The Manifesto.

AI is moving beyond reading papers, analyzing datasets, and designing molecules. It is beginning to interact directly with experiments:

decide what to test → run the experiment → observe the result → learn → choose what to do next

This transition is happening now.

Eli Lilly and NVIDIA are investing up to $1B into an AI co-innovation lab built around continuous learning between agentic wet labs and computation. [1] Genentech describes lab-in-the-loop as the foundation of its R&D strategy. [2][3] GSK already operates an end-to-end automated lab-in-the-loop platform for large-molecule discovery. [4] Lila Sciences is building entire AI Science Factories around models that reason, experiment, and learn. [5]

A new generation of infrastructure companies is emerging around the same shift. C5R is building research facilities where frontier models can execute experiments and learn from reality. [6] Substrate Bio is building a network of autonomous wet labs for biological training data. [7] Worldset plans modular autonomous biology labs designed to generate complete experimental trajectories for frontier models.

And the first results are already here.

OpenAI connected GPT-5 to Ginkgo Bioworks’ automated laboratory. Across more than 36,000 reactions, the system established a new low-cost benchmark for cell-free protein synthesis and reduced protein-production cost by 40%. [8]

Anthropic has now built its own biology laboratory. Claude searched more than 200,000 reverse transcriptases and identified a previously uncharacterized enzyme system associated with CRISPR-like DNA repeats, which Anthropic then investigated experimentally. [9]

The feedback loop between AI and biology is closing.

Laboratory automation is waiting for its intelligence upgrade

Laboratory automation today is similar to humanoid robotics: the physical infrastructure exists, but without sufficiently capable intelligence it remains limited to narrow, pre-programmed tasks.

Liquid handlers, robotic workcells, plate readers, bioreactors, and automated analytical systems already exist across thousands of laboratories.

Most still execute workflows humans designed and programmed in advance.

The opportunity is to turn that automation from infrastructure that executes experiments into infrastructure that learns which experiments to run.

ProviGen provides that intelligence upgrade.

We make existing laboratories adaptive

Rather than building inhouse laboratory infrastructure from scratch, we upgrade the infrastructure that already exists.

Biofoundries, CROs, CDMOs, biopharma companies, and automation providers have already invested billions into physical experimental capacity.

That infrastructure is fragmented, frequently underutilized, and rarely learns across experimental cycles.

ProviGen turns this infrastructure into a distributed network for closed-loop experimentation.

This means smaller laboratories can gain access to predictive capabilities that would otherwise require building massive AI-native infrastructure themselves.

The hardware is already there. We make it adaptive.

Integration is becoming cheaper

Laboratory heterogeneity historically made this type of software difficult to deploy.

That constraint is weakening.

Modern laboratory automation increasingly exposes APIs or machine-readable interfaces.

Schedulers and orchestration platforms abstract much of the underlying hardware complexity, allowing ProviGen to integrate at the workflow layer rather than writing direct drivers for every instrument.

At the same time, AI coding agents materially reduce the engineering cost of connecting heterogeneous software systems.

Mission

The instruments are in place. The models keep getting better. What is missing is the connection between scientific intent and the infrastructure that carries it out.

Scientists should be able to state what they want to learn, and the lab should immediately run it, collect the evidence and improve.

That is how biotechnology will enter an era of widely accessible cures, regenerative medicine and choice over lifespan.

Together with the brightest minds of our time, we want to push that frontier.
We will give people freedom over their own biology.