Two Swedish companies, biotech firm Curve (formerly Re:meat AB) and software company Digital Tvilling AB, have announced a partnership to develop an integrated biomanufacturing platform that uses AI and digital twin technology to reduce the cost of precision fermentation over time.
The collaboration targets one of the core obstacles in alternative protein and ingredient production: unit economics. Precision fermentation holds considerable promise for producing proteins, functional food ingredients, and additives without reliance on animal agriculture, but production costs remain a persistent barrier to commercial viability.
Turning process data into a shared learning system
The two companies are building a system in which each fermentation run feeds data back into the platform, progressively improving efficiency across future batches. Curve contributes its bioproduction hardware and continuous data collection capabilities, while Digital Tvilling adds graph-based modelling, agentic AI, and cross-organisational data infrastructure.
The core premise is that bioprocess data, while expensive to generate, is routinely underused, often sitting in isolated systems with incompatible formats.
“Together, we are building a new kind of biomanufacturing platform where every production run contributes to making the system smarter, more efficient, and more scalable,” said Jacob Peterson, CEO and co-founder of Curve.
Built from the ground up
The hardware and software components are being co-designed rather than retrofitted, with observability and adaptive intelligence built in from the outset.
“Many of our clients operate in environments where systems and operational processes have evolved over decades. Here, we have the opportunity to build a new kind of integrated solution from the ground up, where observability, traceability, and adaptive intelligence are native capabilities from day one,” said Filip Åsblom, CTO and co-founder of Digital Tvilling.
Development is already underway, with the first integrated platform capabilities expected to be demonstrated in 2026.
