What Every Biotech Investor Should Understand About Manufacturing Costs

Manufacturing is usually a footnote in biotech due diligence. For a certain class of companies, it’s actually the thesis.

When evaluating a biotech deal, the biology gets most of the attention. Mechanism. Target validation. Clinical data. Manufacturing is usually a footnote.

But for a certain class of biotech companies, manufacturing economics are the thesis. And even for those where they’re not, a misunderstood cost structure can quietly kill an otherwise good company.

The 50–80% Rule

In biologics manufacturing, downstream processing — the series of purification steps that take crude fermentation broth to drug substance — accounts for 50–80% of total manufacturing cost.

50–80%

of total mfg cost is downstream purification

52%

total recovery at 4 steps × 85% yield each

$30–100/g

typical commercial mAb COGS after 30 years of optimization

3–6 mo

smooth CMO tech transfer; 12–18 months if it goes wrong

Upstream (growing cells, running fermentation) is capital-intensive but efficient. Downstream is different: specialized equipment, consumables that wear out, and technical expertise at every step.

Every time a company tells you “manufacturing is solved,” ask which half they mean.

Why Purification Costs Can’t Be Compressed Easily

Separation steps have a structural cost problem: you’re fighting entropy. Taking a mixture of hundreds of components and isolating one to pharmaceutical purity requires multiple sequential steps — each adding equipment, time, reagents, and yield loss.

          Yield loss surprises people. If a purification train has four steps at 85% yield each, total recovery is 52%. Nearly half the product — and every dollar spent making it — is discarded.
        

A well-designed process gets to 75–85% total recovery. A poorly designed one recovers 30–40%. At commercial scale, that difference is worth tens of millions per year.

The Cost Curve Varies by Molecule Type

COGS varies by over two orders of magnitude depending on molecule type. The key drivers: availability of a standard capture step, process maturity, and production scale economics.

Molecule type	Typical COGS	Key cost driver	Maturity
Monoclonal antibodies (mAbs)	$30–100/g	Protein A resin ($8–15K/L), 30 years of process optimization, standardized platform	High
Recombinant proteins (non-mAb)	$100–2,000/g	No universal capture step; requires custom process development per molecule	Variable
Small-molecule fermentation	Lower per gram	Large volumes, crystallization and conventional separations; margin is volume-driven	High (industry)
Cell & gene therapies (AAV, lentivirus)	Very high	Viral vector purification expensive and yield-limited; not yet well-optimized	Emerging

The implication for diligence: a recombinant protein startup that hasn’t run process development is not comparable to a mAb company on the same platform. The cost uncertainty is structurally different.

The Tech Transfer Risk That Rarely Makes Pitch Decks

Most early-stage biotechs outsource manufacturing to contract manufacturing organizations (CMOs). Tech transfer — moving manufacturing from a startup lab to a CMO — sounds administrative. It frequently isn’t.

Common problems:

Process that worked at small scale doesn’t reproduce at larger scale. Process description not detailed enough for the new team to execute. Equipment differences between sites cause unexpected outcomes. Process was never tested across a range of conditions, so it’s fragile.

Timeline impact: Smooth tech transfer: 3–6 months. Difficult one: 12–18 months, $500K–$2M in additional development costs, and delayed clinical timelines. For a Series A company with 18 months runway, a 9-month delay is existential.

This risk is rarely quantified in decks, but it’s almost always on the critical path to the clinic.

Three Questions Worth Adding to Diligence

You don’t need a process engineering background to ask these. The answers will tell you a lot about manufacturing readiness.

1

“What’s your total process recovery, and what’s the yield at each purification step?”

If they don’t know, the process isn’t ready for tech transfer. Below 50% total recovery is a margin problem at commercial scale. If they know per-step yields, they understand their process. If they only have an end-to-end recovery number, the process hasn’t been characterized properly.

2

“Have you done any scale-up work, or has this only run at bench scale?”

Bench-to-pilot is where manufacturing surprises happen. A process that works in a 1 L bioreactor often behaves differently at 50 L or 500 L — especially in downstream, where fluid dynamics and shear forces change significantly. Any scale-up data, even to 10 L, substantially de-risks tech transfer.

3

“What’s your COGS estimate, and what’s it based on?”

An honest answer acknowledges uncertainty and walks through the assumptions: upstream titer, downstream yield, cycle time, resin cost per cycle, CMO day rates. A confident specific number derived only from bench data should prompt follow-up. COGS at bench scale is not COGS at manufacturing scale.

What Good Process Maturity Looks Like at Series A/B

For a company heading toward a Phase 1 IND, reasonable process maturity means the following are true:

Feed stream composition characterized (major impurities identified, approximate concentrations known)
Defined purification train with 2–3 steps selected based on molecule properties
Yield and purity measured at least once per step (not just at the end)
Written process description detailed enough to hand to someone else and have them run it
At least one robustness experiment run (±20% on a key parameter like pH, temperature, or loading density)

A company missing most of these hasn’t started process development in earnest. That’s not necessarily disqualifying — but the timeline and budget to get through tech transfer should be treated as uncertain.

The Bottom Line

Manufacturing is solvable for most molecules. The question is when you solve it and what it costs to get there.

Companies that start process design early compress that cost. Those that treat manufacturing as something to solve after the biology is proven hit it as a compressed, expensive problem on the critical path to the clinic.

          Manufacturing readiness is a timeline risk, a margin risk, and sometimes a dilution risk. Worth 30 minutes in diligence.
        