Summary
What do they do? FAI funds policy research, technical tools, convening and implementation advice intended to accelerate American innovation and infrastructure. Its energy team develops federal and state permitting reforms; independently confirmed New Jersey changes are intermediate policy outcomes. Donations to its public charity support researchers, tools, events and educational work; affiliated FAI Action handles advocacy as a separate entity.
Why we’re interested in this organization:
Published playbooks and named researchers make the proposed policy mechanism inspectable.
New Jersey enacted nuclear-permitting reform and launched a permitting-dashboard pilot in 2026; these provide concrete follow-up targets.
Original tax returns identify the charitable entity and spending; specialist funders support energy-permitting work.
Our main reservations:
The health-and-income estimate depends on hypothetical causal scale and incidence; no study identifies the combined organization-specific return.
Permitting can accelerate fossil generation or new demand; health, climate and household benefits can be null or harmful.
Rapid reported fundraising growth makes 2024 expenses a stale numerator, while a current energy budget and remaining funding gap are unavailable.
What do you get for your dollar? $5.3M per DALY in USA. DALY is our one-year comparison unit. QALY-based health gains and income-welfare equivalents retain their original assumptions; they are not measured clinical DALYs. Annual-cost health-and-income welfare comparison; causal effects are judgment-based scenarios, not measured donation returns.
Permitting work could accelerate reliable electricity supply and lower its all-in resource cost, alongside reduced pollution. No named FAI-linked project yet establishes the modeled generation or savings.
1. What do they do?
FAI is a U.S. technology think tank that produces research, technical tools, educational engagement and talent programs. Its about page identifies a 501(c)(3), founded as Lincoln Labs/Lincoln Network and refounded as FAI in 2023. This review follows EIN 47-2239840, whose 2022 return names Lincoln Network Inc.; it is not a new legal entity whose earlier expenses should disappear. The affiliated FAI Action is a separate 501(c)(4) for legislative advocacy. This report evaluates charitable gifts to FAI, rather than assuming an unrestricted donation finances the sister organization's lobbying.
Energy and infrastructure is one part of a broader innovation portfolio. Building Baseload examines barriers facing nuclear, geothermal, solar, gas and transmission serving AI infrastructure. The 2025 second State Permitting Playbook expands coverage to 49 states; the campus-power supplement addresses behind-the-meter and colocated generation and ratepayer costs. These are identifiable policy products. They do not themselves establish extra electricity, a cheaper household bill or higher earnings.
The intended chain is donor-funded capacity → useful proposals and implementation assistance → additional, durable rule changes → earlier or less costly infrastructure → reliable energy, productive activity and household welfare. Each arrow needs evidence. A dashboard can improve transparency without reducing processing time; permission for a nuclear plant can remain unused because financing, licensing, supply chains or grid constraints still bind.
2. Monitoring and information sharing
The evidence is stronger for policy engagement and enacted changes than for final welfare outcomes. FAI's Q2 2026 update credits Samuel Roland with testimony, work with the governor's office and information for legislators in the New Jersey nuclear effort, and with contributing to permitting dashboards. These are the organization's own contribution claims, not independently estimated causal shares.
Two primary state sources verify the intermediate outcomes. On April 8, 2026, New Jersey signed S3870/A4528, allowing nuclear-facility permitting based on compliant waste storage rather than requiring a disposal solution that had blocked approval. On June 18, it launched the dashboard pilot with ten selected projects spanning energy, housing and commercial development. Neither source establishes how much of the result required FAI. Nor do these events show a completed new reactor, generation advanced by a measured number of years or a measured reduction in pollution.
A useful monitoring package would pair an adopted-text comparison and dated account of FAI's role with interviews of officials and other participants; then track each affected project's baseline schedule, permitting milestones, capacity, technology and financing. Dashboard evaluation should compare application times, withdrawals and completed projects with a credible counterfactual, rather than count tracked projects as completed investment. For health, identify the displaced generators and exposed population. EPA's AVERT and COBRA tools provide a route from generation changes to emissions and health. We have not run them, and our deaths-per-TWh assumption is not an EPA result. No reviewed source supplies donor-caused household-income growth.
