Arc Institute

What Arc Institute does

Arc Institute is an independent nonprofit research organization founded in 2021 and headquartered in Palo Alto, California, that operates at the intersection of artificial intelligence and biology. The institute is structured around two flagship initiatives — the Virtual Cell Initiative and the Alzheimer's Disease Initiative — supported by 9 Core Laboratories, 5 Technology Centers (Genome Engineering, Multiomics, Cellular Models, Mammalian Models, and Computational), and over 300 scientists ('Arconauts'). Arc's operational model is to provide multi-year, no-strings-attached funding to researchers in partnership with Stanford University, UCSF, and UC Berkeley, enabling long-horizon, high-risk work that traditional grant mechanisms undersupport.

The institute has built a portfolio of AI-native research outputs released as open-source, including Evo 2 (a biological foundation model trained on 9.3 trillion nucleotides from 128,000+ genomes), STATE (a virtual cell model trained on 267M+ cells predicting perturbation responses), Proto (a high-level programming language for generative biology), MULTI-evolve (a protein engineering framework), PerturbSpace (spatial CRISPR screening), and Germinal (generative antibody design). Underlying these models is the Arc Virtual Cell Atlas — a curated data asset combining scBaseCount (500M+ AI-agent-curated single cells) and Tahoe-100M (100M drug perturbation profiles). Outputs are distributed through GitHub, HuggingFace, and the Arc Institute website, with Evo 2 integrated into NVIDIA's BioNeMo framework. Validation is anchored in publications across Nature, Science, and Cell.

Arc Institute is funded entirely through philanthropic donations and grants rather than commercial revenue. Aggregate disclosed commitments exceed $1.6 billion, including the Audacious Project's $1.03 billion raise from 35 donor families, Biohub's $500 million Virtual Biology Initiative, the OpenAI Foundation's $100 million Alzheimer's grant, and founding donations from a coalition including Vitalik Buterin, the Collison brothers, Dustin Moskovitz, Elad Gil, and the Ron Conway family. The institute has no paying customers in the commercial sense; its 'customer base' is the global scientific community, which it engages through academic publications, open-source releases, the annual Virtual Cell Challenge competition (5,000+ registrants from 114 countries), social media, and the AIxBio Fellows Program.

Name

Arc Institute

Legal name

Arc Institute

Website

https://arcinstitute.org

Company type

Private

Founded year

2021

Operating status

Operating

Headcount range

251–500 employees

Short description

Arc Institute is a Palo Alto-based nonprofit research organization founded in 2021 that combines AI and experimental biology through 300+ scientists, 9 Core Labs, and 5 Technology Centers, serving the global scientific community via open-source AI models, datasets, and tools.

Ownership category

akta.pro rank

Product category

Biomedical Research

NAICS

Research and Development in Biotechnology (except Nanobiotechnology) (541714), Research and Development in the Physical, Engineering, and Life Sciences (54171)

SIC

Services-Commercial Physical & Biological Research (8731)

akta.pro primary industry

AI/ML Platforms for Drug Discovery & Experiment Optimization (HLAGAJAO)

akta.pro secondary industries

Digital & AI-Enabled Diagnostics (algorithmic interpretation, decision support tied to assays) (HLAAALAO), AI/ML Engineering & Model Development Services (BPAEAHAF)

Keywords

• Biomedical research institute
• AI biology tools
• Genomic foundation models
• Virtual cell modeling
• Open source research

Headquarters

HQ city

Palo Alto

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Services

Cost components

Personnel, Technology or R&D, Operations, Infrastructure, Others

Revenue model

1. Donations and Grants: Arc Institute is a nonprofit research organization funded entirely through donations and grants from individual philanthropists, foundations, and partnerships. No commercial revenue model.

Go-to-market motion1 record

Distribution channels2 records

Marketing channels6 records

Core offering

Arc Institute conducts fundamental biomedical research and develops open-source AI models, datasets, and experimental tools to accelerate understanding of complex diseases such as cancer, neurodegeneration, and immune dysfunction. Its core deliverables include the Evo 2 genomic foundation model, the STATE virtual cell model, Proto (a programming language for generative biology), the Arc Virtual Cell Atlas (over 600 million curated single cells), and specialized platforms for genome engineering and multiomics. All outputs are released at no cost to the global scientific community through open-source channels.

