From Lab to Scale: The David Friedberg Checklist for Biomanufacturing Readiness

‍

David Friedberg often gets asked a simple but brutal question in biomanufacturing: when is a process really ready to scale.

I hear this question every week from founders and operators who feel stuck between exciting lab data and the harsh reality of pilot runs, capital budgets, and COGS.

In this guide, I lay out a practical, no-nonsense readiness checklist inspired by the way David Friedberg talks about biomanufacturing economics and scale-up discipline.

You’ll learn how to stress-test your fermentation, process development, downstream, cost model, and go-to-market plan before you burn cash on steel and capacity.

I share hard-won lessons, simple tests, and patterns that separate winners from costly science projects.

‍

1) What Is the David Friedberg Biomanufacturing Readiness Checklist

I think of the David Friedberg checklist as a tough-love filter for any biomanufacturing program claiming it’s ready to scale.

It forces clarity on five questions.

• Do you have product/market clarity. What exactly are you selling, to whom, at what spec, and how will they validate it.
• Do you have a repeatable bioprocess. Can you hit the same results three runs in a row on a relevant scale-down model.
• Do you have a credible COGS and capacity model. Can you produce at a margin that supports a business, not a paper.
• Do you have a de-risked scale-up path. Are kLa, OTR, P/V, mixing time, and heat removal accounted for with scale-appropriate correlations.
• Do you have a financing and partner plan. Can you fund the path and secure the right CDMO or facility without losing time or control.

If you can’t answer yes to all five, you’re not ready yet.

That’s not a critique.

That’s the checklist doing its job.

2) Why Lab-to-Pilot Readiness Matters More Than Your Best Flask Result

I’ve watched beautiful bench data die at 300 liters because of one missing link: a validated scale-down model that mimics large-scale hydrodynamics.

At bench scale, oxygen transfer is easy, heat removal is trivial, and mixing is instant.

At pilot scale, oxygen transfer becomes the taxman, heat removal fights you in exothermic runs, and antifoam wrecks downstream yield.

Here’s my short rule.

• Flask wins don’t pay rent. Reactor wins do.
• Biology that only works gentle is biology that won’t ship.
• Three-good-runs in a scale-down model beat one miracle titer every time.

If you’re serious about scale, make your bugs and your DSP earn it in a realistic, mean-spirited model of your future plant.

3) Define the Product and Spec Before You Touch a Reactor

The fastest way to waste a year is to scale a process before you lock the product spec and the customer validation path.

I always ask founders to write a one-pager that answers five things.

• Use case. What job does this molecule, organism, or material do.
• Must-have specs. Purity, potency, activity, particle size, moisture, endotoxin, residual solvents, or any industry-specific metric.
• Acceptance test. How will your customer test and approve a lot.
• Packaging and shelf-life. What format goes to the customer and how long is it stable.
• Regulatory classification. Food, feed, cosmetic, industrial, or pharma-grade, with the matching documentation burden.

When you lock this, your process development gets laser focused and your COGS model turns from fiction into a plan.

4) Strain and Construct Readiness: Stability Beats Hero Titers

I’ve seen founders chase a headline titer and then watch the strain drift during a 14-day run.

That’s not readiness.

Here’s what I look for before scale-up.

• Genetic stability under stress. Run serial subcultures and long batch times with selective pressure off and on.
• Clear copy number and expression control. Avoid designs that depend on perfect induction timing to work.
• Process-tolerant strains. Strains that tolerate shear, antifoam, pH swings, and nutrient limitations will save you time later.
• Freedom to Operate. Do an honest IP scan and lock your FTO early or risk rework.

Stability and robustness beat a fragile race car strain every day at production scale.

5) Fermentation TRY Targets: Define the Line in the Sand

Every bioprocess lives and dies on TRY: titer, rate, and yield.

I ask teams to set a simple target table that ties TRY to COGS and price.

• Titer. How many grams per liter do you need to hit your cost model.
• Rate. Volumetric productivity drives tank turns, revenue per reactor, and capacity planning.
• Yield. Substrate-to-product yield determines raw material cost and waste.

Make three boxes.

• Green. On-model and fundable.
• Yellow. Viable but margin-sensitive.
• Red. Uninvestable without a price or spec change.

Then run your experiments and report against those boxes like a business, not a paper.

6) Scale-Down Models and Scale-Up Criteria: kLa, OTR, P/V, and Mixing Time

Scaling is not a vibe.

It’s math plus scar tissue.

Pick your scale-up criteria and stick to them.

• kLa. If oxygen is limiting, match volumetric oxygen transfer coefficients across scales.
• OTR with safety margin. Calculate OTR needs at peak respiration and leave headroom.
• P/V. Power per volume is a workhorse surrogate for mixing and shear.
• Tip speed and mixing time. Don’t scale shear-sensitive cultures on P/V alone.
• Heat removal. Model duty and check jacket/coil capacity at the pilot and production scale.