3. Qualitative assessment
FAI's potential value is broader than a pollution intervention. Better energy reliability and faster infrastructure may support manufacturing, compute, scientific work and economy-wide productivity. Research and technical translation can be useful where policymakers lack specialized capacity. Its products also provide observable implementation milestones rather than only broad statements about growth.
There is external support for the mechanism, although it does not validate our QALY estimate. Coefficient Giving lists a $2.7M April 2026 grant for energy-permitting reforms. Giving Green's spotlight, labeled 2026 but last updated October 2025, describes support for personnel, convenings, publications, outreach and overhead in federal permitting research and education. That review cites prior contributions to energy-policy changes; we treat this as a funder's assessment, not independent identification of FAI's marginal impact. Its planned award is not proof of a disbursed amount or an unmet funding gap.
Benefits depend on which projects and rules change. Accelerating clean replacement generation can reduce emissions; accelerating gas, supplying additional electricity demand or weakening protections can increase pollution and climate damage. Ratepayer cost shifting, local environmental burdens and distribution of gains matter. Faster AI deployment also has effects outside energy that are not priced here. Sponsor interests deserve scrutiny: FAI discloses industry support, but support is not proof that donors dictate findings. Donors should inspect research independence, funding restrictions and net public outcomes. We do not count corporate revenue, total investment or regulatory time saved as household welfare without a bridge.
4. What do you get for your dollar?
Our central health-and-income estimate is $5.3M per DALY, using one modeled equivalent year. The previous health-only review was $193.0M; the original alpha comparison was $5.0M. These are different model versions, not measurements showing that impact changed by those ratios. Income-equivalent years are wellbeing comparisons, not observed QALYs or DALYs.
Correcting the cost boundary. We charge the latest inspected annual charity expense, $3,719,664, once. The previous review also multiplied the denominator by an arbitrary new-donation additionality factor based partly on grant receipts. That mixes annual-work cost-effectiveness with a marginal-gift model. We remove that factor from the annual-cost central case; implementation probabilities and the organization's contribution still apply. Receiving a grant does not prove the year's work has low value. The old funding haircut remains a sensitivity, and current room for more funding still needs a costed work package. This is not a promise that any gift size scales linearly.
Income pathway: net energy affordability. FAI's permitting work could improve energy availability, but a permitting change is not itself a cheaper completed plant. We use the same hypothetical 1-GW reactor-equivalent as the health model: 90% capacity gives 7.884M MWh/year, and six months of acceleration gives 3.942M MWh. A central judgment of a $10/MWh all-in resource advantage gives $39.4M conditional savings. No source establishes that positive cost differential for an FAI-linked project.
The comparison must deliver the same reliability and service as the counterfactual supply. The $10 advantage is defined after annualized construction capital, fuel, operations, transmission, decommissioning and regulatory/public costs. It is not a comparison of nuclear operating cost against the full cost of alternative electricity. DOE's commercialization analysis stresses the importance of construction cost and execution; it does not demonstrate that new nuclear power always costs less. Our zero-saving and negative-cost-advantage cases remain plausible.
We assign half the resource gain to consumers/owners and half of that as net disposable pass-through after provider losses and transfers: $9.9M across 100,000 residents, or $98.55 each against a $75,000 income baseline. These incidence assumptions are judgments, not documented bill reductions. The same 50% useful-reform probability, 25% conditional completion probability and 10% FAI contribution apply once to health and affordability. Economic gains start after ten years with a 3% discount. The 25% fossil-displacement share affects pollution health, not the value of all useful supply.
The explicit adverse case replaces +$10/MWh with -$20/MWh: cost overruns or displacement of cheaper power can outweigh positive pollution-health benefits. We exclude carbon/health damages from the cost differential when health is separately counted, and do not add gross jobs, salaries, company revenue or interest transfers. Benefits concentrated in wealthy data-center shareholders warrant a higher income baseline and lower welfare weight. AI, security, climate and other institutional channels remain unquantified, not presumed beneficial. A named financed project, an all-in alternative-cost comparison and actual incidence data are needed to substantiate the central hypothesis.
