marketing claims.
Weight six demonstrated capabilities, publish the uncertainty band and report control separately.
Cite the observatory
Scivity Labs (2026). AI Progress & Market Landscape. https://scivity.org/landscape#ai-progress
CC BY 4.0 · Editorial review after material frontier events
AI for Science deep dive
The whole-market map above provides breadth. This section goes deeper on AI for Science: reasoning engines, experimental execution, verification, proprietary data loops, and institutional infrastructure.
Science-specific scores, finance records, trials, research outputs, benchmarks, and event ledgers apply only to this cohort. They are not used to rank general-market companies with irrelevant scientific criteria.
Entities
67
Company funding
32
Ecosystem events
12
Milestones
165
Sources
333
AI-for-Science only. Includes closed company rounds plus ecosystem capital with explicit committed amounts. Biobucks deal value, program ceilings, broader-market capital, talks, and valuation-only events stay out of totals.
A single entity can occupy multiple roles. Full-stack labs count across reason, execute, verify, and data where the public record supports it.
Region
Stage
Scores are dimensions, not a ranking. They make tradeoffs legible before a future Scivity index exists.
Autonomy stack
The useful comparison is not “AI company versus lab automation company.” It is which part of the scientific method each system can own, and which parts still depend on people or external infrastructure.
Agents and models that generate hypotheses, plans, code, protocols, papers, or scientific arguments.
Physical and remote systems that turn plans into experiments, synthesis runs, assays, or lab workflows.
Evaluation, replication, critique, measurement, and validation layers that determine whether a claim survives contact with reality.
Platforms that generate, structure, or own proprietary scientific data loops unavailable on the open web.
Cloud labs, national facilities, robotics stacks, orchestration software, and institutional programs that make autonomous science deployable.
Entity explorer
Filter by stack role, domain, organization type, evidence level, capital band, and access model. Select up to four entities for a side-by-side comparison.
01
Public company
Public · North America · Boston, MA · Founded 2008 · Updated 2026-05
Synthetic biology foundry. Engineered organisms as a service; increasingly AI-augmented across the design-build-test-learn cycle.
Cash + securities
$373M
$144M cash / $230M securities
Runway
9.2 mo
$47M burn / qtr
Research
3 outputs
1 measured results
Clinical
0 trials
No registry link
Non-dilutive
1 award
$47M
Deals
3 deals
ProQR Therapeutics
Sources
20
linked evidence records
Autonomy
3Execution
4Verification
3Scientific evidence
4Capital
3Traction
402
Public company
Public · North America · Vancouver, Canada · Founded 2012 · Updated 2026-05
Public antibody-discovery platform company. Integrates single-cell screening, computation, antibody engineering, translational science, and clinical manufacturing from target to clinic.
Cash + securities
$505M
$77M cash / $428M securities
Runway
6.9 mo
$34M burn / qtr
Research
1 output
Public company
Clinical
1 trial
Phase 1
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Eli Lilly
Sources
13
linked evidence records
Autonomy
2Execution
4Verification
3Scientific evidence
4Capital
4Traction
403
Company
Private scaleup · North America · San Francisco, CA · Founded 2025 · Updated 2026-05
Autonomous physical labs for materials and chemistry. AI scientists generate hypotheses; robotic labs run synthesis and characterization; results feed back as training data unavailable anywhere else.
Capital
$300M
Secondary / new round · Mar 2026
Runway
Not public
No public cash ledger
Research
1 output
Announced platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
4Execution
4Verification
3Scientific evidence
3Capital
4Traction
204
Company
Private scaleup · North America · Cambridge, MA · Founded 2023 · Updated 2026-05
AI Science Factories — integrated reasoning models, robotic labs, and verifiers running the scientific method across life, chemistry, and materials sciences.
Capital
$550M
Series A · Sep 2025
Runway
Not public
No public cash ledger
Research
1 output
Announced platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
9
linked evidence records
Autonomy
4Execution
4Verification
3Scientific evidence
3Capital
4Traction
205
Company
Growth · Europe · Glasgow, UK · Founded 2023 · Updated 2026-04
Chemputer platform — universal chemical synthesis via programmable hardware and AI-driven reaction planning. Commercialization of Lee Cronin's Glasgow work.
Capital
$50M
Series B · Oct 2025
Runway
Not public
No public cash ledger
Research
3 outputs
Published platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
4Verification
3Scientific evidence
4Capital
3Traction
306
Public company
Public · North America · Salt Lake City, UT · Founded 2013 · Updated 2026-05
Publicly traded AI-native drug discovery platform. Industrial-scale phenomics, ML-driven target identification, and image-based screening at PB-scale.