Build a scale-down reactor that mirrors your target plant geometry, baffles, impellers, and spargers.

Run it like the plant will run.

That means real gas flows, antifoam strategy, and CIP/SIP cycles that your biology has to survive.

7) Downstream Processing: Lock the Big Decisions Early

Downstream destroys margins quietly if you ignore it.

Pick a primary recovery and purification strategy that respects physics and scale.

• Cellular vs secreted. Intracellular products drive lysis and solids handling costs, while secreted products shift burden to filtration and concentration.
• Capture step. Affinity is great for purity, bad for resin cost at scale unless you recycle aggressively.
• Solvent and antifoam management. Anything you add upstream shows up downstream as cost and risk.
• Crystallization or formulation. If the market accepts a technical-grade or formulated product, lock that decision early to shrink polishing costs.

Run mass balances and yield ladders for each DSP concept and choose the simplest path that meets spec.

8) Quality by Design and DOE: Map Your CPPs and CQAs

If you can’t state your Critical Quality Attributes (CQAs) and the Critical Process Parameters (CPPs) that control them, you’re not ready.

I use a simple QbD flow.

• Define CQAs. The quality outcomes the customer cares about.
• Hypothesize CPPs. Variables like pH, feed rate, temperature, induction timing, and harvest time.
• Screen with DOE. Use fractional factorials or response surface designs to map sensitivity quickly.
• Set proven acceptable ranges. Lock them in your batch records and train the team.

QbD is not paperwork.

It’s the operating system for reproducibility.

9) PAT and Data Backbone: Measure What Matters in Real Time

Real processes watch themselves.

If you only see what happened after the batch is over, you learn slowly and lose money.

I advise teams to implement lightweight PAT and a data backbone early.

• On-line/at-line sensors. pH, DO, off-gas analysis, capacitance biomass, NIR/Raman for titer proxies.
• Historian. Centralize time-series data for every run and annotate events.
• Golden batch profiles. Build templates and flag deviations in real time.
• Feedback control. Automate feed and induction strategies when signals are trusted.

Small investments in PAT reduce failed batches and accelerate learning curves.

10) Tech Transfer Playbook: How to Not Get Eaten by a CDMO

I love great CDMOs and I’m wary of sloppy tech transfers.

Write a transfer package like you’re explaining your process to a smart stranger who wants to help you win.

• Process narrative. Step-by-step with purpose for every step.
• Bill of materials. With alternates, grades, specs, and qualified vendors.
• Process parameters and ranges. Clear CPPs, alarms, and acceptable deviations.
• Sampling and analytics plan. What to measure, when, and how to decide.
• Change control rules. Who approves adaptations and how you document them.

Visit the site before you sign.

Match your process to their equipment, utilities, and team strengths.

For more on investor-ready documentation, see our blog post: How to Build a Data Room Investors Love.

11) COGS, CAPEX, and Breakeven Volume: The Three-Tab Truth

A credible cost model fits on three tabs and drives real decisions.

• COGS tab. Raw materials, consumables, labor, utilities, yields, and scrap.
• Capacity tab. Tank turns, batch size, uptime assumptions, and bottlenecks.
• CAPEX tab. Fermenters, DSP trains, utilities, commissioning, and validation.

Make three scenarios.

• Base case. Today’s performance and realistic learning curve.
• Upside. With specific process wins, not magic.
• Downside. With expected losses, rework, or supply price shocks.

Back your price and margin goals into TRY targets so the team sees the business impact of every experiment.

For a deeper dive into modeling, see our blog post: COGS Modeling Template for Founders.

12) Supply Chain Readiness: Don’t Let a Valve Gasket Delay Your Launch

I’ve watched six-figure batches slip because a filter grade was backordered and no alternate was qualified.

Supply chain readiness is not a procurement problem.

It’s a process risk problem.

• Dual-source critical materials. Media, resins, filters, and single-use components.
• Qualification packets. Vendor CoAs, change notifications, and quality agreements in place.
• Lead-time buffers. Plan for the longest pole, not the average.
• Obsolescence plan. Know what you’ll do if a vendor pulls a product line.

Run a mock shortage and see how fast your team can swap a resin or filter without blowing your spec.

13) Equipment, Utilities, and Facility Class: The Quiet Constraints

It’s tempting to assume the plant will have what you need.

It won’t unless you check.

• Utilities. Air, steam, chilled water, WFI or PW, and clean gases sized to your worst case.
• CIP/SIP capability. Cycle times, temperatures, and validated recipes that your strain tolerates.
• Containment. Biosafety level, dust control, ATEX if solvents, and waste handling.
• Materials of construction. Are you OK with 316L and elastomer choices for your solvents and pH.