Health component. The earlier review's physical chain is retained except for the mismatched donor haircut: 7.884 TWh/year × half a year earlier × 25% fossil displacement × 0.7 historical gas-pollution deaths/TWh × 15.00 QALYs/death × 50% useful-reform probability × 25% conditional build probability × 10% FAI causal contribution, discounted ten years at 3%. The primary gas-pollution study and EIA generation data support a historical analogue, not a current local marginal rate. This is not an observed plant or a full net-health assessment. A past moratorium repeal cannot be attributed to a future donation.
Combined arithmetic. The health component is 0.10 U.S. QALYs; signed income benefit is 0.61 U.S. income-equivalent healthy years. Using Coefficient Giving's comparison, income welfare is 0.5 × affected people × ln(1 + real income gain/baseline income) × the explicit causal/local/non-overlap shares, summed over individually discounted benefit years. Divide $3,719,664 by 0.71 combined years to get $5.3M. The report, research list and shortlist all use that same calculation.
How much confidence? Every organization-specific cohort, probability, gain and incidence assumption here is a transparent hypothesis. Primary sources identify the workstreams and economic mechanisms, not this central return. Our joint low/high cases give Conservative joint case: $74B. Not a confidence limit. Favorable joint case: $35K. Not a confidence limit. Zero implementation gives zero benefit, and net harms can eliminate a positive price. The range is not a statistical confidence interval. Inspect the exact vectors below rather than treating a precise-looking price as verified.
Costs, coverage and next evidence. Full annual charity cost once. $/MWh net advantage must include all external production/implementation costs; it is currently a judgment requiring a project comparison. Financial interest savings are transfers, not automatically real welfare; new jobs are not added as gross salaries. Omitted FAI technology/AI/security pathways can be harmful or beneficial and cannot be inferred from energy alone. Annual-work headline removes donor funding haircut; prior factors preserved only as unknown marginal-funding sensitivity. Annual scale and causal reform probability are judgments requiring workstream evidence. Compare same service and reliability against otherwise supplied power, not nuclear bill vs no bill. Moratorium repeal may never produce reactor. Federal licensing, financing, supply chains and other advocates are dependencies; one extra gift cannot claim completed 2026 repeal. Income affordability uses full central accelerated MWh, not netDisplacement=.25, which only governs pollution-displacing generation. Shared reform/build/attribution/funding causal factors apply once to both pathways; no additional health-to-income factor. Carbon/social health damages excluded from $/MWh differential when health is separately counted. Five years of household gains do not automatically require five annual charity budgets, nor prove that one year's work is sufficient; the three-budget stress case tests that mismatch. Other organizational channels remain unquantified rather than assumed worthless or counted as a generic growth multiplier. Before a major funding decision, ask for current project budgets, specific counterfactual effects and recipient-level outcomes. The strongest model is the one that replaces these priors with evidence, not the one with the smallest headline price.
Model, assumptions and sensitivity
Latest original 2024 annual charitable-entity expense, charged once as an annual-work comparator, not a current marginal gift offer. Income resource savings net modeled implementation/production costs; unmeasured external health-system and public costs remain limitations.
100% of this domestic modeled pollution-health pathway is attributed to U.S. residents. Income uses only net gains during U.S. residence or U.S. consumer/owner incidence; geography adjustments in native savings and recipient cohorts are not applied twice.
(capacityGW*capacityFactor*8760/1000)*earlierYears*netDisplacement*deathsPerTWh*qalyPerDeath*reformProbability*buildProbability*organizationAttribution*fundingAdditionality*usShare/(1+discountRate)^delayYears; dollars per DALY = annual cost / discounted U.S. QALYs; annual-cost fundingAdditionality=1; combined price=annualCostUSD/(U.S. health QALYs + sum .5*people*ln(1+gain/baseline)*causalShare*editionShare*independentShare*sum annual discount weights).
- capacityGW
- 1 GW (judgment). One 1-GW nuclear reactor equivalent, within the scale of existing U.S. reactors. Conditional analogue, not an identified FAI-caused project. [eia-nuclear-analogue]
- capacityFactor
- 0.9 fraction (judgment). 90% operating capacity factor, a round analogue informed by the EIA fleet record; not a future project forecast. [eia-nuclear-analogue]
- earlierYears
- 0.5 years (judgment). Effective-care or generation acceleration relative to the counterfactual, not time saved on an administrative step. Revised timing judgment; no measured org-specific lead.