Product overview

Arc Institute is a nonprofit research organization offering a comprehensive ecosystem of AI models, tools, datasets, and programs for biological research. The portfolio centers on Evo 2 (a genomic foundation model trained on 9 trillion DNA base pairs), Proto (a programming language for generative biology), and STATE (a virtual cell model predicting cellular perturbation responses). Supporting tools include PerturbSpace for spatial CRISPR screens, MULTI-evolve for protein engineering, Germinal for antibody design, and BioReason-Pro for protein function reasoning. The data foundation includes the Arc Virtual Cell Atlas with scBaseCount (500M+ curated single cells) and Tahoe-100M (100M drug perturbation profiles). Arc also hosts the annual Virtual Cell Challenge competition and runs the AIxBio Fellows Program for undergraduate researchers. All models and code are released as open-source through GitHub and HuggingFace for noncommercial use.

Differentiator

Problem solved

Functional benefit

Brands

• Virtual Cell Initiative (VCI): Initiative to build an accurate ML model that predicts how diverse cell types and states respond to drugs and genetic perturbations
• Alzheimer's Disease Initiative (ADI)
• Virtual Cell Challenge
• Arc Virtual Cell Atlas
• STATE
• Evo 2
• Proto

Products and services

• Evo 2 A biological foundation model trained on 9 trillion DNA base pairs from over 128,000 genomes spanning all domains of life, with a 1 million token context window for predicting pathogenic mutations and designing genome-scale DNA sequences, available as open-source.
• Proto A high-level programming language framework for generative biology that integrates AI models for protein design, RNA engineering, and gene regulation, enabling complex multi-modal biological design through sequences, generators, constraints, and optimizers.
• STATE Virtual Cell Model Arc's first-generation virtual cell model combining State Embedding (SE) and State Transition (ST) modules to predict how cells respond to chemical, genetic, and cytokine perturbations, trained on 167 million observational cells and 100 million+ perturbed cells.
• PerturbSpace A spatial profiling method combining CRISPR perturbations with antibody-based tissue hashing for in vivo Perturb-seq, mapping tissue location alongside transcriptome and CRISPR guide identity at neighborhood-level resolution.
• Arc Virtual Cell Atlas A growing collection of curated open single-cell datasets for virtual cell model training, encompassing over 600 million cells from observational and perturbational experiments.
• MULTI-evolve A rapid evolution framework using protein language models with epistatic modeling to predict and construct hyperactive multi-mutant proteins, achieving up to 256-fold improvements on multiple protein engineering targets.
• Germinal A generative antibody design pipeline that designs antibodies against specific epitopes with nanomolar binding affinities while requiring only low-n experimental testing.
• Bridge Recombinases A programmable, RNA-guided DNA recombinase system that enables precise insertion, inversion, and excision of DNA segments through a universal mechanism, representing a programmable editing approach beyond CRISPR, as published in Nature.
• Stack A foundation model trained on 149 million uniformly preprocessed human single cells, leveraging tabular attention to generate contextual cell representations for in-context learning of single-cell biology.
• scBaseCount An AI agent-curated single-cell RNA-seq database that automatically discovers, annotates, and reprocesses public single-cell data from over 500 million cells spanning 21 organisms and 72 tissues with standardized preprocessing.
• Tahoe-100M A single-cell dataset containing 100 million cells from approximately 60,000 drug perturbation experiments mapping responses of 50 cancer models to more than 1,100 drug treatments.
• BioReason-Pro A multimodal reasoning large language model for protein function prediction that integrates protein embeddings with biological context to generate structured reasoning traces about protein mechanisms and druggability.
• IDLI Iteratively Defined Lengths of Inaccessibility — a computational method that maps single-molecule nucleosome structure at genome scale, classifying nucleosomes into 14 distinct structural states to reveal regulated chromatin accessibility patterns.
• COURIER A cell monitoring system enabling cells to secrete RNA that encodes information about their behavior, supporting non-destructive longitudinal genomics to study dynamic biological processes in living systems.
• Virtual Cell Challenge An annual machine learning competition with a $100,000 grand prize for building models that predict how gene expression shifts in response to perturbations, supported by benchmark datasets and a live leaderboard.

