The-Unfinished-Work-of-Traditional-Bioprocessing

This Danaher Summit 2026 panel features three senior bioprocessing leaders — Ranga Godavarti (Pfizer), John Kaufman (AstraZeneca), and Thomas Saur (Sanofi) — representing a combined 90+ years of biologics development experience, examining the unfinished challenges of traditional bioprocessing. The 55-minute discussion covers the collapse of platform manufacturing under molecular format diversity, the AI and data readiness gap across the industry, the slow adoption of PAT and continuous processing, and the emerging vision for autonomous laboratory operations. Key tensions surface between the desire for rapid AI adoption and the reality of fragmented legacy data, between development perfectionism and manufacturing pragmatism, and between economies of scale and the resilience demands driving distributed manufacturing.

Key Takeaways

Platform biomanufacturing is declared obsolete: molecular format diversity means every new modality requires bespoke development, and the industry is transitioning to a 'toolbox model' combining high-throughput experimentation, process modeling, and AI in place of fixed process platforms.
Most biopharmaceutical companies' historical data is not AI-ready due to fragmented legacy systems and missing unique identifiers; companies with 7+ years of FAIR data (e.g., BMS) can already reduce new molecule process development to approximately 5 targeted experiments via Bayesian optimization — a decisive competitive advantage.
The critical-path window for process development speed is DNA to first GMP toxicology batch (4-6 months); after this point, CMC is rarely rate-limiting, so automation and HTE investment should be concentrated in this early window.
Process Analytical Technology has had FDA guidance since 2004 but remains underadopted due to unclear business cases and organizational risk aversion; the compelling business case is using feedback control of critical quality attributes to reduce deviations and cut phase 2-3 process characterization, which represents approximately two-thirds of total development effort per molecule.
Continuous downstream processing is the key missing enabler: switching from 3 batch chromatography cycles to 10-12 continuous cycles per day unlocks real-time product quality feedback control and addresses the decade-long performance asymmetry between exponentially improving upstream yields and lagging downstream efficiency.
AI interoperability requires shared control vocabularies and ontologies across development, manufacturing, and equipment vendors; equipment must expose programmatic command-line interfaces rather than graphical UIs for AI agents to operate autonomously — a challenge the panel framed as a direct call to arms for equipment manufacturers like Cytiva.
Dual sourcing of all critical raw materials and equipment is a perennial resilience requirement made acutely visible by COVID-19; equipment interoperability across competing suppliers is identified as a competitive differentiator that vendors should treat as a design priority.
Patient affordability — not just technical performance — is framed as a core development obligation; geopolitical pressures and tariffs are challenging the traditional scale-up cost model and accelerating interest in distributed regional manufacturing for biologics.