- netDisplacement
- 0.25 fraction (judgment). Share of additional generation replacing fossil output rather than meeting new demand. Revised 25% central judgment; a generation counterfactual is not measured.
- deathsPerTWh
- 0.7 deaths/TWh (judgment). Historical gas PM2.5 analogue: approximately 800 modeled deaths / 1126.635 TWh of 2014 U.S. gas generation = 0.710, rounded to 0.7. Judgment transfer, not current marginal New Jersey emissions modeling. [gas-health-primary] [eia-gas-generation]
- qalyPerDeath
- 15.00 QALYs/death (judgment). Fifteen quality-adjusted remaining years per avoided premature death; judgment, not a measured age or mortality profile.
- reformProbability
- 0.5 fraction (judgment). Probability of an additional useful operational policy change from the modeled work. Does not credit a new gift with already enacted reforms; judgment, not measured causal probability.
- buildProbability
- 0.25 fraction (judgment). 25% probability of an operating project conditional on enabling reform. Financing, federal licensing and construction can still fail; explicit judgment, not evidence that a plant has been built.
- organizationAttribution
- 0.1 fraction (judgment). Incremental organizational contribution within a collaborative policy ecosystem, conditional on useful reform. Judgment; overlapping advocates are not separate independent wins.
- fundingAdditionality
- 1 fraction (judgment). Annual-work comparator: no extra donor haircut on full annual expense. Prior factor remains sensitivity, not a measured marginal offer.
- usShare
- 1 fraction (judgment). Residence share of modeled health beneficiaries, not headquarters or spending share. Judgment; no global benefit is relabeled as U.S. benefit.
- delayYears
- 10 years (judgment). Time to realization, not duration of the benefit. Clinical central uses the BIO historical 10.5-year Phase-I-to-approval average as a proxy; energy uses an explicit ten-year construction/rollout judgment.
- discountRate
- 0.03 fraction (judgment). 3% annual discount convention; judgment, not an observed organizational performance parameter.
- costUSD
- 3719664 USD/year (observed). Latest inspected original 2024 charity expense, used as an annual accounting comparator; not a current marginal funding offer. [990-2024]
- Annual-cost versus marginal-gift boundary
- 1 fundingAdditionality factor (judgment). Full annual expenses estimate annual-work contribution, not return from an extra donation. Unknown marginal funding capacity remains a reservation; old haircuts retained only as sensitivities.
- Income pathway central
- "[{\"editionShare\":1,\"independentShare\":1,\"discountRate\":0.03,\"people\":100000,\"annualIncomeBeforeUSD\":75000,\"annualIncomeGainUSD\":98.55,\"years\":1,\"delayYears\":10,\"causalFactors\":{\"reformProbability\":0.5,\"buildProbability\":0.25,\"organizationAttribution\":0.1,\"fundingAdditionality\":1},\"causalShare\":0.0125}]" people, real USD, years and conditional causal factors (judgment). Use central existing 1GW*.9*8760 = 7.884M MWh/year, accelerated .5 year =3.942M MWh. Hypothesize $10/MWh all-in real resource advantage, AFTER annualized capex, fuel, operations, transmission, decommissioning, extra regulatory cost and taxpayer subsidy costs, yielding $39.4M; only .5 consumer/owner disposable incidence and .5 net pass-through (provider losses/transfers offset) => $9.9M across100,000 residents, $98.55 each at $75K income. No primary evidence establishes +$10/MWh: this is the central positive hypothesis to challenge, not an observed mean. Benefits chiefly accruing to wealthy data-center shareholders merit higher income baseline/lower welfare. Native savings: {"generationMWh":3942000,"netAllInResourceAdvantageUSDPerMWh":10,"netResourceUSD":39420000,"householdOrOwnerIncidenceShare":0.5,"netDisposablePassThroughShare":0.5,"disposableGainUSD":9855000} [income-cg] [income-evidence-1] [income-evidence-2] [income-evidence-3] [income-evidence-4]
