Celltrion said manufacturing work is focused on smart factories that apply physical AI at new plants. It said it plans to raise automation at drug-substance plants 4 and 5, which are to be built in Songdo, with autonomous mobile robots, an automated warehouse, intelligent and collaborative robot arms, and manufacturing-management software.
A Takeda R&D Quality lead described a multi-agent AI system for preparing CDMO audit briefings. It was shown as a validation-feasibility study on synthetic data, not a production deployment.
An AI quality-intelligence tool that began as a narrow pilot on deviation and complaint trending with suggested root causes, and is now used at several manufacturing sites.
Partners: Aily Labs
J&J's CIO cited AI helping manufacturing teams decide the right time and temperature to add solvent during production.
Merck KGaA's 2025 annual report says its smartfacturing playbook, including modular manufacturing, AI energy optimization and predictive maintenance, cuts time to market by up to 50% and uses around 20% less energy at reference sites.
An in-house AI tool that learns from past and current batch performance to push manufacturing yields higher. Sanofi estimates about EUR 10 million a year in savings from analyzing over 13,000 Lovenox batch runs.
Lilly modelled the steps of a manufacturing process as a high-fidelity digital twin to find ways to raise output, according to its CIDO.
Lilly uses AI to watch product on its lines and catch defects, per its CIDO.
Astellas says it actively uses artificial intelligence, automation and data science to optimize manufacturing processes, enhance quality management and improve operational efficiency. It says it is introducing work-support tools that use generative AI and data mining, and digital twins to simulate manufacturing processes and supply chains.
Vendor case study: Weights & Biases says GSK built manufacturing digital twins at three levels of sophistication, with real-time sensors feeding ML workflows that predict batch trajectories, and that GSK has optimized 13 key products this way.
Partners: Weights & Biases
Weights & Biases says GSK replaced a two-person check of environmental petri dishes with a one-person-plus-AI deep-learning system, operating across nine manufacturing sites and about 3 million images a year.
Partners: Weights & Biases
An AI assistant trained on Takeda's internal investigation standard that helps staff write more consistent deviation investigation reports.
Amgen process development staff use generative AI tools over a curated technical knowledge base of more than a million documents, with evaluations and risk-based validation.
Astellas and YASKAWA Electric signed an agreement to establish a joint venture that will develop a cell-therapy manufacturing platform using the Maholo robot. The planned business includes optimizing the digitized manufacturing process with AI and building a platform that meets GMP conditions. Establishment was subject to closing conditions, including regulatory approvals.
Partners: YASKAWA Electric Corporation
Amgen aims for predictive models and digital twins across its manufacturing network; process-parameter monitoring had been shown only in proofs of concept at the time.
With AWS and Accenture, Merck built a cloud inference pipeline (HawkAVI) that analyses large volumes of inspection-machine data from its manufacturing sites.
Partners: AWS, Accenture