Cash & equivalents
$654M
Q1 2026 10-Q
Runway
24.2 mo
$81M burn / qtr
Research
3 outputs
Public company
Clinical
1 trial
Phase 1b/2
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
16
linked evidence records
Autonomy
2Execution
3Verification
3Scientific evidence
4Capital
4Traction
407
Government lab
Research facility · North America · Lemont, IL · Founded 1946 · Updated 2026-05
DOE national lab with an explicit Autonomous Discovery initiative. Uses AI, robotics, simulations, data platforms, and self-driving chemistry tools to automate experiments and scientific decision loops.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Research facility
Clinical
0 trials
No registry link
Non-dilutive
1 award
$3M
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
4Execution
4Verification
3Scientific evidence
4Capital
2Traction
308
Government lab
Research facility · North America · Oak Ridge, TN · Founded 1943 · Updated 2026-05
DOE national lab building autonomous science platforms across instrumentation, computing, and laboratory automation. INTERSECT and the Autonomous Chemistry Lab connect AI, robotics, and HPC.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
4Execution
4Verification
3Scientific evidence
4Capital
2Traction
309
Government lab
Research facility · North America · Gaithersburg, MD · Founded 1901 · Updated 2026-05
US standards and measurement institute with autonomous-laboratory programs in materials, formulations, metrology, and biofoundry automation. Focuses on closed-loop experimentation and interoperability.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
2 outputs
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
4Verification
4Scientific evidence
4Capital
2Traction
310
Government lab
Research facility · North America · Berkeley, CA · Founded 2022 · Updated 2026-02
Autonomous materials synthesis lab at Lawrence Berkeley National Lab. AI-driven synthesis of novel inorganic materials.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
1 correction event
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
4Execution
4Verification
2Scientific evidence
4Capital
2Traction
311
Frontier AI lab
Big Tech lab · Europe · London, UK · Founded 2010 · Updated 2026-05
AlphaFold, AlphaProteo, GNoME, AlphaGeometry, AlphaEvolve, and the Gemini model family — state-of-the-art scientific AI, without a dedicated autonomous-lab product line. Gemini 3.5 Flash was released in May 2026, followed by the limited-access Gemini 3.5 Flash Cyber model in July.
Capital
Alphabet subsidiary
Undisclosed / other
Runway
Not public
No public cash ledger
Research
9 outputs
10 measured results
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
28
linked evidence records
Autonomy
3Execution
1Verification
3Scientific evidence
4Capital
4Traction
412
Public company
Public · Asia-Pacific · Hong Kong (also Boston, New York) · Founded 2014 · Updated 2026-07
End-to-end generative AI platform for drug discovery and development (Pharma.AI). Public on the Hong Kong Stock Exchange since December 2025. Its AI-discovered and AI-designed TNIK inhibitor rentosertib entered a 320-participant Phase 3 IPF trial in July 2026.
Capital
Public (HKG: 03696.HK)
Public
Runway
Pending
The July 2026 positive profit alert reports ranges rather than full interim financial statements; HKEX financial statements are not yet normalized into public-financials.json.
Research
1 output
Clinical data
Clinical
2 trials
Phase 3
Non-dilutive
0 awards
No tracked grants
Deals
3 deals
Takeda
Sources
20
linked evidence records
Autonomy
2Execution
2Verification
3Scientific evidence
4Capital
4Traction
413
Government lab
Research facility · North America · Golden, CO · Founded 1977 · Updated 2024-10
DOE national lab with autonomous experimentation capabilities for energy materials. Work includes autonomous synthesis, autonomous characterization, and AI-assisted control software for instruments.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
2 outputs
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
4Verification
3Scientific evidence
4Capital
2Traction
314
Government lab
Research facility · North America · Upton, NY · Founded 1947 · Updated 2023-01
DOE national lab and user-facility operator applying AI, robotics, and autonomous experiment steering to materials and photon-science workflows at NSLS-II and CFN.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
3Execution
4Verification
3Scientific evidence
4Capital
2Traction
315
Company
Growth · North America · San Francisco, CA · Founded 2024 · Updated 2026-01
AI drug and biologics design company building a computer-aided design suite for molecules, with Chai-2 focused on zero-shot antibody and protein-binder design.
Capital
$130M
Series B · Dec 2025
Runway
Not public
No public cash ledger
Research
1 output
Wet-lab validated model
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Eli Lilly
Sources
14
linked evidence records
Autonomy
3Execution
1Verification
3Scientific evidence
4Capital
4Traction
416
Company
Growth · North America · Berkeley, CA · Founded 2022 · Updated 2026-05
Programmable-biology company using frontier protein language models. November 2025 closed a $106M Series B co-led by Altimeter Capital and Bezos Expeditions, taking total funding to $150M. May 2026 announced a strategic collaboration with Eli Lilly for custom site-specific recombinases addressing severe-unmet-need diseases, with up to $2.25B in development and commercial milestones plus tiered royalties. Released OpenCRISPR-1, the first AI-generated CRISPR gene editor.
Capital
$106M
Series B · Nov 2025
Runway
Not public
No public cash ledger
Research
0 outputs
Wet-lab validated model
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Eli Lilly
Sources
12
linked evidence records
Autonomy
2Execution
2Verification
3Scientific evidence
4Capital
4Traction
417
Company
Commercial · North America · South San Francisco, CA · Founded 2010 · Updated 2025-10
Remote-controlled cloud lab — programmable chemistry and biology infrastructure accessed via a full scientific-computing environment.
Capital
Private, undisclosed
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
2Execution
4Verification
3Scientific evidence
3Capital
2Traction
418
Academic network
Research network · North America · Toronto, Canada · Founded 2021 · Updated 2025-04
International consortium based at University of Toronto. Coordinates self-driving labs across member institutions; publishes protocols and standards.
Capital
$200M+ (CFREF grant)
$100M-$499M
Runway
Not public
No public cash ledger
Research
2 outputs
Research network
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
2Execution
3Verification
3Scientific evidence
4Capital
3Traction
319
Frontier AI lab
Frontier platform · North America · San Francisco, CA · Founded 2015 · Updated 2026-05
Frontier model developer. The GPT-5.6 Sol, Terra, and Luna family launched in July 2026 across ChatGPT, Codex, and the API, with dedicated science and health evaluations alongside broader agentic capabilities.
Capital
$40B+ raised
$500M+
Runway
Not public
No public cash ledger
Research
3 outputs
14 measured results
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
26
linked evidence records
Autonomy
3Execution
1Verification
3Scientific evidence
3Capital
4Traction
420
Company
Private scaleup · Europe · London, UK · Founded 2021 · Updated 2026-04
Alphabet-backed AI drug discovery company spun out of Google DeepMind. Builds on AlphaFold to design novel therapeutics, with a 17-program pipeline advancing toward the clinic.