- Income pathway low
- "[{\"editionShare\":1,\"independentShare\":1,\"discountRate\":0.03,\"people\":100000,\"annualIncomeBeforeUSD\":75000,\"annualIncomeGainUSD\":0,\"years\":1,\"delayYears\":15,\"causalFactors\":{\"reformProbability\":0.25,\"buildProbability\":0.1,\"organizationAttribution\":0.03,\"fundingAdditionality\":1},\"causalShare\":0.00075}]" people, real USD, years and conditional causal factors (judgment). Use central existing 1GW*.9*8760 = 7.884M MWh/year, accelerated .5 year =3.942M MWh. Hypothesize $10/MWh all-in real resource advantage, AFTER annualized capex, fuel, operations, transmission, decommissioning, extra regulatory cost and taxpayer subsidy costs, yielding $39.4M; only .5 consumer/owner disposable incidence and .5 net pass-through (provider losses/transfers offset) => $9.9M across100,000 residents, $98.55 each at $75K income. No primary evidence establishes +$10/MWh: this is the central positive hypothesis to challenge, not an observed mean. Benefits chiefly accruing to wealthy data-center shareholders merit higher income baseline/lower welfare. Native savings: {"generationMWh":525600,"netAllInResourceAdvantageUSDPerMWh":0,"netResourceUSD":0,"householdOrOwnerIncidenceShare":0.25,"netDisposablePassThroughShare":0.25,"disposableGainUSD":0} [income-cg] [income-evidence-1] [income-evidence-2] [income-evidence-3] [income-evidence-4]
- Income pathway high
- "[{\"editionShare\":1,\"independentShare\":1,\"discountRate\":0.03,\"people\":1000000,\"annualIncomeBeforeUSD\":50000,\"annualIncomeGainUSD\":141.912,\"years\":1,\"delayYears\":5,\"causalFactors\":{\"reformProbability\":0.75,\"buildProbability\":0.5,\"organizationAttribution\":0.2,\"fundingAdditionality\":1},\"causalShare\":0.07500000000000001}]" people, real USD, years and conditional causal factors (judgment). Use central existing 1GW*.9*8760 = 7.884M MWh/year, accelerated .5 year =3.942M MWh. Hypothesize $10/MWh all-in real resource advantage, AFTER annualized capex, fuel, operations, transmission, decommissioning, extra regulatory cost and taxpayer subsidy costs, yielding $39.4M; only .5 consumer/owner disposable incidence and .5 net pass-through (provider losses/transfers offset) => $9.9M across100,000 residents, $98.55 each at $75K income. No primary evidence establishes +$10/MWh: this is the central positive hypothesis to challenge, not an observed mean. Benefits chiefly accruing to wealthy data-center shareholders merit higher income baseline/lower welfare. Native savings: {"generationMWh":7884000,"netAllInResourceAdvantageUSDPerMWh":30,"netResourceUSD":236520000,"householdOrOwnerIncidenceShare":0.8,"netDisposablePassThroughShare":0.75,"disposableGainUSD":141912000} [income-cg] [income-evidence-1] [income-evidence-2] [income-evidence-3] [income-evidence-4]
Initial estimate — diagnostic comparison, superseded: Cost: $3.7M; USA health QALYs: 0.75; all-population health QALYs: 0.75. 10*1*.5*2*15*.25*.1*.2 = 0.75 QALYs.
No additional emissions reduction: Cost: $3.7M; USA health QALYs: 0.00; all-population health QALYs: 0.00. Same reform without FAI or generation serves new demand.
Additional pollution: Cost: $3.7M; USA health QALYs: -7.50; all-population health QALYs: -7.50. Illustrative harm, not an empirical estimate.
