mAbxience, a group predominantly owned by Fresenius Kabi and partially by Insud Pharma, has taken another step forward in applying artificial intelligence (AI) to industrial biotechnology in partnership with HP. Both companies announced a joint project to develop an AI solution capable of optimizing the monoclonal antibody and biosimilar manufacturing process by creating a digital twin of the biological process.
The initiative sits at the intersection of advanced biomanufacturing and high-performance computing, with a clear goal: to improve predictability, consistency, and efficiency in large-scale biological drug production—a key segment for the sustainability of healthcare systems.
A digital twin to understand and optimize the biological factory
The project is based on building a digital twin of the cell culture manufacturing process. Essentially, it’s a virtual model that replicates the behavior of the bioreactor and various stages of the production process, fed by real plant data and advanced neural network models trained and validated in an industrial environment.
Thanks to this approach, mAbxience teams can simulate, analyze, and optimize different manufacturing scenarios before implementing them on the actual line. The digital twin allows, for example, evaluating how changes in temperature, airflow, or nutrient composition will affect culture performance, product quality, or process stability—reducing the risk of deviations and out-of-spec batches.
According to the company, this tool enables optimization of different stages of the cell culture manufacturing process with the ultimate goal of increasing yields and reducing variability, both critical variables in biosimilar production.
AI with purpose: from efficiency to patient access
HP highlights the “purpose-driven” nature of the project: AI is not applied here merely as an innovation exercise, but as a direct pathway to ensure complex and costly treatments reach more patients more quickly and efficiently. As technical leaders explain, the use of digital twins and neural network models allows for shorter process tuning times, minimizing errors, and better leveraging each manufacturing campaign—ultimately translating into lower costs and increased medication availability.
In the case of biosimilars, this is especially relevant. These drugs are lower-cost alternatives to original biologics, but their development and production remain technically complex and heavily regulated. Any advances that enhance manufacturing robustness and repeatability can make a significant difference in ensuring supply and maintaining competitive prices.
A competitive boost for mAbxience and its partners
Apart from technological advancement, the AI solution has a direct impact on mAbxience’s business model. The company acts both as a biosimilar developer—licensing to partners in multiple markets—and as a Contract Development and Manufacturing Organization (CDMO), providing development and manufacturing services to third parties.
Integrating AI and digital twins into its manufacturing operations allows mAbxience to:
- Optimize plant resource utilization (equipment, raw materials, energy).
- Improve campaign planning and responsiveness to demand changes.
- Provide additional data and advanced analytics to its CDMO clients and licensees, strengthening its value proposition.
Company executives emphasize that the project reinforces their position as a reliable, cost-effective global biomanufacturing partner, while also helping improve patient access to safe, high-quality biological medicines.
HP brings computing muscle and AI expertise applied to industrial processes
HP contributes its experience in computing infrastructure, data analysis, and AI models tailored to industrial environments. Developing a digital twin based on real production data involves handling vast amounts of information, training sophisticated neural networks, and validating models against plant results—challenging without robust computing capabilities and specialized software.
The collaboration resulted in an initial prototype that has already been validated in an industrial setting, allowing mAbxience to simulate more efficient, resilient, and controlled manufacturing campaigns. Building on this, both companies aim to refine the model further, incorporate more process parameters, and extend the approach to other product lines and facilities.
AI and biomanufacturing: an expanding trend
mAbxience and HP’s project is part of a broader trend: adopting AI across the entire value chain of biologics and biosimilars—from molecule design to production planning and real-time process control.
In biomanufacturing, AI begins to be used for:
- Identifying complex correlations between process parameters and product quality.
- Detecting anomalies early to prevent deviations before they compromise a batch.
- Recommending optimal operational adjustments based on performance and quality goals.
The current challenge is integrating these capabilities within regulated frameworks, with explainable models and traceable decisions, demonstrating to authorities that AI can coexist with the sector’s strict safety and quality standards.
A step closer to “smart” biopharmaceutical factories
With this initiative, mAbxience and HP showcase how combining historical plant data, digital twins, and neural networks can transform the design, scaling, and operation of bioprocesses.
If the solution proves successful and scalable, it could become a core component of future “smart” biosimilar factories—facilities where processes adapt in real time, anticipate issues, and leverage each production cycle to improve the next.
Ultimately, the end beneficiaries will be patients, who could have access to more biological treatments across more countries, with more stable supplies and at more sustainable prices.
Frequently Asked Questions about the mAbxience and HP AI Project
What is a biosimilar, and why is its production so complex?
A biosimilar is a highly similar biological medication to an already authorized reference product, with regard to quality, safety, and efficacy. Since it’s produced in living cells, any process change can affect the final product, requiring extremely strict control over all cultivation and purification parameters.
What does “digital twin” mean in biomanufacturing?
A digital twin is a virtual replica of a physical process or system. In biomanufacturing, it models cell culture behavior and various production stages based on real data, enabling simulation of scenarios, troubleshooting, and optimization before implementing changes on the actual plant.
How does AI help improve biosimilar production?
AI analyzes large volumes of process data to uncover patterns not visible to the naked eye. This can enhance performance predictability, reduce batch-to-batch variability, suggest operational adjustments, and make production more efficient and robust.
What does this project mean for patients and mAbxience partners?
For patients, a more efficient and controlled production process can mean greater availability of biological treatments and biosimilars at more affordable prices. For licensees and CDMO clients, it translates into optimized processes, better resource utilization, and a technologically advanced industrial partner.
Sources: mAbxience; HP; press releases and corporate content on AI and digital twins in biomanufacturing.