Walk the facility, ask stupid questions, and bring your process engineer who loves piping and instrumentation diagrams.

14) GMP, Non-GMP, and Regulatory Pathways: Choose Early and Align Docs

Your documentation burden depends on your market and claim.

Pick the pathway early and set the work plan accordingly.

• Industrial and food-grade. Focus on specs, traceability, and safety testing.
• Cosmetic-grade. Add claims substantiation and micro controls.
• GMP. Full GxP documentation, validation, and QA oversight from day one.

Don’t half-build GMP and call it speed.

You’ll pay twice later.

15) EHS and Biosafety: Safety Is a Schedule Risk, Not Just a Value

Environmental, Health, and Safety can make or break your throughput.

Plan it like you plan titer.

• HAZOP. Run hazard and operability studies before first pilot.
• Waste streams. Characterize, treat, and dispose legally with buffers.
• Solvent management. Fire code, ventilation, and operator exposure limits locked.
• Biosafety. Kill steps validated and documented.

One incident can end your program and your fundraising in the same week.

16) Digital Stack: LIMS, MES, ERP, and the Data Thread

Spreadsheets don’t scale to production.

You don’t need a giant system either.

Here’s a pragmatic stack that grows with you.

• LIMS. Sample tracking, methods, and results so you stop chasing notebooks.
• MES or eBR. Electronic batch records with controlled recipes and deviations.
• ERP. Materials, purchasing, and inventory for clean audits.
• Historian and analytics. Time-series capture and visualizations to learn faster.

Connect them with IDs so a lot, a batch, and a data trace are the same truth everywhere.

For more on communicating operational excellence to investors, see our blog post: The Investor Update that Gets Replies.

17) Team and Org Design: Who You Need Before You Scale

Great biomanufacturing is a team sport with clear roles and fast handoffs.

I’d hire these roles before I sign a CDMO SOW.

• Process engineer. Owns scaling criteria, utilities, and P&IDs.
• Downstream lead. Owns yield ladder and cost drivers post-harvest.
• QA lead. Owns specs, change control, and supplier qualifications.
• Supply chain lead. Owns vendor risk and alternates.
• Tech transfer lead. Owns the package and on-site execution.

If one person is wearing three of those hats, slow down and hire.

18) Pilot Plant Economics and Financing Strategy

Pilot plants are learning machines, not factories.

Budget them like experiments with a CFO’s eye.

• Learning milestones. Tie every campaign to a decision gate that changes your valuation or risk.
• Cost per decision. If a campaign doesn’t unlock a decision, don’t run it.
• Partner leverage. Share-of-wallet with a CDMO can get you priority slots when it matters.

Investors care about clear use-of-proceeds tied to process milestones and commercial proof points.

For more on aligning capital with milestones, see our blog post: Unit Economics for Deep Tech Startups.

19) Kill Switches: Red Flags to Stop Before You Spend Millions

Saying no is part of readiness.

I keep a list of kill switches that trigger a pause or pivot.

• Unstable TRY after three consecutive runs. If you can’t repeat, you can’t scale.
• DSP cost overwhelms upstream wins. If polishing eats half your price, rethink spec or product format.
• No customer validation path. If a buyer can’t test and accept your product, you don’t have a business.
• Supply risk you can’t hedge. If a single-source input controls your schedule, redesign now.

Kill switches save companies.

Use them without emotion.

20) Go-to-Market Meets Scale-Up: The Weekly Sync That Changes Everything

When process and GTM work in two separate rooms, months get lost.

I run a weekly 45-minute readiness sync with three standing agenda items.

• Spec alignment. What the market wants vs what the process can deliver today.
• Sample and pilot use. Who is testing what, when, and how feedback loops to development.
• Pricing gates. What price and MOQ the market will accept at current COGS and lead time.

This meeting turns experiments into revenue faster than any other process I’ve seen.

For more on fundraising enablement, see our blog post: AI for Fundraising: Tools and Playbooks.

The Readiness Checklist: A One-Page Summary You Can Run Weekly

I keep readiness visible as a single-page checklist that the team updates weekly.

• Product spec. Locked and validated with two external customers.
• Upstream. Three green runs in a scale-down model with defined scale-up criteria.
• Downstream. Yield ladder and unit operations locked to meet spec.
• PAT and data. Historian on, golden profiles defined, alarms tuned.
• COGS. On-model at planned volumes with raw material alternates qualified.
• Supply chain. Dual-sourced criticals and lead times inside plan.
• Tech transfer. Package complete, site fit confirmed, change control defined.
• Regulatory. Quality system and documentation aligned to market.
• EHS. HAZOP complete, waste plan validated.
• Capital. Runway covers next two decision gates with buffer.

Green means go, yellow means debate, red means stop and fix.

Stories from the Trenches: Two Patterns You Can Use Tomorrow

Pattern one is the antifoam tax.