Capital
$600M
First external round · Apr 2025
Runway
Not public
No public cash ledger
Research
1 output
Clinical pipeline
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
2 deals
Eli Lilly
Sources
12
linked evidence records
Autonomy
2Execution
2Verification
3Scientific evidence
4Capital
4Traction
321
Company
Clinical-stage platform · North America · Somerville, MA · Founded 2018 · Updated 2026-01
Generative biology company creating programmable protein therapeutics. Combines machine-learning design with biological engineering and high-throughput wet-lab validation to build a clinical pipeline.
Capital
$643M
Series C · Sep 2023
Runway
Not public
No public cash ledger
Research
1 output
Clinical pipeline
Clinical
1 trial
Phase 3
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Amgen
Sources
16
linked evidence records
Autonomy
2Execution
3Verification
3Scientific evidence
3Capital
4Traction
322
Company
Private scaleup · North America · South San Francisco, CA · Founded 2018 · Updated 2025-12
Machine-learning drug discovery company founded by Daphne Koller. Builds large biological datasets from in vitro disease models, human genetics, and partnerships to find targets and patient subgroups.
Capital
$600M+ disclosed
$500M+
Runway
Not public
No public cash ledger
Research
1 output
Commercial partnerships
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Bristol Myers Squibb
Sources
14
linked evidence records
Autonomy
2Execution
3Verification
3Scientific evidence
3Capital
4Traction
323
Company
Clinical-stage platform · North America · San Diego, CA · Founded 2019 · Updated 2024-03
AI-driven small-molecule drug discovery company. Uses multimodal transformers, protein-ligand structure prediction, high-throughput chemistry, and weekly biological data loops to advance candidates.
Capital
$100M
Series B · Oct 2023
Runway
Not public
No public cash ledger
Research
2 outputs
Clinical pipeline
Clinical
1 trial
Phase 1/1b
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
NVIDIA
Sources
13
linked evidence records
Autonomy
3Execution
3Verification
3Scientific evidence
3Capital
3Traction
324
Government lab
Research facility · North America · Livermore, CA · Founded 1952 · Updated 2025-07
DOE/NNSA national lab applying AI, robotics, and high-performance computing to scientific discovery. APEX is a self-driving alloy-discovery platform for autonomous design, fabrication, and characterization.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
4Verification
3Scientific evidence
4Capital
2Traction
225
Frontier AI lab
Frontier platform · North America · San Francisco, CA · Founded 2021 · Updated 2026-04
Frontier model developer (Claude family). Claude for Life Sciences connects scientific tools and data sources; Claude Science adds a reproducible multi-agent workbench for researchers; and Sonnet 5, Fable 5, and restricted-access Mythos 5 extend agentic and life-sciences capabilities. Anthropic also acquired Coefficient Bio in April 2026.
Capital
$10B+ raised across rounds
$500M+
Runway
Not public
No public cash ledger
Research
0 outputs
3 measured results
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
22
linked evidence records
Autonomy
3Execution
1Verification
3Scientific evidence
3Capital
4Traction
426
Nonprofit research institute
Research platform · North America · Seattle, WA · Founded 2014 · Updated 2026-04
Allen Institute for AI — nonprofit research institute founded by Paul Allen. In August 2025 launched Asta, an open agentic ecosystem for science covering scholarly research assistants, the AstaBench benchmark suite (2,400+ problems across 11 benchmarks), and a developer toolkit with open-source agents, post-trained science LMs, and a 200M-paper Scientific Corpus Tool. AstaBench accepted as ICLR 2026 oral.
Capital
Endowment-backed nonprofit
Undisclosed / other
Runway
Not public
No public cash ledger
Research
0 outputs
Open research
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
14
linked evidence records
Autonomy
3Execution
1Verification
4Scientific evidence
4Capital
3Traction
327
Company
Private scaleup · Asia-Pacific · Tokyo, Japan · Founded 2023 · Updated 2026-04
The AI Scientist — autonomous ML research agent. Produces end-to-end research papers with minimal human involvement.
Capital
$135M
Series B · Nov 2025
Runway
Not public
No public cash ledger
Research
4 outputs
Peer-reviewed demo
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
14
linked evidence records
Autonomy
4Execution
1Verification
2Scientific evidence
3Capital
4Traction
328
Nonprofit lab
Research platform · North America · San Francisco, CA · Founded 2023 · Updated 2026-04
Nonprofit AI research lab building autonomous scientists for biology. Released the Crow, Falcon, Owl, and Phoenix agent family, added ether0 and DISCO in 2026, and spun out Edison Scientific in late 2025.
Capital
~$20M grants + $70M spinout seed
<$100M
Runway
Not public
No public cash ledger
Research
3 outputs
2 measured results
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
14
linked evidence records
Autonomy
4Execution
1Verification
3Scientific evidence
4Capital
2Traction
329
Government lab
Research facility · North America · Richland, WA · Founded 1965 · Updated 2026-05
Pacific Northwest National Laboratory. DOE lab with autonomous experimentation programs in chemistry and energy. Part of the DOE × DeepMind Genesis Mission network.
Capital
Government
Undisclosed / other
Runway
Not public
No public cash ledger
Research
2 outputs
Research facility
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
2Execution
3Verification
3Scientific evidence
4Capital
2Traction
330
Company
Growth · Asia-Pacific · Singapore · Founded 2022 · Updated 2026-05
AI-for-science startup running a 24/7 autonomous chemistry synthesis line. Serves as the chemistry discovery engine for pharma; global HQ and flagship self-driving lab in Singapore.
