State by Arc Institute: status and timeline

State

State is a virtual cell model that predicts how cells respond to genetic, chemical and biological perturbations. Trained on nearly 170 million observed cells and 100+ million perturbed cells across 70 human cell types.

DeployStage 4 of 5

Released June 2025; trained on 270M+ cells for perturbation prediction.

Updated Jun 2025Checked 10 Oct0 updates this week

Milestones

Next · Validation against experimental data

State model releasedJun 2025Complete.
Validation against experimental dataTarget 2026Current milestone.

Most important updates

Arc Institute releases State, a virtual cell model trained on 270 million cellsJun 2025

Upcoming

2026Validation against experimental data (next)

Current obstacles

Capturing rare cell statesCells spend most time in common states; rare responses to stress or treatment are underrepresented in training data.
Cross-batch variabilityData from different labs and experiments have systematic biases; State must bridge them accurately.

Physics limits

Cell complexityA cell has thousands of genes interacting in networks; State captures patterns but not all causal mechanisms.
Single-cell resolutionState predicts at the individual cell level, but real tissues have spatial structure and cell-cell interactions that matter.

How it works

3 parts

Embedding

State Embedding (SE)

Maps cell gene expression to a smooth vector space, allowing comparison across cell types and batches.

Transition

State Transition (ST)

A bidirectional transformer that models how chemical, genetic or cytokine perturbations shift a cell's gene expression in that space.

Applications

Drug discovery and cell therapy

State lets researchers predict how cells will respond to treatments before wet-lab experiments, speeding drug discovery.

Spec sheet

Spec	State
Total training data	270M+ cells (170M observed, 100M+ perturbed)^R (reported)
Human cell types covered	70 cell types^R (reported)
Architecture modules	State Embedding (SE) + State Transition (ST)^R (reported)

^R reported by the company

Update log

1 update

Jun 2025

Major: Arc Institute releases State, a virtual cell model trained on 270 million cellsPressResearch

About Arc Institute

The team behind State

Arc Institute

Evo DNA AI, virtual cell, bridge editing

Nonprofit research institute in Palo Alto founded in 2021 by Patrick Hsu, Silvana Konermann and Patrick Collison, with Stanford, Berkeley and UCSF. It funds scientists without grant applications and builds shared tools like the Evo DNA models.

Evo 2 (Feb 2025, with NVIDIA) learned from DNA of all known life and can predict effects of mutations and write new sequences.
In 2024 Arc researchers showed 'bridge RNAs' that guide enzymes to insert, delete or flip DNA.
The second Virtual Cell Challenge launched on 20 Aug 2026.

arcinstitute.org (opens arcinstitute.org)Company profile →

Founded: 20215 yrs
Headquarters: United States
Status: Private
Valuation: private
Works in: BiotechSynthetic biologyAI
Coverage: 4 programs · 6 updateslatest 20 Aug 2026checked 25 Sep
Partners: NVIDIA (opens nvidia.com)Stanford University (opens stanford.edu)
People: Patrick Hsu (opens arcinstitute.org)Co-founder and core investigatorSilvana Konermann (opens arcinstitute.org)Co-founder and executive directorPatrick Collison (opens en.wikipedia.org)Co-founder

Spec

State

Total training data

270M+ cells (170M observed, 100M+ perturbed)^R (reported)

Human cell types covered

70 cell types^R (reported)

Architecture modules

State Embedding (SE) + State Transition (ST)^R (reported)