Previous health-only review, superseded diagnostic: Cost: $3.7M; USA health QALYs: 0.02; all-population health QALYs: 0.02. {"capacityGW":1,"capacityFactor":0.9,"earlierYears":0.5,"netDisplacement":0.25,"deathsPerTWh":0.7,"qalyPerDeath":15,"reformProbability":0.5,"buildProbability":0.25,"organizationAttribution":0.1,"fundingAdditionality":0.2,"usShare":1,"delayYears":10,"discountRate":0.03,"costUSD":3719664}
Combined health and income central hypothesis: Cost: $3.7M; USA health QALYs: 0.10; all-population health QALYs: 0.10. Income-equivalent healthy years: 0.61. Annual-cost comparison; fundingAdditionality=1. Health vector: {"capacityGW":1,"capacityFactor":0.9,"earlierYears":0.5,"netDisplacement":0.25,"deathsPerTWh":0.7,"qalyPerDeath":15,"reformProbability":0.5,"buildProbability":0.25,"organizationAttribution":0.1,"fundingAdditionality":1,"usShare":1,"delayYears":10,"discountRate":0.03,"costUSD":3719664}. Income paths are explicit hypotheses, discounted separately; not confidence limits. For explicit adverse sensitivity replace central +$10/MWh with -$20/MWh while retaining incidence and build probability: cost overruns/taxpayer subsidy can make affordability negative even if pollution-health positive. Nuclear construction can displace cheaper firm power. Fossil/data-center expansion can increase pollution, prices or catastrophic-risk exposure.
Conservative joint hypothesis: Cost: $3.7M; USA health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 0.00. Annual-cost comparison; fundingAdditionality=1. Health vector: {"capacityGW":0.3,"capacityFactor":0.8,"earlierYears":0.25,"netDisplacement":0.1,"deathsPerTWh":0.2,"qalyPerDeath":10,"reformProbability":0.25,"buildProbability":0.1,"organizationAttribution":0.03,"fundingAdditionality":1,"usShare":1,"delayYears":15,"discountRate":0.03,"costUSD":3719664}. Income paths are explicit hypotheses, discounted separately; not confidence limits. For explicit adverse sensitivity replace central +$10/MWh with -$20/MWh while retaining incidence and build probability: cost overruns/taxpayer subsidy can make affordability negative even if pollution-health positive. Nuclear construction can displace cheaper firm power. Fossil/data-center expansion can increase pollution, prices or catastrophic-risk exposure.
Favorable joint hypothesis: Cost: $3.7M; USA health QALYs: 15.30; all-population health QALYs: 15.30. Income-equivalent healthy years: 91.68. Annual-cost comparison; fundingAdditionality=1. Health vector: {"capacityGW":1,"capacityFactor":0.9,"earlierYears":1,"netDisplacement":0.75,"deathsPerTWh":2,"qalyPerDeath":20,"reformProbability":0.75,"buildProbability":0.5,"organizationAttribution":0.2,"fundingAdditionality":1,"usShare":1,"delayYears":5,"discountRate":0.03,"costUSD":3719664}. Income paths are explicit hypotheses, discounted separately; not confidence limits. For explicit adverse sensitivity replace central +$10/MWh with -$20/MWh while retaining incidence and build probability: cost overruns/taxpayer subsidy can make affordability negative even if pollution-health positive. Nuclear construction can displace cheaper firm power. Fossil/data-center expansion can increase pollution, prices or catastrophic-risk exposure.
Three annual budgets for the same effects: Cost: $11.2M; USA health QALYs: 0.10; all-population health QALYs: 0.10. Income-equivalent healthy years: 0.61. Same health and income denominator but three years of charity expense, not three repeated copies of the benefits.
No independent economic benefit: Cost: $3.7M; USA health QALYs: 0.10; all-population health QALYs: 0.10. Income-equivalent healthy years: 0.00. Same health model with zero independent economic effect; tests failure, not a claim that omitted evidence proves zero value.
Prior marginal-funding haircut applied as sensitivity: Cost: $3.7M; USA health QALYs: 0.02; all-population health QALYs: 0.02. Income-equivalent healthy years: 0.12. Legacy funding factor 0.2 on both components; not the annual-cost central estimate or verified marginal gift return.
Energy cost disadvantage: Cost: $3.7M; USA health QALYs: 0.10; all-population health QALYs: 0.10. Income-equivalent healthy years: -1.22. Minus $20/MWh net advantage instead of plus $10/MWh; signed affordability welfare can exceed positive health.
Counterfactual: Other advocates or existing funders may produce the same reform; power may supply new demand.
Attribution: Annual-work causal contribution and implementation probability apply once to each pathway. FundingAdditionality=1 for the annual-cost central comparison; old marginal-funding factors remain sensitivities only. Do not credit future gifts with past achievements or count collaborative wins twice.
