The Integrated Biology Environment is a speciation event
Have you ever found yourself building something impossible to explain that simultaneously feels inevitable?
Biologists do extraordinary science inside extraordinary friction. The friction is infrastructural. A working biologist might touch PubMed for literature, Elicit for extraction, a LIMS for sample tracking, Jupyter for analysis, and a separate platform for predictions, all before lunch. Each of these tools is useful in isolation; together they create a tax on every change of context. Data are trapped between PDFs, lab notebooks, and oral history. Months are lost due to information asymmetry.
Last week, Phylo announced their recent seed raise and introduced a term I think will matter: the Integrated Biology Environment, or IBE. The concept borrows from software engineering’s Integrated Development Environment (the IDE), where writing, execution, and debugging converge in a single workspace. Phylo is building the IBE for biomedicine, backed by a16z and Menlo Ventures. The analogy and the ambition are right.
But I think the IBE is bigger than biomedicine. I think it’s a category.
Tool speciation rewires workflows
Genuine speciation events in tooling introduce a new abstraction layer that permanently changes how work is done. The relational database changed the question from “how do I find this data?” to “what do I want to do?” The spreadsheet changed computing from a specialist task to one for anyone with a question about numbers. CAD software collapsed design, simulation, and specification into one process, where the drawing became the model.
Biology is no exception. PCR changed molecular biology by making DNA abundant, turning a scarcity problem into a selection problem. CRISPR changed gene editing from a multi-year technical feat to a programmable operation.
The IBE is next.
Scientists don’t need to be middleware
An IBE is a unified environment where biologists and machines collaborate across the full arc of research and development, from search to hypothesis to experiment to analysis to iteration, without leaving the workspace.
That sounds simple, but it’s not. The reason biology hasn’t had its IDE moment is that biological work is irreducibly heterogeneous. A single project easily spans literature review, sequence analysis, cell culture protocols, imaging data, statistical modeling, and physical experimentation. The tools for each of these have been built in isolation. The biologist is the integration layer—and that’s the problem.
An IBE replaces the biologist-as-middleware with an environment that holds everything together. But the IBE is not primarily about the agent. Agentic interfaces are table stakes. What defines the IBE is the tooling underneath: the APIs, data structures, and domain-specific infrastructure available to both humans and machines as collaborators.
The agent is the inhabitant. The IBE is the environment.
Three criteria define the IBE
If the IBE is a category, it needs to be classifiable. Here is my framework. If a platform doesn’t meet all three, it is not an IBE.
Ontological unity. The platform must enforce a shared data language across the entire scientific lifecycle—from request to process to result. A paper, a protocol, an experiment, and a prediction must share the same underlying structure. If a platform requires manual data translation between its own modules, it fails this criterion.
A single execution surface. The platform must collapse knowledge discovery, lab operations, and computational analysis into a continuous workspace. Transitions between literature discovery, wet-lab execution, and predictive modeling must occur without loss of context. If a scientist must manually re-enter or reformat data to move between core steps of the DBTL cycle, it is not an IBE.
A recursive loop. The platform must be capable of feeding experimental results forward into the next hypothesis. It cannot function as a static ledger of past work. The environment wires physical execution into a predictive layer so that accumulated data can compound and be shared across teams, institutions, and the broader scientific community. If the platform records history but cannot compound it, it is not an IBE.
Ontological unity gives biology a common grammar. A single execution surface makes that grammar operational. A recursive loop makes it generative and social. Remove any one and you have a tool, not an IBE.
The bioeconomy requires a shared foundation
The conversation around IBEs has so far centered on biomedicine—drug discovery, genomics, clinical workflows. That makes sense. Biomedicine is where data infrastructure is most mature and where the pain of fragmentation, as a result, is most visible.
But the IBE is not a biomedical concept. It’s a biological one.
Jennifer Kan has offered a useful framework for navigating the overlapping labels in the life sciences (biotech, techbio, synbio, genbio) by separating three questions: where biology is being applied (health, food & ag, industrial, environment, defense), how the work is being done (engineering principles, software and automation, generative models), and how value is captured (asset-led, platform-led, tools, services). The framework matters because it reveals that most of the confusion in our industry comes from conflating sector, approach, and business model.
It also reveals something about IBEs. An IBE is not confined to any single cell of this matrix. It is infrastructure that operates across sectors, approaches, and business models simultaneously.
The IBE is the missing layer. Not for one sector. For the bioeconomy at large. And like the bioeconomy itself, it won’t be monolithic. There will be as many IBEs as there are domains of biology, each built on the same three criteria, shaped by its own organisms, workflows, and data.
The environment collapses the scientific loop
For years, the life sciences have tried to make computation and biology work together. What’s been missing is the infrastructure to remove context switching between biology’s scientific loops. The IBE is that infrastructure: the environment that collapses the DBTL cycle into one place. The question is not whether biology will get its IDE moment. The question is when.
Today, no biology needs an IBE more than plants. Threats to production are rapidly evolving, and plant industries can’t keep pace. For our most sophisticated plant vertical, it takes 13 years and >$100M to bring an improved row crop to market. Decades for tree species. Meanwhile, the US loses over $20B in crops to new pests and pathogens each year.
Agriculture, forestry, health & beauty, medicine, restoration. Lagging plant innovation puts half of global GDP at risk. An abundant future requires innovative solutions to the way we produce and adapt. And delivering those innovations requires an evolution in how we build.
If biologists will be our agents of change, it’s time to build the habitat that ushers a metamorphosis.