A team I worked with doubled titer by adding an aggressive antifoam regime.

DSP crashed because the product partitioned into micelles and the filter train blinded early.

We ran a quick DOE upstream to trade a small titer loss for a massive DSP cost reduction and ended ahead on COGS.

Pattern two is the heat removal surprise.

At 50 liters, exotherm looked harmless.

At 1000 liters, the jacket hit limits during peak growth and the batch wandered off spec.

The fix was to slow the feed profile and adjust kLa targeting to curb oxygen demand peaks while staying inside temperature control limits.

Both wins came from aligning upstream and downstream with simple scale-aware models, not hero experiments.

Financing the Path: Capital Efficiency for Biomanufacturing

I coach teams to raise for learning, not for the press release.

Map your capital to two or three irreversible decisions that change your risk profile.

• Decision A. Process reproducibility at pilot scale.
• Decision B. Customer validation on production-representative samples.
• Decision C. COGS on-model with signed MOUs or POs.

Each decision is a fundraising milestone with a new story and a higher probability of success.

For more on non-dilutive options, see our blog post: Venture Debt for Startups: A Practical Guide.

Common Mistakes and Fast Fixes

I’ve made most of these mistakes and watched others make the rest.

• Mistake. Letting research-grade variability into production-line experiments. Fix. Define release criteria for every input, even at bench scale.
• Mistake. Ignoring cleaning and turnaround times in capacity models. Fix. Model CIP/SIP cycles and downtime honestly.
• Mistake. Over-relying on affinity capture. Fix. Explore precipitation, ion exchange, or crystallization if the market tolerates technical grade.
• Mistake. Shipping beta samples that don’t match future spec. Fix. Only send production-representative material or label it explicitly as exploratory.
• Mistake. Underestimating analytics burden. Fix. Staff QC early and automate where possible.

Generative AI Assist for Readiness

I use AI as a force multiplier, not a replacement for domain expertise.

• Protocol normalization. Standardize SOPs and highlight step conflicts.
• COGS scenario sweeps. Rapid what-ifs with parameter sweeps tied to TRY and yield ladders.
• Deviations triage. Summarize batch records and suggest likely root causes tiered by impact.
• Vendor intelligence. Draft supplier comparison matrices to review with humans.

AI helps you move faster if you feed it clean data and keep a human in the loop for plant realities.

The Biomanufacturing Readiness Score: A Simple Self-Test

Score yourself from 0 to 2 on each line and add them up.

• Product spec locked and validated.
• Upstream reproducible at scale-down.
• Downstream yield ladder locked.
• Defined scale-up criteria.
• PAT and historian in place.
• COGS model with three scenarios.
• Supply chain dual-sourced.
• Regulatory path aligned.
• EHS cleared.
• Capital covers next two gates.

18–20 is investable now.

12–17 is close with targeted fixes.

Below 12 means pause, focus, and rebuild the plan.

FAQs

What is biomanufacturing readiness.

It’s the point where your process is reproducible, your cost model works, and your customers can accept your product.

Why is David Friedberg associated with this topic.

He consistently highlights the economics and discipline required to make biomanufacturing commercially viable, which inspired this practical checklist.

What TRY targets do I need before pilot.

Set TRY tied to COGS and price, then run three consecutive green runs in a scale-down model to validate readiness.

How do I choose scale-up criteria.

Match constraints to biology and equipment, usually kLa or OTR for oxygen demand, plus P/V, mixing time, and heat removal capacity.

When should I lock downstream steps.

As soon as you define the product spec and run a yield ladder that meets cost and quality in three independent runs.

Do I need GMP to start.

Not unless your market requires it. Choose the regulatory path early and align your documentation to that path.

What’s a reasonable pilot success metric.

Reproducible batches that match scale-down profiles, on-spec output, and learning that de-risks production scale economics.

How do I build a credible COGS model.

Use three tabs for COGS, capacity, and CAPEX, with base, upside, and downside scenarios tied to measured TRY and yields.

When should I engage a CDMO.

After you can document your process narrative, CPPs, ranges, analytics, and supply plan, and after a facility fit check.

What are the top kill switches.

Irreproducible TRY, DSP cost dominance, no customer validation path, and unhedgeable single-source inputs.

Conclusion

Biomanufacturing only works at scale when product clarity, process reproducibility, and unit economics converge, and the David Friedberg checklist helps you force that convergence before you spend big.

Use this readiness checklist to align your fermentation, downstream, COGS, and financing strategy, and you’ll move from lab to scale with fewer surprises and stronger leverage with partners and investors.

If you remember one thing, it’s this: measure what matters, run the hard scale-down model, and let the numbers tell you when you’re ready to scale like David Friedberg would insist.

Subscribe to Capitaly.vc Substack (https://capitaly.substack.com/) to raise capital at the speed of AI.