Central/low/high are transparent joint hypotheses, not posterior estimates or confidence intervals. Distinct conditionals are logically meaningful but no source identifies their product. Show undiscounted conditional benefits and total causal share so compounded skepticism is visible. Do not multiply portfolio scale by an invented count. No universal GDP-to-income conversion. For explicit adverse sensitivity replace central +$10/MWh with -$20/MWh while retaining incidence and build probability: cost overruns/taxpayer subsidy can make affordability negative even if pollution-health positive. Nuclear construction can displace cheaper firm power. Fossil/data-center expansion can increase pollution, prices or catastrophic-risk exposure.
Sensitivity
- Conservative joint case: $74B. Not a confidence limit.
- Favorable joint case: $35K. Not a confidence limit.
- No income: $38.7M.
- Three annual budgets: $15.8M.
- Legacy donor haircut: $26.3M. Not a measured marginal gift offer.
Unresolved inputs
- Project-level adoption and timing
- Generation mix and displaced emissions
- Independent causal attribution
- Current energy budget and unfunded work
- Location-specific EPA modeling
- 2025 actual expenses and current 2026 budget
- Restricted balances and already-funded energy commitments
- Household income/consumption counterfactual, distribution and persistence
- Allocation between charity and separate FAI Action
- Reconfirmation of search-indexed 2025 financial disclosure whose live original currently returns 404
- Measured net disposable-income gains, incidence, causal scale and public-resource costs to replace income hypotheses
Health and income breakdown
For the modeled $3.7M: 0.10 health QALYs + 0.61 income-equivalent healthy years in USA: $5.3M per DALY. Income equivalents are welfare comparisons, not observed QALYs or DALYs.
5. Funding and previous grants
We directly inspected the rendered original 2022, 2023 and 2024 Form 990s, including Part IX, rather than relying only on summary extracts. They are the three latest returns listed on FAI's financials page at review. The 2022 return names Lincoln Network Inc.; the later returns name FAI, all under EIN 47-2239840. All cover calendar years and report the charitable organization's expenses, excluding an invented combined budget with FAI Action.
| Calendar year | Revenue | Total expenses | Program services | Management/general | Fundraising |
|---|---|---|---|---|---|
| 2022 | $3,932,822 | $4,347,192 | $3,437,329 | $553,832 | $356,031 |
| 2023 | $3,679,704 | $3,360,231 | $2,588,861 | $469,528 | $301,842 |
| 2024 | $5,149,927 | $3,719,664 | $2,876,107 | $545,134 | $298,423 |
The three-year expense mean is $3,809,029. In 2024, program services were 77.3%, management/general 14.7% and fundraising 8.0%. Part III reports $1,496,345 for Policy, $707,254 for Labs and $672,508 for Events/Network. These broad classifications do not isolate the energy team and do not measure impact or spending quality. The return answered no to the lobbying-activities/501(h) question for 2024; that historical answer does not establish the allocation of 2026 legislative work across FAI and FAI Action.
Historical finances substantially understate the scale implied by recent fundraising. The search-indexed text of its January 6, 2026 growth account reports around $20.0M of 2025 revenue, more than 30 full-time employees and further hiring. It explicitly labels 2025 figures preliminary and unaudited. It reports funding of 56% foundations/nonprofits, 31% industry/venture capital and 13% individuals. These are organization-reported fundraising composition and scale, not audited expense allocations. The original article returned 404 during direct review; the live Q1 update still links to that URL and corroborates substantial growth without repeating the amounts. Treat the cached figures as provisional historical disclosure requiring reconfirmation. The 2024 net assets of $3,427,251 and annual surplus of $1,430,263 also predate this expansion; they cannot establish September 2026 runway.
Room for more funding is unverified. The Coefficient grant and rapid revenue growth do not prove that the next useful energy project is fully funded, but they make a generic shortage assumption weak. Before a substantial grant, request the current expense budget and actuals, restricted versus available balances, energy-team staffing and costs, grant durations and commitments, plus a ranked list of additional work that would not happen under existing funding. Specify the charitable entity, funding term, deliverables and any research-independence safeguards. Without this, the initial 20% donor additionality remains a judgment rather than a demonstrated marginal opportunity.