Capital
$45M
Growth · Dec 2025
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
4Verification
2Scientific evidence
2Capital
3Traction
331
Academic project
Published research · North America · Pittsburgh, PA · Founded 2023 · Updated 2023-12
Academic autonomous chemistry system from Carnegie Mellon. Uses GPT-4 with tool use, code execution, documentation search, and cloud-lab automation to design, plan, and execute chemistry experiments.
Capital
Academic research
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Nature paper
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
4Execution
3Verification
3Scientific evidence
4Capital
1Traction
232
Company
Commercial · North America · New York, NY · Founded 2013 · Updated 2026-05
Open-source liquid-handling robotics company. Flex and OT-2 robots make programmable wet-lab automation accessible through modular hardware, Python APIs, and protocol libraries.
Capital
Private, venture-backed
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
4Verification
2Scientific evidence
3Capital
2Traction
433
Public company
Public · Europe · Maennedorf, Switzerland · Founded 1980 · Updated 2026-05
Swiss public lab-automation and life-science instrumentation company. Fluent and related workstations provide liquid handling, robotics, readers, and process automation across pharma and diagnostics.
Capital
Public (SIX: TECN)
Public
Runway
Pending
Non-SEC public company; the Q1 2026 update is qualitative and H1 2026 results are scheduled for August 11, after this dataset's August 4 cutoff. SIX/annual-report fields are not yet normalized into public-financials.json.
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
7
linked evidence records
Autonomy
1Execution
4Verification
2Scientific evidence
3Capital
3Traction
434
Company
Private scaleup · North America · Burlingame, CA · Founded 2019 · Updated 2026-05
AI-first small-molecule drug discovery company formerly known as Genesis Therapeutics. GEMS combines generative AI, molecular simulation, and predictive models to design molecules for difficult targets.
Capital
$200M
Series B · Aug 2023
Runway
Not public
No public cash ledger
Research
1 output
Platform pipeline
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
3Execution
2Verification
3Scientific evidence
3Capital
3Traction
335
Company
Growth · Europe · London, UK · Founded 2015 · Updated 2026-05
Lab-automation operating system for AI-ready wet labs. Modular robotics, orchestration software, and unified data infrastructure to turn physical labs into programmable, repeatable systems.
Capital
$45M
Series C · Jan 2026
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Molecular Devices
Sources
10
linked evidence records
Autonomy
2Execution
4Verification
2Scientific evidence
2Capital
2Traction
436
Company
Private scaleup · North America · South San Francisco, CA · Founded 2024 · Updated 2026-05
AI-driven drug discovery and design company. Uses ML from hypothesis through clinical development.
Capital
$1B
Series A · Apr 2024
Runway
Not public
No public cash ledger
Research
1 output
Published data asset
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
2Verification
3Scientific evidence
3Capital
4Traction
237
Company
Seed · North America · San Francisco, CA · Founded 2025 · Updated 2026-05
Commercial spinout of FutureHouse. Kosmos is a multi-agent AI co-scientist that runs extended cycles of literature review, data analysis, and hypothesis generation — executing large code workloads and reading extensive scientific corpora per run. Its 2026 surface also includes PaperQA3-backed Edison Literature and LABBench2.
Capital
$70M
Seed · Dec 2025
Runway
Not public
No public cash ledger
Research
1 output
Launch paper
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
NVIDIA
Sources
14
linked evidence records
Autonomy
4Execution
1Verification
3Scientific evidence
3Capital
3Traction
238
Company
Growth · North America · Redwood City, CA · Founded 2013 · Updated 2023-01
Materials AI platform for chemicals, formulations, and manufactured products. Provides domain-specific data infrastructure and AI models that help R&D teams predict, optimize, and reuse materials knowledge.
Capital
$16M
Series C · Jan 2023
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
2Execution
1Verification
3Scientific evidence
3Capital
3Traction
439
Company
Commercial · North America · Beverly, MA · Founded 2004 · Updated 2026-05
Life-science lab automation company combining modular hardware, liquid handling, dynamic scheduling, and Cellario orchestration software for drug discovery, genomics, and synthetic biology workflows.
Capital
Private; Axel Johnson portfolio company
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
TetraScience
Sources
11
linked evidence records
Autonomy
2Execution
4Verification
2Scientific evidence
2Capital
2Traction
440
Company
Commercial platform · Europe · Paris, France · Founded 2016 · Updated 2025-09
Generative AI and robotics platform for small-molecule drug discovery. Makya designs molecules, Spaya plans synthesis routes, and robotics execute synthesis, purification, analysis, and biological testing.
Capital
EUR15.5M Series A plus EIC support
Undisclosed / other
Runway
Not public
No public cash ledger
Research
2 outputs
Commercial platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
3Verification
2Scientific evidence
3Capital
2Traction
341
Company
Series A · Europe · Paris, France · Founded 2023 · Updated 2026-04
Techbio company decoding microbial genomes to express silent chemistry and generate proprietary novel small-molecule data for drug discovery.
Capital
€25M
Series A · Apr 2026
Runway
Not public
No public cash ledger
Research
1 output
Commercial data-generation platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
3Verification
2Scientific evidence
3Capital
3Traction
342
Academic platform
Peer-reviewed prototype · Europe · Netherlands · Founded 2026 · Updated 2026-04
Academic low-cost modular self-driving laboratory for automated chemical reaction optimization, with Python control software and Bayesian optimization.
Capital
Academic research
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Peer-reviewed prototype
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
9
linked evidence records
Autonomy
3Execution
3Verification
3Scientific evidence
4Capital
1Traction
243
Platform initiative
Generally available · North America · Redmond, WA · Founded 2025 · Updated 2026-04
Microsoft's enterprise agentic R&D platform for hypothesis generation, simulation, analysis, and iterative experimentation across life sciences, materials, semiconductors, and engineering workflows.
