The promise of a self-driving lab is a simple one: replace the slow, manual, and often irreproducible workflows of a life-sciences bench with a system that runs itself. For a field where a single experiment can take weeks and a misplaced decimal can invalidate months of work, the appeal is profound. Foundational Bio, a New York-based startup, is building precisely that, an AI-native, fully automated laboratory designed to increase the speed and reliability of discovery [Foundational Bio, retrieved 2024]. The company, co-founded by Tomaz Berisa and Jahan-Yar Parsa, has secured backing from deeptech investors SOSV and Boost VC, placing its bet on a future where the physical act of experimentation is increasingly driven by software.
The Bet on Automation
Foundational Bio's core thesis is that the next leap in biotech productivity will come from integrating automation not just at the level of individual instruments, but across the entire experimental lifecycle. This moves beyond robotic liquid handlers to a system where AI agents design experiments, orchestrate robotic hardware, analyze results, and then plan the next iteration. The concept, often called a "cloud biofoundry," is gaining academic and commercial traction as a way to compress discovery timelines [ScienceDirect, 2025]. For Foundational Bio, the goal is to build a turnkey platform that can be deployed for a range of life-science tasks, from basic research to early-stage drug discovery, handling the tedious and error-prone steps that currently occupy skilled technicians.
The competitive landscape for automated discovery is crowded with well-funded players, but approaches vary. Foundational Bio appears to be positioning itself as a pure-play infrastructure provider, distinct from companies that use automation primarily for their own internal pipelines.
- Formation Bio operates as a tech-driven pharma company, using AI and data to develop drugs in-house [Formation Bio, Unknown].
- Recursion Pharmaceuticals has built a massive automated phenomics platform to fuel its drug discovery efforts, a model centered on proprietary data generation [Crunchbase, Unknown].
- Blank Bio, a Y Combinator alum, is an applied AI research lab building foundation models for biology, focusing on the computational layer rather than physical automation.
Foundational Bio's wedge, then, may be selling the automated lab itself as a service or platform to other research organizations, aiming to own the hardware and software stack that enables faster iteration for its customers.
An Honest Counterfactual
The ambition is significant, and so are the hurdles. Building a truly reliable self-driving lab is a profound integration challenge, requiring flawless communication between AI planning systems, robotic actuators, and analytical sensors. The regulatory path for discoveries made within such a system is also untested; while the lab itself may not need FDA clearance, any therapeutic or diagnostic output certainly will, and regulators will scrutinize the reproducibility and data integrity of automated workflows. Furthermore, the company's public footprint is currently minimal, with no detailed technical specifications, named pilot customers, or disclosed funding amounts beyond its investor roster. Success will depend on moving from a compelling concept to a validated, operational system that can demonstrate clear advantages over traditional, and increasingly sophisticated, manual methods.
The ultimate test for any platform in this space is its impact on patient outcomes. Foundational Bio is targeting the broad and fundamental disease state of slow discovery. The patient population is, indirectly, anyone waiting for a new therapeutic or diagnostic,a wait measured in years and billions of dollars. The standard of care today is a fragmented, manual process. A biologist or chemist develops a hypothesis, designs an experiment, and then spends days or weeks executing it by hand at the bench, pipetting solutions, incubating samples, and manually recording results in a lab notebook. This process is not only slow but inherently variable, with reproducibility crises plaguing published research. Foundational Bio's vision is to replace that manual cadence with a closed-loop, automated system, aiming to turn the slow, artisanal craft of experimentation into a reliable, scalable engineering discipline.
Sources
- [Foundational Bio, retrieved 2024] Foundational Bio | https://foundational.bio/
- [ScienceDirect, 2025] Cloud biofoundries and multi-AI agents advance self-driving labs via collaboration
- [Formation Bio, Unknown] Formation Bio | About Us | https://www.formation.bio/about-us
- [Crunchbase, Unknown] Recursion Pharmaceuticals - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/recursion-pharmaceuticals