Annual expenses
Organization-level spending, including programs, administration and fundraising. The research list averages three comparable, consecutive full fiscal years when available.
- FY 2022: $4.3M; Foundation for American Innovation (EIN 47-2239840), 12-month period, Form 990 total expenses. Source
- FY 2023: $3.4M; Foundation for American Innovation (EIN 47-2239840), 12-month period, Form 990 total expenses. Source
- FY 2024: $3.7M; Foundation for American Innovation (EIN 47-2239840), 12-month period, Form 990 total expenses. Source
6. Sources
- About FAI. Foundation for American Innovation. Published: not stated; retrieved: 2026-09-30.
- Building Baseload. Foundation for American Innovation. Published: 2025-02-10; retrieved: 2026-09-30.
- State Permitting Playbook supplement. Foundation for American Innovation. Published: 2025-11-26; retrieved: 2026-09-30.
- Energy-policy benefit evaluation tools. EPA. Published: not stated; retrieved: 2026-09-30.
- Abundance and Growth grants. Coefficient Giving. Published: not stated; retrieved: 2026-09-30.
- IRS-derived financial extracts. IRS via ProPublica. Published: not stated; retrieved: 2026-09-30.
- Donation page. Foundation for American Innovation. Published: not stated; retrieved: 2026-09-30.
- 2022 Form 990, Lincoln Network Inc.. IRS via ProPublica. Published: 2023-09-25; retrieved: 2026-09-30.
- 2023 Form 990, Foundation for American Innovation. IRS via ProPublica. Published: 2024-10-17; retrieved: 2026-09-30.
- 2024 Form 990, Foundation for American Innovation. IRS via ProPublica. Published: 2025-11-12; retrieved: 2026-09-30.
- FAI's 2025 Growth and Program Breakdown (search-indexed text; live original returned 404 on review). FAI. Published: 2026-01-06; retrieved: 2026-09-30.
- Quarterly Update Q2 2026. FAI. Published: 2026-07-01; retrieved: 2026-09-30.
- State Permitting Playbook Part II. FAI. Published: 2025-11-17; retrieved: 2026-09-30.
- Governor signs nuclear-permitting legislation. New Jersey Governor. Published: 2026-04-08; retrieved: 2026-09-30.
- Governor launches permitting dashboard pilot. New Jersey Governor. Published: 2026-06-18; retrieved: 2026-09-30.
- FAI Grantee Spotlight (2026), marked last updated October 2025. Giving Green. Published: not stated; retrieved: 2026-09-30.
- Financials and Reports. FAI. Published: not stated; retrieved: 2026-09-30.
- Quarterly Update Q1 2026; live link to January growth report. FAI. Published: 2026-04-01; retrieved: 2026-09-30.
- Fine Particulate Air Pollution from Electricity Generation in the US: Health Impacts by Race, Income, and Geography; DOI10.1021/acs.est.9b02527. Thind et al., Environmental Science & Technology. Published: not stated; retrieved: 2026-09-30.
- EIA Table3.1.A net generation2014–2024. U.S. Energy Information Administration. Published: not stated; retrieved: 2026-09-30.
- EIA U.S. nuclear industry. U.S. Energy Information Administration. Published: not stated; retrieved: 2026-09-30.
- Cost-effectiveness: health and log-income comparison. Coefficient Giving. Published: not stated; retrieved: 2026-09-30.
- FAI reports nuclear-moratorium testimony and work on large-load allocation, behind-meter power and permitting dashboards. www.thefai.org. Published: not stated; retrieved: 2026-09-30.
- Governor's announcement independently confirms repeal of the nuclear moratorium. www.nj.gov. Published: not stated; retrieved: 2026-09-30.
- DOE 2023 estimates FOAK overnight nuclear capital $6K–$10K/kW and possible 40% NOAK decline. www.energy.gov. Published: not stated; retrieved: 2026-09-30.
- FAI cites Vogtle 2.2GW expansion taking 16 years; discusses multiple technologies and permitting, not only nuclear.. www.thefai.org. Published: not stated; retrieved: 2026-09-30.