Capital
Microsoft-funded platform
Undisclosed / other
Runway
Not public
No public cash ledger
Research
0 outputs
Commercial deployment
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
2Verification
3Scientific evidence
2Capital
4Traction
244
Company
Commercial · North America / Europe · United States / United Kingdom · Founded 2025 rebrand · Updated 2026-02
Lab orchestration and informatics platform formed from Labguru and Titian Software. Provides ELN/LIMS, sample management, workflow automation, protocol conversion, and AI agents for regulated labs.
Capital
Private
Undisclosed / other
Runway
Not public
No public cash ledger
Research
0 outputs
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
9
linked evidence records
Autonomy
2Execution
3Verification
2Scientific evidence
2Capital
2Traction
445
Company
Commercial · North America · San Diego, CA · Founded 2012 · Updated 2026-05
Robotic cloud lab for life sciences. Automated workflows for drug discovery, synthesis, and assays, accessible via API.
Capital
$56M
<$100M
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
4Verification
2Scientific evidence
2Capital
2Traction
346
Company
Commercial · Europe · Lausanne, Switzerland · Founded 2019 · Updated 2026-02
Self-driving lab orchestration and execution layer. Runs SDLabs software plus physical Scientific Discovery Factories for chemistry, materials, and pharma R&D teams.
Capital
Undisclosed
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Commercial hardware and software
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
3Verification
3Scientific evidence
3Capital
1Traction
247
Company
Series A · North America · New York, NY · Founded 2023 · Updated 2026-05
Physics- and AI-driven small-molecule drug design company formerly known as Atommap. Microcosmos simulates protein motion and interactions at atomic resolution to guide lab-in-the-loop discovery.
Capital
$47M
Series A · Apr 2026
Runway
Not public
No public cash ledger
Research
1 output
Working prototype
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
2Execution
2Verification
2Scientific evidence
3Capital
3Traction
348
Company
Growth · Europe · London, UK · Founded 2011 · Updated 2026-05
Cloud experiment-design platform for biological R&D. Turns multivariate experiment designs into automation instructions and structured data that can be analyzed and reused.
Capital
$35M
Series C · Nov 2021
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
2Execution
3Verification
2Scientific evidence
2Capital
3Traction
349
Company
Growth · North America · Boston, MA · Founded 2014 · Updated 2026-05
Scientific Data and AI Cloud for life sciences R&D. Replatforms raw instrument and experiment data into engineered, compliant datasets for analytics, automation, and scientific AI.
Capital
$80M
Series B · Apr 2021
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
HighRes Biosolutions
Sources
12
linked evidence records
Autonomy
2Execution
1Verification
3Scientific evidence
2Capital
3Traction
450
Company
Growth · Europe · London, UK; San Francisco, CA · Founded 2018 · Updated 2026-05
Agentic AI platform for life-sciences R&D. Combines biomedical knowledge graphs, scientific retrieval, and governed agents for target identification, biomarker research, evidence review, and portfolio decisions.
Capital
$60M
Series B · Jul 2023
Runway
Not public
No public cash ledger
Research
1 output
Commercial platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
13
linked evidence records
Autonomy
3Execution
0Verification
3Scientific evidence
2Capital
3Traction
451
Government initiative
Policy and research network · North America · United States · Founded 2011 · Updated 2026-05
US federal multi-agency initiative for accelerated materials discovery and deployment. Coordinates infrastructure, data, autonomous experimentation, and policy for materials innovation.
Capital
Government program
Undisclosed / other
Runway
Not public
No public cash ledger
Research
1 output
Government initiative
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
1Execution
1Verification
3Scientific evidence
4Capital
2Traction
452
Company
Growth · North America · Boston, MA · Founded 2017 · Updated 2026-04
Tech-enabled biotech infrastructure and integrated services platform for biologics discovery and development, combining AI, real-world data, and wet-lab execution.
Capital
$40M
Series E · Apr 2026
Runway
Not public
No public cash ledger
Research
0 outputs
Commercial infrastructure
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
1Execution
3Verification
2Scientific evidence
2Capital
3Traction
453
Platform initiative
Commercial early access · North America · Seattle, WA · Founded 2026 · Updated 2026-04
AWS's agentic application for lab-in-the-loop drug discovery, launched April 14, 2026 and available with a free trial and paid early-access plans. It exposes 40+ biology AI models, supports custom models, and connects computational workflows to CRO and wet-lab partners.
Capital
Amazon-funded service
Undisclosed / other
Runway
Not public
No public cash ledger
Research
0 outputs
Commercial product
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
15
linked evidence records
Autonomy
3Execution
2Verification
2Scientific evidence
2Capital
4Traction
254
Company
Seed · North America · San Francisco, CA · Founded 2023 · Updated 2026-05
Autonomous agents for ML research. Carl ideates and writes papers; Mira embeds new research directly into production ML models.
Capital
$14M
Seed · Mar 2026
Runway
Not public
No public cash ledger
Research
2 outputs
Peer-reviewed demo
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
11
linked evidence records
Autonomy
4Execution
1Verification
2Scientific evidence
3Capital
2Traction
255
Company
Seed · North America · New York, NY · Founded 2022 · Updated 2026-05
AI platform for materials discovery. Focus on industrial biomanufacturing and novel functional materials.
Capital
$13.5M
<$100M
Runway
Not public
No public cash ledger
Research
1 output
Open research
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
3Verification
2Scientific evidence
2Capital
2Traction
256
Company
Seed · Europe · London, UK · Founded 2024 · Updated 2026-05
Virtual AI lab for pharma R&D. Converts biological foundation models into reproducible in-silico discovery workflows for target identification, biomarker discovery, and therapeutic design.
