Why Abzena?
Trust our integrated approach.
Following CPhI Milan 2026, the pharmaceutical industry is reflecting on how increasingly complex medicines are reshaping development, manufacturing and global supply. From the continued growth of biologics and advanced therapies to the practical adoption of AI, greater supply chain resilience and the push for more sustainable manufacturing, the conversations at this year’s event pointed to an industry increasingly focused on turning scientific innovation into robust, scalable and commercially viable processes. In this roundtable, our CSO, Dr Campbell Bunce, alongside Minjeong Seong, Director of Market Intelligence at Samsung Biologics; Martin Meeson, CEO at Axplora; and Christiane Bardroff, Chief Business Officer, Advanced Manufacturing and Development Business Unit at FUJIFILM Biotechnologies, shared their perspectives on the developments shaping pharmaceutical R&D and manufacturing, and what they could mean for the industry ahead.
Campbell’s insights:
What stood out most was the progression of the conversation from capacity to capability. Next-generation modalities, such as bispecifics, multispecifics, ADCs, AOCs are now driving the manufacturing agenda. The question has moved beyond just “can you make it?” to “can you make it reproducibly, at scale and fast enough?” We’re seeing that CDMOs are being judged on their development and innovation capability, not just on availability of capacity and scale. AI also got its own zone this year, but the key discussion was about validation and governance rather than the “promise” of what it could deliver. I was also struck that sustainability now has its own summit rather than being a side session.
Overall, this tells me the industry is maturing around complexity. Molecules are getting more complex, so the partnerships that deliver them need to be more integrated, more scientifically led and more transparent. Simply being “an alternative” provider is no longer enough. Sponsors want partners who make better scientific decisions earlier and remove the white space between development steps.
The biggest challenges sit where molecular complexity meets compressed timelines. Scientifically, many of the newer formats don’t behave like standard antibodies. Fusion proteins, multispecifics and conjugates with new or dual payloads need to be expressed, characterised, and potency strategies built for the molecule rather than borrowed from a platform. On the manufacturing side, the historical split between producing the antibody intermediate and doing the conjugation creates handoffs, tech-transfer risk, and delay.
In the supply chain, the pressure points are specialist materials, HPAPI handling and payload capacity, and geopolitical exposure. Many sponsors are reassessing their dependence on single regions and are actively looking at onshoring and diversification. A particular challenge is a shortage of experienced people who understand both the biology and the chemistry.
The most important thing is to design for manufacturability from day one. Many costly failures start in early design and CMC, such as poorly chosen conjugation chemistry or inadequate analytical characterisation, that we may be unable to fix in Phase 3. A developability-led approach asks two questions early: can we make it, and does it work? It then designs the liabilities out before they become expensive.
Second, we need platforms that perform consistently across scales. Stable pools and clones should behave the same from 50L through 2,000L, so there are no performance cliffs. Third, analytics have to lead development rather than follow it. Robust, phase-appropriate methods are what give you reproducibility and regulatory confidence. Finally, commercial-readiness thinking, covering scalable processes, analytical robustness and supply planning, now needs to happen before pivotal studies, not after them. Quality and speed aren’t a trade-off when the science is right at the start.
Resilience is built by reducing the number of interfaces a programme has to cross. Every handoff between providers adds risk, time and another point where materials, data or knowledge can be lost. Integrated models that combine design, cell line development, linker-payload work, conjugation, analytics, and GMP manufacture under one accountable team make the supply chain both shorter and more transparent. Geographic diversification seems to be an important factor with sponsors too – rebalancing towards Western and dual-footprint partners to reduce geopolitical exposure.
We should be careful that resilience doesn’t come at the expense of capability. Back up sourcing for specialist materials such as novel payloads, linkers and HPAPIs should be considered alongside strategic partnerships and early supply planning, and not just-in-time thinking. Just as important is direct access to experts and clear data sharing, so that when something changes, decisions are made quickly and on good science. In my view, resilience is as much about knowledge continuity as it is about where the material is sourced.
We see significant value from AI at the front end, in design and developability. Tools such as in silico immunogenicity and developability prediction help us triage candidates and flag liabilities before we commit expensive lab and manufacturing resources. As timelines tighten, high-throughput front-end design that can reliably predict productivity early is becoming more valuable. In manufacturing, the promise lies in predictive process control and intelligent batch records, but that is where the bar rises sharply.
To be appropriate for GXP environments, AI needs well-curated, high-quality training data, transparent models, defined intended use and proper validation, just like any other analytical method. Governance and human expert oversight remain essential. I’d also argue that AI is only as good as the experimental data that feeds it. Wet-lab biology and chemistry generating good-quality data will remain the foundation. The next step is to move AI from proof of concept into validated, production-scale workflows.
Sustainability can partly be addressed through reducing manufacture risk and number of failed batches. Getting the science right early, through developability, robust analytics and right-first-time process design, cuts failed runs, rework and wasted materials. That is a major and often overlooked sustainability gain. Integration helps too. Fewer tech transfers mean less duplicated method development, fewer shipments of intermediates between sites and less material spent on bridging studies. Platform approaches that perform consistently across scale reduce development cycles and the resources they consume.
Single-use systems, process intensification and smarter solvent and reagent use in conjugation chemistry all contribute, but each has to be weighed against quality and regulatory requirements. The industry is now treating sustainability as a strategic priority, as this year’s CPhI Sustainability Summit showed. The key is to build environmental measures into process design from the outset rather than retrofitting them. That way, sustainability and quality support each other.