How David Friedberg Evaluates Technical Risk: A Founder's Guide to De-Risking Bio Ventures

‍

David Friedberg put a spotlight on a hard truth most bio founders quietly worry about on day one. Technical risk can kill a great idea long before the market gets a say.

I wrote this guide to show you how to evaluate, sequence, and crush technical risk the way top investors do, using practical tools you can apply this week.

You will learn how to build a validation plan, map technology readiness levels to bio, design killer experiments, and communicate diligence-ready results with authority.

I will keep it plain, actionable, and rooted in what I have seen work for bio startups raising from sophisticated investors.

‍

How David Friedberg Evaluates Technical Risk: A Founder's Guide to De-Risking Bio Ventures

What “technical risk” means in bio ventures

Technical risk is the set of unknowns that might make your biology not work, not scale, not be repeatable, or not be economical.

It is different from product or market risk, and it precedes them in bio.

David Friedberg often frames investment risk as a stack you peel back layer by layer, and in biology the first layers are scientific validity, engineering feasibility, and process economics.

As a founder, I define technical risk in four buckets.

Proof of mechanism: Does the biology do what we think it does under controlled conditions.
Proof of repeatability: Can we reproduce the effect across runs, operators, and instruments.
Proof of scaling: Does the system behave at 10x, 100x, or 1000x without catastrophic loss of performance.
Proof of economics: Can we reach cost, quality, and yield targets that beat existing alternatives.

These four proofs create the backbone of your validation plan and your fundraising story.

Friedberg’s mental model for de-risking: sequence the stack

I think of the Friedberg approach as ruthless sequencing.

You attack the hardest assumptions first with small, fast, and decisive experiments.

Each experiment is designed to resolve a binary risk quickly and cheaply.

Here is a simple way to stack and sequence.

Physics first: Is there a plausible mechanism based on known science.
Biology next: Does the pathway, enzyme, or edit produce the expected change in a controlled assay.
Engineering constraints: Can you control key parameters at larger scale.
Economics early: Do a rough technoeconomic analysis before you spend on scale-up.

Good investors will push you to kill weak branches quickly so you can concentrate capital on what works.

Technology Readiness Levels for bio: a practical mapping

Traditional TRLs are vague for biology, so I use a bio-specific mapping investors understand.

Bio-TRL 1–2: Hypothesis and in vitro signal in a single lab setup.
Bio-TRL 3: Repeatable assay with controls and statistical power.
Bio-TRL 4: Bench-scale system integration across modules.
Bio-TRL 5: Engineering prototype in a bioreactor or preclinical model.
Bio-TRL 6: Pilot-scale or GLP preclinical package with process controls.
Bio-TRL 7: Demo-scale with defined quality attributes and release criteria.
Bio-TRL 8: Manufacturing readiness with validated process and supply chain.
Bio-TRL 9: Commercial operation meeting cost, quality, and regulatory requirements.

When I present a plan, I show current Bio-TRL, target Bio-TRL by milestone, and the experiments that bridge the gap.

Designing a validation plan investors trust

Your validation plan should make it obvious what you are trying to prove, how you will prove it, and the go or no-go criteria you will use.

I use a one-page template that every investor gets in the first data room folder.

Hypothesis: One sentence that states the scientific or engineering claim.
Experiment: The minimum test that can falsify the claim.
Metrics: Pre-registered thresholds for success and confidence.
Time and cost: Calendar time, burn, and equipment needs.
Decision: If met, what new option value is unlocked and what is the next experiment.

This clarity shortens diligence and positions you as a disciplined operator.

For more on building investor-grade materials, see our blog post: The AI-Ready Data Room for Deep Tech Diligence.

Assay to system: laddering proof from unit to field

I walk investors through proof ladders because it mirrors how engineering teams work.

Assay level: Prove the effect in a controlled, high-signal environment.
Unit level: Prove a critical step under realistic conditions.
Module level: Integrate two to three units and control variability.
System level: Run the full process at bench or pilot scale.
Field level: Validate with real use, real constraints, and real customers.

Each rung de-risks a specific failure mode and produces data you can reuse in manufacturing and regulatory filings.

Reproducibility and statistical power without wasting runway

Nothing triggers diligence alarms faster than fragile data.

I set a default rule for early bio startups.

Triplicate runs per condition at minimum, with at least two operator days.
Pre-register the analysis and effect size you need to believe the result.
Show raw data and instrument logs alongside processed figures.
Use positive and negative controls, not historical baselines only.

Simple power calculations can keep you from over- or under-spending on replication.

If you cannot reproduce a result internally, fix that before any investor meeting.

Scale-up risk: the bio equivalent of gravity

Scaling biology is where great pitches go to die, and sophisticated investors like David Friedberg look for early signs you understand this.

I frame scale-up across three axes.

Mass transfer and mixing: Oxygen, shear, and gradients change behavior.
Time constants and control: Heat removal and sensor latency matter.
Contamination and quality: Clean-in-place and bioburden become dominant risks.

