Repligen made a quiet bet on biopharma's next intellectual property war. It is not about molecules.
In September 2025, the Massachusetts-based biologics equipment maker took its first ever outside investment in Novasign, a Vienna-based software startup Dealroom. The deal came just two months after Repligen publicly announced a strategic partnership with the company, and the timing matters: the equity came first, the press release second Repligen press release. Novasign makes digital twins for microbial protein factories, software that simulates bioprocesses before companies have to run them. Repligen (NASDAQ: RGEN) is not buying a licensing agreement. It is buying a position in a data layer most investors have not named yet.
The core technology is real. Novasign's digital twin uses hybrid modeling, combining mechanistic equations that describe the physical boundaries of a bioreactor with artificial intelligence that learns from each experiment run GEN News. According to published data from a collaboration with Boehringer Ingelheim, this approach reduced the number of required experiments by more than 66 percent compared to the standard method, called Design of Experiments PubMed. The company claims 70 percent in its marketing materials. The peer-reviewed result anchors that number. At the Bioprocessing Summit Europe last year, Novasign's chief executive Mark Duerkop demonstrated the software running a continuous protein production process for 30 consecutive days without manual intervention GEN News.
That is not yet what the FDA calls autonomous manufacturing. Regulators require extensive validation before approving self-optimizing production systems, and Duerkop himself told GEN News that remains years away GEN News. The 30-day run was a proof of concept, not a licensed process.
But the commercial logic is already in motion. Repligen's press release described the partnership as aimed at deploying digital twins across Repligen's filtration systems, with the stated goal of reducing development timelines and costs Repligen press release. The company sells the equipment and consumables that biopharmas use to manufacture protein drugs at scale. If process modeling becomes standard in those facilities, Repligen wants to own the software layer.
This is the shift that matters for anyone building or investing in biologics manufacturing. For three decades, the intellectual property battle in biopharma centered on molecules: who owned the patent on the antibody, the enzyme, the gene therapy construct. The manufacturing process was treated as infrastructure, not IP. That assumption is breaking.
The reason is data. A digital twin does not just simulate one process; it learns from every experiment a company runs inside it. The model gets more accurate over time. It captures the specific behavior of a company's cell line, their raw material suppliers, their operating ranges. That accumulated learning is not easily replicated. It is not a patent that can be designed around. It is more like a search engine's index: the value compounds as more people use it, and a competitor starting from scratch cannot simply catch up by hiring the same engineers.
Novasign's ECOnti consortium, funded by Austria's Research Promotion Agency and led by recombinant protein manufacturer enGenes Biotech, has been building this capability since roughly 2023 GEN News. The consortium's prototype currently runs at one-liter scale. enGenes has said a shipping-container-sized manufacturing unit could be ready within two years if investment materializes. That timeline is aspirational, as timelines at the lab-to-production boundary tend to be.
The competitive landscape is not empty. Thermo Fisher, Cytiva (Danaher), and Sartorius are all investing in digital manufacturing platforms, and the continuous bioprocessing market is projected to grow at roughly 17 percent annually through the early 2030s, according to industry analysts. What distinguishes the Repligen-Novasign bet is the explicit focus on filtration systems, which represent one of the most critical and cost-intensive steps in downstream protein manufacturing Pharma Manufacturing.
The unanswered question is who owns the process model when the engagement ends. In the current contract structure at most contract development and manufacturing organizations, the customer pays for the experiment but rarely retains the model, according to people who negotiate these agreements. That is the fault line. If digital twins become the primary intellectual property in biologics manufacturing, the party that holds the model holds leverage in every future negotiation. Repligen is apparently betting that position will be worth more than any single molecule patent on its balance sheet.
The technology is early. The regulatory path is long. But the direction of travel is clear, and Repligen just moved first.