Quantifiable outcome

• Evo 2 achieves over 90% accuracy in distinguishing benign from pathogenic mutations in BRCA1 benchmark tests
• +3 more outcomes

Named customers1 record

Segments2 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration3 records

AI capability24 records

Feature6 records

Partnerships

14 partnerships are on record, tiered flagship and core.

• NVIDIAflagshipTechnology or Integration · 2 May 2026NVIDIA serves as a key technology partner providing accelerated computing infrastructure (H100 GPUs) and integrated Evo 2 into their BioNeMo framework for the Virtual Biology Initiative.
• Biohub (Mark Zuckerberg and Priscilla Chan)flagshipStrategic or Co-development Partner · 29 April 2026Biohub is backing the Virtual Biology Initiative with $500 million ($100M for global data collection, $400M for internal data generation and technologies). Arc Institute is a key participating institution alongside Allen Institute, Broad Institute, Human Cell Atlas, Human Protein Atlas, and Wellcome Sanger Institute.
• Tahoe Therapeutics (formerly Vevo Therapeutics)coreStrategic or Co-development Partner · 25 February 2025Tahoe generated the Tahoe-100M dataset (100 million cells from ~60,000 drug perturbation experiments) for Arc's Virtual Cell Atlas using their Mosaic platform. This collaboration provides computational-ready single-cell data for training virtual cell models.
• NVIDIA (Virtual Cell Challenge Sponsor)coreTechnology or Integration · 1 January 2025NVIDIA served as a key sponsor of Arc's Virtual Cell Challenge competition, providing support for the $100,000 grand prize and advancing the virtual cell modeling field.
• 10x Genomics (Virtual Cell Challenge Sponsor)coreTechnology or Integration · 1 January 202510x Genomics served as a sponsor of Arc's Virtual Cell Challenge competition, supporting innovation in single-cell perturbation prediction modeling.
• Ultima Genomics (Virtual Cell Challenge Sponsor)coreTechnology or Integration · 1 January 2025Ultima Genomics served as a sponsor of Arc's Virtual Cell Challenge competition, supporting the advancement of virtual cell modeling.
• Allen InstitutecoreStrategic or Co-development Partner · 1 January 2025Partner in the Virtual Biology Initiative, collaborating with Arc Institute, Broad Institute, and other leading institutions to build the data foundation for AI-accelerated biology.
• Broad InstitutecoreStrategic or Co-development Partner · 1 January 2025Partner in the Virtual Biology Initiative, contributing to the global effort to create the open data foundation for AI-accelerated biology.
• Wellcome Sanger InstitutecoreStrategic or Co-development Partner · 1 January 2025Partner in the Virtual Biology Initiative, participating in the coordinated worldwide data generation effort for building AI models of the human cell.
• Human Cell AtlascoreStrategic or Co-development Partner · 1 January 2025Partner in the Virtual Biology Initiative, contributing single-cell reference data and expertise for AI-accelerated biology.
• Human Protein AtlascoreStrategic or Co-development Partner · 1 January 2025Partner in the Virtual Biology Initiative, contributing protein reference data for the global effort to create the data foundation for AI models of human biology.
• Stanford UniversitycoreStrategic or Co-development Partner · 1 January 2021Founding university partner with dual-appointment faculty opportunities. Core Investigators may hold faculty positions at Stanford, and Stanford graduate students can pursue PhD studies at Arc labs.
• University of California, San Francisco (UCSF)coreStrategic or Co-development Partner · 1 January 2021Founding university partner with dual-appointment faculty opportunities. Core Investigators may hold faculty positions at UCSF, and UCSF students can pursue PhD studies at Arc labs.
• University of California, Berkeley (UC Berkeley)coreStrategic or Co-development Partner · 1 January 2021Founding university partner with dual-appointment faculty opportunities. Core Investigators may hold faculty positions at UC Berkeley, and Berkeley students can pursue PhD studies at Arc labs.

Scale indicators8 records

Recent moves6 records

Expansion highlights6 records

Market position

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/arcinstitute
• GitHubgithub.com/ArcInstitute
• YouTubeyoutube.com/@arcinstitute
• LinkedInlinkedin.com/company/arc-institute-org

Revenue estimate

Valuation estimate

Number of profiles

Profiles12 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Arc Institute