Run a 1 L to 10 L to 100 L step-up plan with explicit acceptance bands for yield, titer, and productivity.

Show how you will hold critical parameters constant in dimensionless terms, not just in absolute values.

Technoeconomic analysis early and often

Investors want to see a credible path to unit economics that beat the incumbent.

You do not need perfect numbers to start, but you do need a TEA based on first-principles and vendor quotes.

Bridge from biology to dollars using yield, titer, rate, and downstream recovery.
Show the cost stack across feedstock, media, utilities, labor, and depreciation.
Plot a learning curve that links process improvements to cost reductions.

I include a tornado chart to show which parameters move the economics most.

For more on TEA fundamentals, see our blog post: TEA 101: Building the Cost Curve for Bio Manufacturing.

Data room artifacts that speed diligence

Your data room should let an investor recreate your claims in their head in under an hour.

Validation plan one-pager with current and target Bio-TRLs.
Assay protocols, SOPs, and raw datasets for key figures.
Scale-up plans with control strategies and hazard analyses.
TEA workbook with scenarios and clearly labeled assumptions.
IP landscape, filings, and freedom-to-operate notes.
Regulatory memo with target quality attributes and path.

Label files by hypothesis so diligence reviewers can map data to claims without guessing.

Milestone-based financing and kill criteria

Founders who define kill criteria earn trust.

I describe milestones as binary gates tied to explicit thresholds and dates.

Gate A: Achieve 35 percent yield at 10 L with less than 15 percent CV by Q2.
Gate B: Demonstrate 80 percent recovery in downstream at bench scale.
Gate C: Hit $7.50 per kg modeled COGS with sensitivity bounds.

Kill criteria protect your team from sunk-cost fallacy and show investors you are a steward of capital.

For more on milestone-driven raises, see our blog post: Milestone Sprints: How to Raise on Experiments, Not Hype.

IP strategy aligned to technical risk

I prioritize filings that block the highest-value part of the process or product, not the easiest claims.

Draft provisionals that cover compositions, methods, and use cases with clear enablement.

Pair IP filings with lab notebooks, timestamped data, and third-party corroboration when possible.

Investors will ask how your IP maps to unit economics advantages or regulatory exclusivity.

Regulatory path and quality by design from day one

Regulatory readiness is part of technical risk because it determines what you must prove and how you control variation.

I write a one-page Quality by Design plan showing the critical quality attributes, critical process parameters, and control strategies.

For therapeutics, outline GLP and GMP pathways and the indicative timeline to IND or pivotal studies.

For food, materials, or industrial biotech, show GRAS, novel food, or environmental permitting paths.

Manufacturing readiness and supply chain realities

Map Manufacturing Readiness Levels alongside Bio-TRLs so investors see you can make what you claim at quality and cost.

Show vendor quotes for key equipment, media, resins, and consumables with second sources.

Call out long-lead items and strategies to derisk them, like leasing pilot capacity or using a CDMO.

Include a risk register with mitigation for shortages, price swings, and regulatory changes.

Team culture for de-risking: speed with discipline

The best bio teams have two habits.

Pre-registration: State what success looks like before doing the experiment.
Postmortems: Learn fast from failed runs without blame or spin.

I hire for curiosity, comfort with controls, and the ability to communicate uncertainty clearly.

In interviews, I ask candidates to design a kill experiment for one of our claims to test their de-risking mindset.

Working with CROs and CDMOs without losing your edge

External partners can accelerate validation if you structure the work right.

Send locked protocols and acceptance criteria.
Own the assay transfer and reserve time for tech troubleshooting.
Require raw data access, not just summary reports.
Split work to reduce vendor lock-in and compare performance.

Use CROs for speed and capacity, not for thinking through first-of-kind problems you must own.

Pilot deployments and field validation

Even early, I look for ways to put the product in the wild under controlled risk.

For industrial bio, that could be a demo skid at a partner site with defined KPIs and sampling plans.

For therapeutics, it means robust preclinical models and a clear path to first-in-human with risk controls.

For bio-based materials or food, run limited launches to gather sensory, stability, and shelf data.

Risk-adjusted valuation and term sheets

Investors price rounds based on what you have de-risked and what remains.

I anchor valuation to option value created by each milestone, not to vanity comps.

Show the delta in risk after each gate and how that reduces the discount rate or increases the probability of technical success.

Transparent math makes negotiations faster and friendlier.

Communicating technical risk to investors

Great founders sound like their own toughest diligence committee.

I use a three-slide format for the technical section of the deck.

Slide 1: The claim, the controlling variables, and the data that supports it.
Slide 2: The biggest open risks, what could go wrong, and the experiment to resolve them.
Slide 3: The economic significance of success, tied to the TEA and the market.

Speak plainly, avoid hedging, and let the data carry the weight.

Common failure modes and how to avoid them

I see the same patterns sink promising bio startups.

