Hamilton Process Analytics makes the case for real-time PAT to scale cultivated and fermented foods

On Thursday 25 September 2025, Protein Production Technology International hosted an ICYMI-worthy webinar on how process analytical technology can accelerate scale-up for cultivated and fermented foods, with Hamilton Process Analytics joining leaders from the Good Food Institute, Opalia, and BioPharm Designs to outline what works now and what comes next. More than 350 people tuned in live from over 40 countries – spanning the USA, UK, Germany, India, Canada, Spain, the Netherlands, Israel, Australia, Brazil, South Africa, and many more, and heard at the outset how precision fermentation could surpass US$30 billion by the mid-2030s, and cultivated meat could supply up to 10% of global meat by 2035. The common hurdle is not promise, but performance at scale – producing efficiently, consistently, and safely.

Hamilton’s Scientific Content Manager, Charlotte Hughes, set the foundation by linking PAT to the FDA’s Quality by Design framework. “The idea is simple but powerful: by measuring critical parameters like pH or oxygen in real time, you gain deeper understanding and tighter control of your process,” she said. “I often describe PAT as the eyes and ears of the bioreactor. Without these tools, the process is essentially a black box.” She highlighted mature sensor platforms for pH, dissolved oxygen, carbon dioxide, conductivity, and cell density as near-term tools novel food producers can deploy to shorten development cycles and build trust with regulators and investors.

From the cultivated dairy front, Jennifer Côté, Opalia’s Co-founder & CEO, explained why milk made from cow cells follows different rules than meat. “Cell doubling doesn’t matter as much for us. What matters more is cell-specific productivity,” she said. Opalia uses a fully continuous process in which cells grow and secrete milk in the same culture “for extended periods of time”, improving cost efficiency. On consumer acceptance, she added that cultivated dairy can benefit from familiarity. “In some ways, it’s closer to precision fermentation… that product is milk,” which can be turned into cheese, butter, chocolate, and other everyday formats. Ultimately, “what really wins is taste”.

Turning PAT theory into plant-floor practice, Hamilton’s Strategic Business Development Manager, Aniekan Esenam, described how multiple dissolved oxygen probes reveal heterogeneity that a single sensor can miss in large stainless-steel bioreactors. That richer picture of oxygen uptake supports better decisions on mixing and refeed timing. Pairing oxygen with CO₂ adds the 'ears' to those 'eyes', he said, helping operators detect nutrient shortfalls or shear stress and correct in real time. Asked where founders should begin, Esenam was direct: make pH and dissolved oxygen standard, then add inline viable cell density. “We see cell density as the single most critical parameter to measure,” he said, because it underpins both unit economics and investor confidence.

Cost gravity framed the discussion throughout. “In biopharma, you have very high margins and low volumes. Here, it’s the complete reverse,” said BioPharm Designs Principal Consultant, Bert Frohlich. For cultivated meat, “scale-up is critical”,with raw materials likely to represent 70-80% of cost. In microbial systems such as mycoprotein, he sees downstream purity and off-flavor removal as the tougher challenge. On sensors, Frohlich praised capacitance-based viable cell concentration monitors that read total biovolume rather than simple counts. He called pH and dissolved oxygen “absolutely essential”, often used redundantly so a failed probe does not jeopardize a batch.

For models and data, the Good Food Institute’s Principal Scientist of Cultivated Meat, Elliot Swartz, pointed to two barriers that limit the impact of AI and predictive analytics today: access and standardization. Companies can now capture continuous data streams, but “the data just isn’t freely available”, he said. Even in consortia, incompatible taxonomies make it hard to learn across organizations. He noted efforts such as BioMADE’s work on a modular, interoperable digital backbone and argued that better measurement plus shared standards would accelerate tech transfer between facilities.

Audience questions pushed the panel into specifics. On measuring biomass in heterogeneous, semi-particulate media, Frohlich called capacitance probes the “gold standard”, cautioning that light scattering is easily confounded by other particulates. On probe hygiene, Hughes explained that Hamilton uses retractable housings and supports CIP and SIP so sensors can remain in place through cleaning and sterilization. For spent media, Swartz said residual nutrients are common – lactic acid, glucose, amino acids, ammonia, and other nitrogen sources – and noted emerging work on circular use as fermentation feedstock, though composition varies with formulation and cell metabolism. In differentiation stages, especially with fixed-bed systems, Frohlich warned that single-point sensing will not suffice; multiple sensors placed throughout the reactor are needed to understand gradients across scaffolds.

Safety and trust were recurring themes. “Regulators want to see that companies truly understand their processes and the products they’re producing,” Hughes said, arguing that continuous datasets document both the product and its production journey. Côté emphasized transparency with authorities and consumers, adding that safety must be communicated clearly and consistently. “Safety really is the keyword.”

Looking five to 10 years ahead, the panel converged on faster, richer feedback loops. “The ability to measure critical quality attributes of the final product in real time” would let manufacturers control processes proactively and cut waste, said Frohlich. Swartz hopes for shared, predictive models that give the field a common foundation. Esenam envisioned real-time multi-omic analysis inside the tank to steer taste, texture, and nutrition. Hughes stretched the horizon to a self-optimizing, near-autonomous system – an evolution from eyes and ears to a central nervous system for the bioreactor.

You can watch the full playback here.