Biology has never lacked data, tools, or ambition. What it has lacked—until now—is integration. Phylo, an applied AI research lab focused on accelerating biological discovery, is betting that agentic AI can finally close that gap.
The company today announced the launch of Biomni Lab, a fully integrated, AI-native workspace designed to help scientists plan, write, execute, and collaborate on complex biological research using state-of-the-art agentic AI. Alongside the product launch, Phylo revealed a $13.5 million seed round, co-led by Andreessen Horowitz and Menlo Ventures’ Anthology Fund, with participation from Zetta, Conviction, SV Angel, and others.
Together, the announcements signal Phylo’s transition from influential open-source research to a commercial platform aimed at reshaping how biological research actually gets done.
From Open-Source Breakthrough to Integrated Lab Environment
Biomni Lab builds on Biomni, an open-source project Phylo introduced in June 2025 that quickly gained traction as the first Integrated Biology Environment (IBE). Biomni combined agentic AI with hundreds of domain-specific tools to create a generalized biomedical AI agent—an ambitious attempt to unify fragmented biological workflows.
Biomni Lab takes that concept several steps further. It merges a significantly upgraded agent architecture with an expanded, production-grade environment, giving scientists a single place to orchestrate AI agents reliably, reproducibly, and at scale. The goal isn’t to replace biologists, but to offload the tedious, error-prone work that slows discovery to a crawl.
Kexin Huang, Ph.D., Phylo’s co-founder and CEO—and first author of the original Biomni research paper—frames the shift as inevitable. Coming from a computer science background, she saw clear parallels between biology and pre-AI software engineering: powerful tools, stitched together by fragile workflows.
Biomni Lab, she argues, applies the same AI-native productivity leap to biology that tools like GitHub Copilot brought to code.
What Makes Biomni Lab Different
At the heart of Biomni Lab is a new agent architecture that Phylo says outperforms existing agent systems by more than 20% across standard benchmarks and real-world, long-horizon scientific evaluations. That performance gap matters in biology, where experiments often span days or weeks and depend on tightly coupled computational and physical steps.
Unlike generic AI copilots, Biomni Lab is designed specifically for biological research. It allows scientists to:
- Plan and author complex experimental workflows
- Execute computational analyses across multiple tools and datasets
- Reproduce prior experiments without rebuilding environments from scratch
- Collaborate across wet-lab and computational teams in a shared workspace
The system is built to handle both physical and computational biology, reflecting how modern research actually happens. Scripts, dependencies, datasets, and experimental logic live together, reducing the constant context-switching that defines much of today’s lab work.
Yuanhao (Jerry) Qu, Ph.D., Phylo’s co-founder and president, points to familiar pain points: re-creating experimental setups, debugging brittle pipelines, and manually wrangling tools that were never designed to work together. Biomni Lab, he says, introduces a new default—one where an expert AI agent is always available to handle these tasks, compressing week-long cycles into minutes without sacrificing rigor.
Early Validation: From Weeks to Hours
This isn’t just a conceptual promise. Biomni Lab has already been tested in production settings.
In a case study with Ginkgo Bioworks, the platform accelerated more than ten complex cell-painting and transcriptomic analyses. According to Ginkgo scientists, the results were publication-quality—and workflows that typically took weeks were reduced to just hours.
Ayla Ergun, Senior Director of Data Science at Ginkgo Bioworks, described Biomni Lab as accessible to scientists across disciplines, capable of automating bioinformatics analyses, generating figures, and even benchmarking results against external datasets. Ginkgo Datapoints is now exploring making Biomni an everyday part of its workflows.
That kind of endorsement is notable in an industry known for skepticism toward new tools—especially those promising dramatic productivity gains.
AI-Native Biology Meets Investor Conviction
The funding round underscores how strongly investors are leaning into AI-native scientific platforms. Andreessen Horowitz and Menlo Ventures’ Anthology Fund—created in partnership with Anthropic—are both making concentrated bets on agentic AI systems that move beyond chat interfaces into real-world workflows.
Jorge Conde, General Partner at Andreessen Horowitz, highlighted one metric that stands out in life sciences: adoption. Biomni, as an academic project, has already been adopted by more than 7,000 labs, biopharma, and healthcare organizations, a level of real-world usage that’s rare for research-born tools.
Menlo Ventures partner Matt Kraning echoed that sentiment, framing Phylo’s work as a fundamental change in how biological research happens—bringing AI-native productivity directly into the lab, not just the cloud.
Rooted in Open Science, Aimed at Scale
Despite launching a commercial product, Phylo isn’t abandoning its academic roots. The company spun out of the open-source Biomni project at Stanford and says it will continue to maintain and invest in that community.
That dual-track approach—advancing open research while commercializing production systems—reflects a broader trend in AI tooling, where open ecosystems often drive faster adoption and trust, especially in scientific domains.
Phylo’s founding and advisory roster reinforces that blend of credibility and ambition. Alongside Huang and Qu, the company counts Stanford professors Jure Leskovec and Le Cong as scientific co-founders, with advisors including Nobel laureate Carolyn Bertozzi, CRISPR pioneer Feng Zhang, and computational biology leader Fabian Theis.
It’s a lineup that signals Phylo’s intent to operate at the intersection of cutting-edge AI research and real-world biological impact.
The Bigger Picture: Biology’s Workflow Problem
For decades, biology has advanced through breakthroughs in tools—sequencing, imaging, automation—but workflows have remained stubbornly fragmented. Data moves between scripts, spreadsheets, and specialized software, often with manual glue holding everything together.
Biomni Lab is part of a growing push to rethink biology as an AI-native discipline, where agents don’t just answer questions but actively manage and execute research processes. If successful, that shift could change not just productivity, but the pace at which hypotheses are tested and discoveries made.
The comparison to AI-assisted software engineering is telling. Once developers experienced AI copilots that understood context, tools, and intent, there was no going back. Phylo is betting biologists will feel the same way.
A New Default for the Lab?
Biomni Lab is launching in research preview, but its ambitions are clear. By unifying agentic AI, domain-specific tools, and collaborative workflows, Phylo wants to become the operating system for modern biological research.
Whether it succeeds will depend on adoption beyond early innovators and elite labs. But with strong open-source roots, early industrial validation, and backing from some of Silicon Valley’s most influential investors, Biomni Lab is positioning itself as more than another AI assistant.
It’s aiming to redefine how science gets done.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