Skipping controls to save time, which backfires during diligence.
Scaling before reproducibility is locked at bench.
Optimizing for a metric that does not move the TEA.
Underestimating contamination control and QA systems.
Confusing promising science with a manufacturable process.

Make a stop-doing list and review it weekly with your technical leads.

Quantifying uncertainty: sensitivity and scenario analysis

Investors appreciate when you quantify uncertainty instead of hand-waving it.

I run one-page sensitivity analyses on key parameters like yield, recovery, and media cost.

Then I add a simple Monte Carlo in the TEA workbook to show expected value and tail risks.

Use this to decide which experiments collapse the most uncertainty per dollar of burn.

Safety, biosafety, and ethical risk as technical risk

Include biosafety level requirements, gene drive or environmental risk considerations, and containment strategies in your plan.

For therapeutic work, address immunogenicity, off-target effects, or toxicity with appropriate models and controls.

This is not just compliance. This is part of the engineering reality investors expect you to manage.

Case-style example: the three proofs in 120 days

Here is a simple pattern I helped a bio manufacturing team run before a Seed+ raise.

Days 0–30: Achieve 2x baseline titer with a single genetic modification in triplicate, with defined QC and control conditions.
Days 31–75: Transfer to 10 L, hold kLa constant, and show less than 20 percent performance loss with three runs.
Days 76–120: Improve downstream recovery from 55 percent to 75 percent using resin screening and define a control plan.

The team then updated the TEA, cut modeled COGS by 38 percent, and closed the round at better terms because the risk story had changed.

Investor-ready narratives: from data to story

A disciplined narrative aligns the science, the engineering, and the economics.

I use this sentence to tie them together in the deck.

Because we improved yield by X and recovery by Y in a configuration we can scale with control strategy Z, our modeled cost drops below the incumbent, which unlocks market segment A with margin B.

Everything else becomes supporting detail.

Integrating AI and automation into your de-risking loop

Use design of experiments software and Bayesian optimization to explore conditions faster and with fewer runs.

Automate data capture from instruments to reduce transcription errors and accelerate learning.

Run nightly analyses that update your TEA based on the latest lab results so your team sees the economic impact in real time.

For more on AI-powered fundraising workflows, see our blog post: AI-First Fundraising: From Experiments to Term Sheets.

Your 90-day de-risking sprint checklist

Here is a tight plan you can start Monday.

Week 1: Finalize validation plan, pre-register metrics, and build the data room skeleton.
Weeks 2–4: Run assay-level proofs with triplicates and solid controls.
Weeks 5–8: Execute unit and module tests and lock reproducibility.
Weeks 9–12: Do a 10x scale step and update the TEA and risk register.
Week 13: Ship the investor update with clear gates, results, and next risks.

Then you raise on the back of truth, not talk.

How Capitaly.vc evaluates technical risk

At Capitaly.vc, we look for founders who show discipline in their validation plans and who can tie each technical win to economic impact.

We favor milestone-based financing with clear go or no-go criteria, and we love seeing TEA updates in monthly investor notes.

If you can show a credible path from Bio-TRL 3 to 6 with quantified uncertainty, you will find our diligence process fast and fair.

For more on investor expectations, see our blog post: The Technical Diligence Playbook for Bio Startups.

FAQs

How does David Friedberg think about technical versus market risk.

He emphasizes sequencing by resolving technical feasibility and economics before betting big on market scale, especially in bio where technical risk dominates early.

What is the fastest way to build a validation plan.

Write hypotheses, design minimal falsifying experiments, pre-register acceptance criteria, and tie each win to a TEA update.

How many replicates do I need to satisfy investors.

Triplicates with proper controls are a practical baseline, but use power analysis for critical claims.

When should I run a TEA.

Immediately, even if rough, and update it after each technical milestone to show economic movement.

How do I present TRLs to bio investors.

Use a bio-specific mapping and show your current level, the next gate, and the experiments to get there.

What are common red flags in technical diligence.

Irreproducible data, missing controls, overfitted models, and scaling assertions without control strategies.

How do I pick milestones for a Seed round.

Choose binary gates that remove the biggest uncertainties in mechanism, scale-up, and cost.

Should I use CROs or build in-house.

Use CROs for speed and capacity but own the core problem-solving and require raw data access.

How do I handle negative results.

Publish them internally, adjust your plan, and show investors how they saved time and money by closing dead ends.

What belongs in the data room for technical diligence.

Validation plan, SOPs, raw data, TEA, IP docs, regulatory memos, and scale-up plans with hazard analyses.

Conclusion

Investors like David Friedberg reward founders who de-risk methodically, measure honestly, and tie every technical win to economics.

If you adopt the validation plan mindset, map your Bio-TRLs, and keep your TEA live, you will shorten diligence and improve your odds of raising on great terms.

Build your story on proofs, not promises, and you will stand out in the crowded field of bio startups focused on de-risking and technical risk the way David Friedberg does.

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