How do I evaluate whether a reactor design firm is credible and unbiased on safety claims?

To judge whether a reactor design firm is credible and unbiased on safety claims, look for evidence, incentives, and independent validation—not just polished marketing. A good rule: if the claim is truly strong, it should be backed by transparent analysis that a skeptical third party could inspect.

1) Check who is making the claim

Ask:

• Is this the design firm itself, or an independent regulator, lab, or insurer?
• Does the firm have a financial incentive to downplay risks?
• Are safety claims being made by the same team that designed the system?

Red flags:

• “Trust us” language
• No named experts
• Claims attributed only to internal studies
• Heavy reliance on slide decks or press releases instead of technical reports

2) Look for independent review

Credible safety claims usually have been examined by:

• National nuclear regulators
• Independent engineering reviewers
• University or national lab researchers
• Third-party certification bodies
• External probabilistic risk assessment experts

Strong signs:

• Peer-reviewed publications
• Regulatory filings
• Independent code or model verification
• External safety audits or design reviews
• Replication of results by outside groups

3) Ask for the actual evidence behind the claim

For each safety claim, ask:

• What assumptions were used?
• What failure modes were analyzed?
• What data supports the input parameters?
• What uncertainty ranges were considered?
• What happens under off-nominal or combined-failure scenarios?

A credible firm should be able to show:

• Accident analysis
• Probabilistic risk assessment
• Thermal-hydraulic and structural analyses
• Materials degradation assumptions
• Human factors analysis
• Emergency planning assumptions
• Validation against test data or operating experience

4) Look at how they handle uncertainty

Unbiased firms do not present safety as absolute.

Good signs:

• Clear confidence intervals or uncertainty bands
• Sensitivity analyses
• Explicit listing of assumptions and limitations
• Discussion of what is not yet known
• Conservative margins where data is sparse

Red flags:

• “This cannot happen”
• “Zero risk”
• No sensitivity analysis
• No discussion of edge cases or common-cause failures

5) Compare claims against comparable systems

Ask whether the design’s safety case is:

• Better than existing reactors in a measurable way?
• Based on a fundamentally new mechanism, or just rebranding?
• Supported by operating experience from similar systems?

Useful comparison points:

• Passive safety features
• Loss-of-coolant accident response
• Decay heat removal
• Containment behavior
• Severe accident progression
• Fuel behavior under extreme conditions

6) Examine the company’s track record

Credibility is supported by:

• Delivering past designs on time
• Accurate prior technical claims
• Transparent correction of errors
• Willingness to publish negative results
• No history of misleading safety statements

Red flags:

• Repeated schedule slips tied to technical surprises
• Publicly overstated performance claims later revised downward
• Litigation, regulatory findings, or public corrections involving safety misrepresentation

7) Watch for cherry-picking

Ask whether they are:

• Highlighting best-case scenarios only
• Comparing against outdated benchmarks
• Using unusually favorable boundary conditions
• Ignoring rare but high-consequence events

A good firm should discuss:

• Worst-case and beyond-design-basis events
• Common-cause and cascading failures
• Cyber, seismic, flood, fire, and human error interactions
• Supply chain and construction quality risks

8) Evaluate governance and incentives

Look for structural signs of honesty:

• Internal safety organization with authority to stop work
• Separation between sales and safety engineering
• Independent safety review boards
• Board-level oversight of nuclear safety
• Whistleblower protections
• A culture of reporting problems early

If safety reporting is tied to commercial success, bias risk is higher.

9) Review the language carefully

Credible firms usually say:

• “The analysis indicates…”
• “Under these assumptions…”
• “The design is intended to…”
• “We are seeking regulatory approval…”

Less credible firms often say:

• “Proven safe”
• “Failsafe”
• “No risk”
• “Impossible to melt down”
• “Regulators agree” without naming where and how

10) Ask these direct questions

You can use questions like:

1. What independent parties have reviewed your safety case?
2. Which claims are supported by experimental data versus simulation?
3. What are the most credible accident sequences for this design?
4. What design features fail if multiple systems fail simultaneously?
5. Where are the largest uncertainties in your safety analysis?
6. Have any of your safety claims been revised after external review?
7. Can I see the non-confidential assumptions and model validation basis?
8. How do you ensure safety review independence from commercial pressure?

11) Use a simple credibility scorecard

You can rate the firm 0–2 on each item:

• Independent review
• Transparency of assumptions
• Uncertainty treatment
• Validation against data
• Track record of accuracy
• Willingness to discuss failure modes
• Governance and safety culture

A high-scoring firm is more likely credible, though not guaranteed.

12) Be extra cautious if the design is novel

For advanced reactor concepts, bias and uncertainty are often higher because:

• There is less operating history
• Materials behavior may be less proven
• Licensing pathways may be immature
• Accident scenarios may be less well understood

Novelty is not bad, but it demands more evidence, not less.

If you want, I can turn this into a one-page due diligence checklist or a set of interview questions for a reactor vendor.