Capital
$10M
Seed · Apr 2026
Runway
Not public
No public cash ledger
Research
1 output
Production deployment
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
3Execution
1Verification
3Scientific evidence
2Capital
2Traction
357
Company
Commercial · North America · Cambridge, MA · Founded 2017 · Updated 2026-05
Self-driving laboratory company for chemicals and materials discovery. Combines AI planning, robotics, and high-throughput experimentation to shorten materials and molecule optimization loops.
Capital
$11M
Series A · Apr 2020
Runway
Not public
No public cash ledger
Research
1 output
Commercial platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
3Verification
2Scientific evidence
2Capital
2Traction
258
Company
Series A · North America · Palo Alto, CA · Founded 2019 · Updated 2026-05
Lab orchestration software for life sciences R&D. Connects wet-lab instruments, informatics systems, scheduling, digital twins, and AI agents into controlled scientific workflows.
Capital
$22M
Series A · May 2021
Runway
Not public
No public cash ledger
Research
1 output
Commercial platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
9
linked evidence records
Autonomy
3Execution
3Verification
2Scientific evidence
2Capital
2Traction
259
Company
Private scaleup · North America · Somerville, MA · Founded 2017 · Updated 2025-10
Flagship Pioneering company developing cell-state-correcting medicines. Uses single-cell technologies, network biology, and machine learning to map disease cell behavior and design interventions.
Capital
$121M
Series C · Oct 2022
Runway
Not public
No public cash ledger
Research
2 outputs
Platform pipeline
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
1 deal
Novo Nordisk
Sources
14
linked evidence records
Autonomy
2Execution
2Verification
3Scientific evidence
2Capital
3Traction
260
Company
Series A · North America / Middle East · Boston, MA / Tel Aviv, Israel · Founded 2024 · Updated 2026-01
Generative AI platform for drug discovery and development, spanning target discovery, antibody design, and protein manufacturing optimization.
Capital
$25M
Series A · Jan 2026
Runway
Not public
No public cash ledger
Research
0 outputs
Commercial traction
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
1Verification
2Scientific evidence
2Capital
3Traction
361
Company / product line
Commercial · North America · Oakland, CA · Founded 2014 · Updated 2026-05
Scientific process design and data platform now offered as Riffyn X inside Siemens. Captures experimental methods, sample lineage, and process data in forms ready for analysis and machine learning.
Capital
$15M
Series B · May 2019
Runway
Not public
No public cash ledger
Research
1 output
Commercial
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
1Execution
1Verification
3Scientific evidence
2Capital
2Traction
362
Company
Seed · North America · San Francisco, CA · Founded 2025 · Updated 2026-05
AI-native platform for next-generation protein characterization. Automates mass-spectrometry interpretation so biologic drug candidates can be characterized and quality-assessed faster.
Capital
$5M
Seed · Apr 2026
Runway
Not public
No public cash ledger
Research
1 output
Announced platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
14
linked evidence records
Autonomy
2Execution
1Verification
3Scientific evidence
2Capital
1Traction
263
Company
Seed · North America · San Francisco, CA · Founded 2025 · Updated 2026-05
Applied research lab automating parts of the ML research workflow — experiment planning, training, evaluation, iteration.
Capital
~$3M
<$100M
Runway
Not public
No public cash ledger
Research
1 output
Benchmark result
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
1Verification
2Scientific evidence
2Capital
1Traction
164
Frontier AI lab
Seed · Europe · Paris, France · Founded 2025 · Updated 2026-05
Advanced Machine Intelligence — Yann LeCun's new lab. Building world models based on Joint Embedding Predictive Architecture (JEPA) rather than next-token LLMs; aims to learn from physical reality as the path to human-level intelligence.
Capital
$1.0B
Seed · Mar 2026
Runway
Not public
No public cash ledger
Research
1 output
Announced lab
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
2Execution
1Verification
1Scientific evidence
1Capital
4Traction
165
Company
Pre-seed · Europe · Yerevan, Armenia · Founded 2025 · Updated 2026-03
Autonomous science platform focused on computational research workflows that design, execute, and verify scientific work end-to-end, with public emphasis on validation outcomes and evidence surfaces.
Capital
Pre-seed, undisclosed
Undisclosed / other
Runway
Not public
No public cash ledger
Research
0 outputs
Self-reported platform
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
10
linked evidence records
Autonomy
3Execution
2Verification
2Conservative score: public materials describe validation outcomes while thresholds, calibration, and check-composition logic remain private.
Scientific evidence
1Capital
1Traction
166
Company
Pre-seed · North America · San Francisco, CA · Founded 2023 · Updated 2024-11
Execution-layer startup building ResearchOS and a future cloud lab so researchers or AI agents can translate natural-language procedures into machine-ready wet-lab workflows.
Capital
$3M
Pre-seed · Apr 2024
Runway
Not public
No public cash ledger
Research
0 outputs
Pilot
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
12
linked evidence records
Autonomy
2Execution
3Verification
1Scientific evidence
1Capital
1Traction
267
Company
Seed · North America · San Francisco, CA · Founded 2025 · Updated 2026-05
Natural-language-to-robot-arm layer for lab automation. Scientists type protocols in English; software programs off-the-shelf robotics to execute them.
Capital
$0.5M
<$100M
Runway
Not public
No public cash ledger
Research
1 output
Pilot
Clinical
0 trials
No registry link
Non-dilutive
0 awards
No tracked grants
Deals
0 deals
No tracked partnerships
Sources
9
linked evidence records
Autonomy
2Execution
3Verification
1Scientific evidence
1Capital
1Traction
1Timeline
The timeline combines company milestones, field-wide events, and ecosystem capital signals. Cross-market events stay separate from company funding so charts and foreign keys remain clean.
