Model to molecule. Molecule to margins.
Every biotech product makes the same journey. It begins as an idea, and increasingly that idea starts as a model's output. It becomes a molecule that has to be made, tested and understood. And it only matters at scale, when it can be produced reliably, released with confidence and sold at a margin.
AI is reshaping the first step. We think the more durable opportunity lies in the steps that follow: the software that carries a design from model to molecule, and from molecule to margins. That's where Swea Ventures invests, at the early stage, in the US and the Nordics.
Why now
The hard problems in biotech are moving from the model to everything that comes after it.
- Ideas are cheap. Evidence isn't. Models can propose more candidates than labs can make and test.
- Lab data trains the next model. But only when it's captured with its full context, and most labs aren't set up for that.
- Pharma needs more from every program. Expiring patents and price pressure make speed and yield matter more.
- New plants need new systems. Drug manufacturing is being rebuilt in the US and Europe, and a new site is often when software gets chosen.
- AI has to show its work. Regulators expect AI used in drug development and manufacturing to be traceable.
AI also makes software cheaper to build, so small, expert teams can now take these problems on.
Where we invest
We back the software layer at each step of that journey, from the model to the bench to the plant.
- Lab data and infrastructure. Platforms that capture experiments, instrument output and metadata in a form both scientists and models can use, with a clear record of where every result came from.
- Lab automation and orchestration. Software that plans, runs and tracks experiments across instruments, robotic workcells and cloud labs, and closes the loop between design, testing and learning.
- AI for design and development. Tools for protein and ligand design, formulation and process development, sold to the teams that develop drugs.
- Manufacturing and quality. Software for manufacturing execution, batch records, process analytics and quality, where a validated design becomes a product.
Our buyers range from therapeutics pipelines and research organizations to contract manufacturers and industrial, food and agricultural biotech. We back the tools, not the pipeline. Companies that own their own drug candidates, and hardware-first businesses, fall outside our focus.
What we look for
The companies we back tend to share a few traits:
- Founders who know the bench. Someone on the team has run the lab, data or manufacturing workflow.
- A clear buyer. An R&D, platform, manufacturing or quality team that controls a budget and feels the problem today.
- Proof beyond the benchmark. Results that hold up in a customer's lab or plant, not only on a public dataset.
- Integration friendly. A product that works with the tools customers already use, from LIMS and ELN to MES and ERP.
- Data that compounds. A product that gets more useful as customers run more experiments through it.
- Ready for regulated work. Software that can pass a customer's security, quality and validation review where it needs to.
- Room to grow. A capital-efficient path to the next milestone, and a market beyond the home country.
Two markets, one bridge
The US is our home market. It has most of the buyers our companies need, the deepest pool of follow-on capital, and biotech clusters from San Diego to Boston.
The Nordics bring something different: a strong base of pharma, biomanufacturing and industrial biotech, public and foundation funding that backs science early, and a habit of building lean. Our roots there run through Lund and Copenhagen, two cities on either side of the Øresund that work as one life-science region. Founders in the region build excellent companies, but they often start far from US customers and investors.
The bridge runs both ways. We help Nordic companies land their first US customers and investors, and US companies reach buyers in Europe.
Our research looks at the evidence behind this thesis and the questions we're still working through.