AI Drug Discovery Advances as EMA Backs New Haemophilia Treatment
AI agents design validated cancer therapies, EMA recommends haemophilia approval, and the US debates Medicare reimbursement for medical AI software.
By Dr. Asher Knippel
Today's roundup covers landmark developments at the intersection of artificial intelligence and medicine, alongside a European regulatory milestone for haemophilia A.
Thursday, 18 September: Stanford's Virtual Biotech Deploys 37,000 AI Agents to Design Validated Lung Cancer Therapy
A Stanford Medicine spin-out published findings in Science showing that a coordinated network of 37,000 AI agents can predict drug trial success and independently design new therapies. Analysing 55,984 clinical trials, the system found that drugs targeting genes specific to particular cell types were 48% more likely to reach market and caused 32% fewer adverse events. One therapy designed by the AI — targeting the protein B7-H3 in lung cancer — was later independently validated by a major pharmaceutical company. The study is peer-reviewed and represents one of the largest computational drug-discovery analyses published to date. This is a promising proof-of-concept, though large-scale clinical validation of AI-led pipelines remains ahead.
Thursday, 18 September: EMA Committee Recommends EU Approval of Denecimig (FREHEMGO) for Haemophilia A
The European Medicines Agency's Committee for Medicinal Products for Human Use (CHMP) issued a positive opinion on 17 September 2026, recommending marketing authorisation for FREHEMGO (denecimig), developed by Novo Nordisk. Denecimig is a next-generation factor VIIIa mimetic bispecific antibody — it mimics the action of the clotting factor missing or deficient in haemophilia A, without being a replacement factor itself. It is indicated for routine prophylaxis to prevent or reduce bleeding episodes in haemophilia A (with or without inhibitors) in adults and children alike. Haemophilia A affects approximately 1 in 5,000 male births. A positive CHMP opinion typically leads to European Commission approval within two months; patients and physicians should await the final Commission decision before expecting prescriptions to change.
Thursday, 18 September: AI-Designed Drug Rentosertib Reduces Biological Age in Phase IIa Pulmonary Fibrosis Trial
Insilico Medicine announced on 17 September that its AI-discovered drug candidate rentosertib (ISM001-055) reduced biological age across six independent proteomic aging clocks in a Phase IIa clinical trial for idiopathic pulmonary fibrosis (IPF), a progressive lung disease with limited treatment options. The results were reported in Nature Biotechnology. Rentosertib is now advancing to Phase III. Insilico also published a companion cover study in Cell introducing LongevityBench, an open AI toolkit for aging research, and announced a Longevity Vaccines initiative targeting cellular drivers of age-related disease. These are early-to-mid-stage results; Phase III data will be needed to confirm clinical benefit definitively.
Monday, 15 September: U.S. Officials Accelerate Medical AI Reimbursement Amid Safety Concerns
The Trump administration is developing a new Medicare payment category to reimburse companies for AI software used in medical care or diagnosis. The FDA has already authorised more than 1,500 AI-enabled medical devices. Some federal officials have raised concerns that the pace of deployment outpaces the safety and efficacy evidence base. The debate carries particular weight for underserved communities, where AI tools are increasingly positioned as a substitute for specialist access rather than a supplement. Regulatory oversight of AI in clinical settings remains an evolving challenge across major health jurisdictions, including the EU.
Tuesday, 16–17 September: Novo Nordisk and Anthropic Partner to Deploy Claude AI in Drug Discovery
Novo Nordisk — maker of semaglutide and a leading company in diabetes, obesity, and cardiovascular medicine — announced a collaboration with Anthropic to deploy Claude AI for targeted drug discovery workflows and R&D software development. This is Novo's third major AI deal in 2026, following agreements with OpenAI and Amazon Web Services. An initial project will test Claude in specific research workflows. Novo's CEO stated the partnership will "supercharge" R&D for chronic diseases. The collaboration is at an early stage; clinical outcomes from AI-assisted pipelines typically emerge over years, not months.
This article is for journalistic and informational purposes only. It does not constitute medical advice. Readers should consult a qualified clinician before making any change to their treatment, medication, or care plan.