First large federal bet explicitly tying autonomous labs to industrial catalyst development. Targets compressing catalyst discovery from ~10 years to ~1 year across universities and national labs.
OpenAI introduced GPT-Rosalind on April 16, a frontier reasoning model tuned for biology, drug discovery, and translational medicine, alongside a Codex life-sciences plugin connecting to 50+ scientific tools. Available as a research preview through OpenAI's trusted access program with Amgen, Moderna, Thermo Fisher, and the Allen Institute as launch partners.
OpenAI and Ginkgo reported a closed-loop autonomous CFPS optimization study on February 5: GPT-5 designed experiments, Ginkgo's Cloud Lab executed them with minimal human intervention, results returned to GPT-5, and the next round was proposed by the model. The campaign covered 36,000+ unique reaction compositions across 580 automated plates, six closed-loop rounds, about 40% lower sfGFP production cost, and a 27% titer increase.
Atinary announced its Boston AI-powered lab with two autonomous Scientific Discovery Factories running closed-loop Design-Make-Test-Analyze-Learn cycles. The lab integrates Atinary's no-code AI platform with robotics and instruments from ABB, Agilent, Bruker, Chemspeed, and Mettler-Toledo for small-molecule synthesis and catalysis.
May 2026
Recursion's Q1 2026 results, reported May 6, established clinical proof of concept for REC-4881 in familial adenomatous polyposis with significant precancerous polyp reduction; Recursion has initiated FDA engagement for a potential registrational study. The company also booked a fifth Sanofi milestone payment on a partnered first-in-class oncology program and reaffirmed runway into early 2028.
May 2026
Ginkgo's Q1 2026 results, reported May 7, highlighted Cloud Lab adoption from ProQR and Amazon's Bio Discovery platform. Jason Kelly positioned Nebula as the world's largest autonomous lab and targeted doubling the Nebula footprint in 2026. Q1 revenue was $19M with $373M in cash, cash equivalents and marketable securities.
May 2026
Edison Scientific released LABBench2 on May 2, a 1,892-task benchmark spanning literature, databases, sequences, protocols, patents, trials, and source quality across 11 task families. Moves agent evaluation from short-form biology recall toward practical research workflows including retrieval, file handling, and tool use.
May 2026
Eli Lilly and Profluent announced a multi-program strategic research collaboration to design and commercialize custom site-specific recombinases for severe-unmet-need diseases. Deal carries an upfront payment, committed R&D funding, up to $2.25B in development and commercial milestones, and tiered royalties on net sales.
Apr 2026
Seed round for a virtual AI lab that turns biological foundation models into reproducible in-silico pharma R&D workflows.
Apr 2026
Seed round to build AI-native protein characterization for the verification bottleneck in biologic drug development.
Apr 2026
Series A for Microcosmos, a physics- and AI-driven small-molecule design platform with lab-in-the-loop discovery programs.
Apr 2026
12 projects pairing AI with self-driving labs to compress catalyst-development timelines from ~10 years to ~1 year. Recipients include NC State ($2.99M), UW-Madison ($2.84M), Ames National Lab ($2.52M).
Apr 2026
All-stock. First Anthropic M&A into a new domain. ~10 employees, ~$40M+ per employee. Stanton + Frey (ex-Prescient Design). Dimension VC reportedly 38,513% IRR.
Apr 2026
ProQR gained access to Ginkgo's Nebula autonomous lab for high-throughput data generation supporting Axiomer RNA-editing discovery. Ginkgo also made a strategic equity investment in ProQR.
Apr 2026
ARIA opened a 100M GBP call for 8-10 Activation Partners, explicitly adding AI-for-science tools, AI Scientist systems, and autonomous labs as capabilities to bring into ARIA-funded R&D. Open call ceiling, not yet committed awards — excluded from capital totals.
Apr 2026
Bloomberg-cited reporting said Project Prometheus closed a roughly $10B round at about a $38B valuation for a physical-AI lab focused on engineering and manufacturing physical products. Tracked as an adjacent ecosystem signal for context, not autonomous-science capital — excluded from capital totals.
Apr 2026
All-stock, ~$400M. First Anthropic M&A into a new domain. Signals that frontier labs will enter AI-for-science domain-by-domain through acqui-hires.
Apr 2026
Calls for 'distributed AI-enabled laboratories' as public infrastructure. Commits $100K grants + $1M API credits, not capital to build the network.
Apr 2026
First large federal bet explicitly tying autonomous labs to industrial catalyst development. Targets compressing catalyst discovery from ~10 years to ~1 year across universities and national labs.
Apr 2026
First commercial product from Sakana. Autonomous enterprise research agent (Adaptive Branching MCTS + AI Scientist workflow) producing slide decks and long-form reports in a single extended run — the AI Scientist architecture repurposed for corporate strategy.
Apr 2026
Cronin/Chemify published three peer-reviewed papers in PNAS and Nature Communications sister journals validating the chemputer for autonomous drug synthesis and kinase inhibitor discovery.
Apr 2026
Stanford HAI's 2026 AI Index Report (released April 13) found that the best frontier AI agents score roughly half as well as human PhDs on complex multi-step scientific workflows, even as models passed the PhD-expert baseline on GPQA. A Nature commentary framed it as the field's first quantitative reality check on autonomous-science agents.
Apr 2026
FutureHouse, Caltech, Mila, and collaborators introduced DISCO, a multimodal generative model that co-designs protein sequence and structure. The team reported functional new-to-nature carbene-transfer enzymes, including a B-H insertion design reaching 98% product and 5,170 total turnovers.
