Valid Therapeutics | Predictive human biology at scale

Valid in, Valid out.

Every model is a proxy. Our models begin with the right biology.

1

Drugs fail because models don’t reflect human biology.

Today, efficacy often isn’t tested until months or years into development, with programs instead relying on simple, scalable models to identify targets and hits.

2

Others chase compute. We fix the bottleneck of biological truth.

Our human stem cell-derived systems recreate disease biology with unprecedented fidelity. Scaled production of human-relevant models and automated, high-throughput perturbation screening generate the data AI needs to predict what will work in humans.

3

We build the most predictive experimental and AI models at scale.

The hypotheses that emerge are the ones most likely to matter in humans.

VALID is where biological truth becomes predictive intelligence.

The human origin

Each model is anchored in biological truth to capture regional specificity of disease and reproduction of complex human microenvironments.

VALID starts where disease begins, inside real human systems. Our human iPSC-derived co-culture disease models replicate feedback loops and cellular architecture of organs and tissues, from the midbrain to muscle.

Diagram showing differentiation of IPSC cells into brain, muscle, adipose, and liver tissues connected to respective human body regions.

The predictive sweet spot

Every model trades fidelity for scale.

VALID operates in the narrow region where human

and

intersect, using hiPSC-derived co-culture and reporter systems to preserve disease-relevant complexity while remaining experimentally scalable.

Graph showing scalability versus predictive validity of scientific models, with simplistic models featuring cell lines having high scalability but low validity, Valid co-culture models with reporter strategy in the middle, and complex models like organoids and assembloids with higher predictive validity but lower scalability, depicted alongside a human silhouette.

TechBio V2.0

VALID has created the experimental engine for predictive human disease modeling. We are able to test existing hypotheses experimentally and generate new, human-relevant hypotheses that traditional models cannot surface.

Our platform integrates human co-culture models, automated perturbation, and multimodal measurement to convert phenotypic biology into AI-ready data. Imaging, RNA-seq, and proteomics are captured from the start, at scale.

Flowchart showing co-culture models labeled 'Valid' leading to cellular environment and perturbations, which then lead to phenotypes and further to imaging, transcriptomics, proteomics, and AI.

The old funnel tests efficacy last.

Visual funnel diagram of the drug discovery pipeline. Large pools of targets and leads narrow through discovery and optimisation stages into preclinical efficacy and clinical studies, with the VALID platform shown expanding the funnel by improving lead quality and validation.

We Test it first.

Diagram showing interconnected gears labeled Better Biology and Better Data driving Better AI, which leads to Better Hits, Better Leads, Better Candidates, and Better Drugs.

Programs in motion

VALID’s discovery engine is expanding across multiple human systems from neuronal co-cultures for neurodegeneration to peripheral models for metabolic disease. Each dataset compounds knowledge and strengthens the next, fueling a continuous cycle of predictive discovery.

Silhouette of a human body showing brain, heart, and liver with labeled microscopic images for brain (neurodegeneration, neuropsychiatric & metabolic), aging & muscle preservation, and liver health.

Where most systems break, VALID executes.

Human iPSC systems are fragile, complex, and hard to scale. Many have tried. Most have failed. VALID has built the infrastructure to industrialize experimental biology—creating a reproducible, data-generating engine for modern drug discovery.