Antibody-Drug Conjugate Manufacturing: Cost vs. Sustainability, 5 Years On
Antibody-Drug Conjugates · Manufacturing Strategy
Antibody-Drug Conjugates, Five Years On: What Our Modeling Got Right — and the One Thing It Didn’t See Coming
In 2021, we asked how the cost of making ADCs could be monitored and improved. Half a decade later, the surprise isn’t the cost. It’s what happens when you put cost and sustainability on the same page.
Back in January 2021, we published a blog observing and arguing that antibody-drug conjugates weren’t simply “too expensive”; they were not properly understood. The field was buoyant: six approvals in three years, around eighty molecules in the clinic, and a growing anxiety about whether anyone could make these things at a blockbuster scale.
So, we built a model. Using BioSolve® Process, we set up a protein conjugation process. We ran a representative case, a HER2-targeting, cysteine-linked trastuzumab ADC, to ask a deceptively simple question. What scale of facility, and what size of investment, would it take to supply a commercial ADC? A dedicated conjugation facility at roughly 200 kg/year, producing something like 0.6 million 100 mg vials, enough to treat around 20,000 patients, gave us a concrete basis for reasoning. And the headline finding was clear: raw materials, the purchased antibody, and the drug-linker accounted for the majority of costs. Everything else was a rounding error by comparison.
That was the right question for 2021. The industry needed to know it could make ADCs at scale. Five years on, the question has changed. It’s no longer “can we make it?” It’s “how do we make it well — economically and ecologically, at once?”
What changed
The most consequential shift since 2021 isn’t the chemistry of any single molecule. It’s the manufacturing platform underneath the whole enterprise. Over these five years, ADC conjugation has moved decisively away from stainless steel with clean-in-place (CIP) toward single-use systems, with parallel changes in solvent handling (DMF giving way to DMSO) and containment (OEB-5 isolators becoming standard for high-potency work). These aren’t cosmetic. They rewrite the process’s cost and environmental profile from the ground up.
When we rebuilt our 2021 case to reflect a 2026 process, we expected the switch to single-use to look like a clean win. It’s the modern choice; it’s flexible; it removes the enormous burden of cleaning validation. What we found instead was more interesting, and it’s the reason I’m giving a talk about it this August.
The thing we didn’t see coming
Here’s the tension, stated plainly. On the single most important process decision — stainless with CIP versus single-use — cost and sustainability point in opposite directions.
Modeled change from the 2021 stainless + CIP process to a 2026 single-use process (representative HER2 ADC case, ~100 kg/year). One switch, two opposite results.
Our modeling indicates that switching to single-use raises the process’s conversion cost by roughly 20 percent. The driver is disposables: you’re buying assemblies rather than cleaning and reusing equipment, and with one reuse, those consumables become the dominant line in the conversion cost.
And yet the same switch cuts the process mass intensity (PMI, the total mass of inputs per unit of product) by about 75 percent. The reason is almost entirely one thing: eliminating CIP removes an enormous cleaning-water burden. In the stainless case, cleaning water is the single largest mass stream in the entire process. Take CIP away, and it simply vanishes.
Process mass intensity breakdown by input stream. Cleaning water — the CIP burden of a stainless facility — accounts for the bulk of 2021 throughput and disappears with single-use.
Sit with that for a moment. The item that makes single-use expensive, the disposable plastic, is nearly weightless. The thing that makes stainless steel heavy, all that cleaning water, is relatively cheap. Cost accounting and mass accounting rank the very same decision in opposite orders. Energy and carbon fall too, though with a nuance worth noting: some carbon shifts from operational emissions to the embodied, supply-chain footprint of the disposables themselves.
There’s a quieter gain hiding in the numbers as well. The newer process uses a few percent less antibody per gram of ADC — a small efficiency that matters more than you’d think once you remember how much the antibody costs.
Why this is a system problem, not a chemistry problem
I want to be careful here. These are BioSolve® Process modeling outputs from one representative case, not universal laws. In fact, whether single-use or stainless “wins” on cost isn’t fixed at all; it depends on the antibody input price a given program actually faces. There’s a crossover point, and where your program sits relative to it changes the answer. (That’s a detail I’ll walk through properly at the Summit.)
Schematic: the two processes trade places as antibody price rises. Where the lines cross — and which side your program sits on — is the story I’ll tell in Boston.
A word on the raw materials. Throughout this comparison, we hold both the antibody and the toxin-linker at their 2021 prices, so what you’re seeing is a shift in technology, not in the market. In reality, the toxin-linker has moved too: our figure is drawn from AI-inferred should-cost data rather than supplier quotes, and on that basis, a realistic 2026 auristatin toxin-linker sits at around $1,650/g, comfortably below the price we hold fixed for the comparison. We’ve deliberately excluded that delta from the like-for-like analysis to keep it honest, which means the real 2026 economics are better than the numbers here imply.
But the general lesson holds, and it’s the whole point. If you had looked only at cost, you’d have hesitated over single-use. If you had looked only at ecological sustainability, you’d have adopted it without a second thought. Either metric alone would have led you somewhere; the combined picture suggests it is too simple.
This is why we keep insisting that ADC economics is a system-level manufacturing problem, not just a chemistry problem, and why cost and ecological sustainability have to be modelled together, not as separate afterthoughts bolted on at the end. A late-stage spreadsheet built after the process is fixed can only tell you what you’ve already committed to. Predictive, integrated modeling connects molecular design, process choices, facility impact, cost, and environmental footprint early, while the decisions are still live, before capital and process are locked in. That’s where the value is: not better estimates, but better decisions, sooner.
And the ground is about to move again. The horizon beyond 2026 — continuous-flow conjugation, site-specific and enzymatic chemistries, and AI-assisted process control — will reshape the cost architecture a second time. BioSolve® Process is built to model these shifts before anyone commits capital to them.
Come and see the full picture
I’ll be unpacking all of this — the crossover, the trade-offs, and how the molecule-to-facility model actually works — at The Bioprocessing Summit, August 10–13, 2026, at the Omni Boston Hotel at the Seaport in Boston, MA (and virtually). My talk, “From Molecule to Facility: Predictive Modeling for ADC Cost and Sustainability,” sits in the Intensification & Digitalization / Digital Transformation streams. If ADC manufacturing strategy is on your desk, I’d love to see you there.