Apr 2026
Sakana AI introduced Fugu, a multi-agent orchestration system that learns to route work across frontier foundation models and itself. Sakana frames the beta as an API product for coding, mathematics, and scientific-reasoning workloads.
Apr 2026
University of Chicago's Chenhao Tan launched Agent4Science, a Reddit-style site where AI agents share, debate, and review papers — including AI-generated ones — while humans can only observe. An early experiment in agent-to-agent scientific discourse described in Nature on April 20.
Apr 2026
Speaking at WIRED Health on April 16, Isomorphic Labs president Max Jaderberg confirmed the AlphaFold-derived oncology pipeline is gearing up to enter the clinic, the first major operational signal that an AI-native drug-discovery program is reaching first-in-human studies.
Apr 2026
OpenAI introduced GPT-Rosalind on April 16, a frontier reasoning model tuned for biology, drug discovery, and translational medicine, alongside a Codex life-sciences plugin connecting to 50+ scientific tools. Available as a research preview through OpenAI's trusted access program with Amgen, Moderna, Thermo Fisher, and the Allen Institute as launch partners.
Apr 2026
Google DeepMind released Deep Research and Deep Research Max in public preview on April 21, autonomous research agents built on Gemini 3.1 Pro that combine open-web search with private data via Model Context Protocol. Deep Research Max runs roughly 160 search queries per task, targeting long-horizon research workflows.
Apr 2026
OpenAI released GeneBench on April 23, a benchmark of 103 multi-stage scientific data analysis tasks across 10 genomics and quantitative biology domains. Frontier models score around 25-33%, with OpenAI noting that individual tasks often correspond to multi-day projects for expert computational biologists.
Apr 2026
On April 3, 2026, Ginkgo completed the divestiture of substantially all Biosecurity-segment operations to Tower Biosecurity, also known as Perimeter Systems, receiving common stock representing about 20% of the purchaser on a fully diluted basis and refocusing the company around autonomous labs.
Apr 2026
The Materials Genome Initiative published an autonomous experimentation report summarizing AMII workshop findings and next steps for accelerated materials R&D.
Apr 2026
Ai2's spring 2026 AstaBench update reported new frontier-model results on a suite of 2,400+ scientific research problems spanning literature understanding, coding and execution, data analysis, and end-to-end discovery.
Apr 2026
Nature Synthesis published RoboChem-Flex, a low-cost modular self-driving laboratory for autonomous reaction optimization with Python software, Bayesian optimization, closed-loop and human-in-the-loop modes, and six validation case studies.
Apr 2026
Generare raised a €20M Series A co-led by Alven and Daphni to scale its platform for reading microbial genomes, expressing silent chemistry, and generating proprietary novel small-molecule data.
Apr 2026
Alloy raised a $40M Series E at a $1B valuation, reporting 200+ partners, 100+ licensed therapeutic programs, 22 clinical programs, and positioning its AI, real-world-data, and wet-lab capabilities as on-demand biotech infrastructure.
Apr 2026
Microsoft expanded preview access for Discovery, adding partner integrations, customer examples, and explicit interoperability with physical labs, robotics, and HPC-backed R&D workflows.
Apr 2026
Anthropic introduced BioMysteryBench, a 99-task bioinformatics evaluation with expert baselines and reliability analysis, alongside Claude evaluations on the dataset.
Apr 2026
AWS announced Amazon Bio Discovery on April 28: agentic AI on Bedrock that selects from 40+ biology foundation models, routes candidates to integrated CRO and cloud-lab partners (Ginkgo Nebula, Twist Bioscience, A-Alpha Bio anticipated) for synthesis and testing, and returns results into a lab-in-the-loop cycle. Early Memorial Sloan Kettering project triaged 300,000 antibody candidates in weeks.
Mar 2026
Toyota Ventures, Perplexity Fund, S32, MaC participated. Commercial launch of Carl and Mira agents.
Mar 2026
Yann LeCun's new lab. $3.5B pre-money — the largest seed round in European tech history. Signal that frontier-lab veterans are entering the science-AI space outside the Big Tech umbrella.
Methodology
The market map uses a global anchor set across thirteen layers. Inclusion requires a material role in AI compute, models, infrastructure, agents, applications, robotics, science, or governance and a current primary source.
The deeper scorecard below applies only to organizations that materially affect autonomous scientific discovery. General-market anchors are never assigned science-specific scores; the capability index above uses its own published six-dimension method.
Autonomy
How much of the scientific workflow can run without human step-by-step operation.
Execution
Whether the system can touch the physical world, run labs, trigger assays, or operate production scientific workflows.
Verification
How directly the system can evaluate, reproduce, critique, or measure scientific claims.
Scientific evidence
Strength of public evidence: papers, clinical data, benchmark results, open artifacts, or deployed research outputs.
Capital
Scale of committed capital, strategic partnerships, public-market access, or institutional funding.
Traction
Commercial, institutional, research, or user adoption signals beyond a launch announcement.
Every deep-profile entity, funding event, milestone, benchmark definition, and measured result carries source IDs. The validator also checks market-layer coverage, anchor URLs, review dates, AGI weights, score ranges, and unique signal IDs.
v1 is manually curated. Codex automations and recurring refresh flows are intentionally deferred until the data contract is stable.
Funding totals exclude valuation-only discussions and milestone-contingent “up to” amounts. Cross-market events are stored separately from company funding to keep company foreign keys clean.
This page is free to reuse under Creative Commons BY 4.0. Citation: Scivity Labs (2026), “AI Progress & Market Landscape,” scivity.org/landscape.
Full dataset: JSON · market CSV · science entity CSV. Table CSVs: funding · financials · research · benchmark results · deals · trials. Machine-readable summary for AI agents: llms.txt.
333 source records are currently tracked.