One To Watch

PerturbAI

A technology-driven biotechnology company developing AI-enabled drug discovery systems based on large-scale in vivo perturbation data. Perturb AI focuses on mapping causal relationships between genes, cells and disease to train predictive biological models.

Company Overview

PerturbAI is building an integrated platform that combines experimental biology and artificial intelligence to accelerate drug discovery. Its core strategy is to generate large-scale datasets linking genetic perturbations to cellular and disease outcomes, and use these data to train predictive models of biology.

The company’s approach is based on the idea that understanding causal gene function in living systems—rather than relying solely on observational data—can significantly improve target identification and therapeutic development.

Perturb AI operates as a platform-centric biotech, with its value derived from data generation, model training and downstream application to drug discovery rather than a traditional disclosed clinical pipeline.

Headquarters and Global Presence

headquarters not broadly disclosed publicly
operates as a research-focused organization with platform development and potential partnerships

The company’s model is inherently global, as its datasets and AI models are designed to support drug discovery across multiple disease areas.

Founding and History

recently established AI-native biotechnology company
built around advances in CRISPR-based perturbation biology and machine learning

The company emerged as part of a broader wave of AI-driven drug discovery platforms aiming to move beyond correlation-based models toward causal biological understanding.

Therapy Areas and Focus

PerturbAI is not organized around specific therapeutic areas.

Its platform is applicable across:

oncology
neurological disorders
immunology
other complex diseases driven by genetic and cellular networks

The company’s focus is on enabling target discovery rather than advancing disease-specific pipelines.

Technology Platforms and Modalities

The company’s platform integrates experimental perturbation biology with AI.

Key components include:

in vivo CRISPR-based gene perturbation at scale
generation of organism-level “perturbation atlases” linking genes to biological outcomes
AI models trained on causal datasets
prediction of how genes and drugs reshape cellular systems

This approach differs from traditional AI drug discovery models by emphasizing causal data generation rather than relying solely on observational or historical datasets.

Key Platform Outputs and Programs

Perturbation atlas generation

Modality: in vivo CRISPR perturbation datasets
Scope: organism-scale mapping of gene function
Application: linking genes to cells, circuits and disease states

AI biological models

Modality: machine learning models trained on perturbation data
Application: prediction of gene function and drug effects
Role: enabling target discovery and therapeutic hypothesis generation

Discovery applications

Modality: AI-guided drug discovery
Indication focus: broad, depending on partner or internal priorities
Status: platform development

The company’s outputs are designed to feed into downstream drug discovery rather than represent discrete clinical assets.

Strategic Partnerships

PerturbAI is expected to operate through partnerships typical of platform companies.

Key elements include:

potential collaborations with pharmaceutical companies for target discovery
integration of platform outputs into drug development pipelines
partnerships leveraging data and AI models

Specific partnerships have not been widely disclosed.

FAQ Section

The central strategic issue is whether a single-platform, dermatology-focused pipeline can achieve sufficient clinical and commercial validation. The company’s value is concentrated in the success of QTORIN and its ability to expand across multiple indications.

Perturbation biology involves actively modifying genes or pathways and observing the resulting effects. This provides causal insight into biological systems, which is more informative than observational data alone.

The company combines large-scale in vivo CRISPR experiments with AI model training. This creates datasets that capture cause-and-effect relationships between genes and disease, rather than correlations.

Many AI drug discovery companies rely on existing datasets. Perturb AI generates its own experimental data at scale, which may improve model accuracy and biological relevance.

A perturbation atlas is a dataset mapping how changes to specific genes affect cells, tissues and disease processes across an organism. These atlases are used to train predictive models.

PerturbAI is an early-stage platform company focused on building its data and AI infrastructure rather than advancing clinical-stage drug candidates.

Key issues include: