The $16.6 Billion Infrastructure Gap: A 12-Claim Cross-Reference

NEXA Advisors and UAM Geomatics published As eIPP Launches, AAM Aircraft are Nearly Ready. The $16.6 Billion Infrastructure Gap Remains in March 2026 — a bottom-up forecast of the capital required to deploy Advanced Air Mobility infrastructure across 62 US metropolitan areas through 2045. The paper's structural argument is a three-pillar model: aircraft certification, operational frameworks, and infrastructure deployment. If any one pillar lags, certified aircraft enter markets without the network facilities to serve them.

AirIndex sits inside that frame. We are the live operational measurement layer for the third pillar — refreshed daily, scored per facility, identifier-keyed, and audit-chain anchored. This cross-reference dispositions twelve claims in the paper against the AirIndex substrate using four categories: validate where the substrate independently confirms the claim, refine where we add resolution, reframe where the substrate suggests a different binding constraint, and coverage gap where the paper extends past what AirIndex measures. The intent throughout is complementary, not competitive: the macro forecast and the operational substrate sit on adjacent axes.

Headline at a glance

Per-claim disposition

Claim 1 — $16.6B headline CAPEX through 2045

"Analysis conducted by UAM Geomatics indicates that through 2045 approximately $16.6 billion in infrastructure investment will be required to support peak demand across 62 key U.S. metropolitan locations…" (p.1)

Split: $11.2B ground infrastructure + $5.4B ATC/airspace modernization.

Disposition: COVERAGE GAP.

AirIndex measures live operational readiness, not 20-year capital forecasts. There is no AirIndex equivalent to the $16.6B figure to validate or contradict directly. AirIndex's contribution to this claim is resolution on the components — site count, geography, regulatory deployability — rather than a competing aggregate.

Claim 2 — 62 metros represent the AAM economic concentration

"…62 key U.S. metropolitan locations, representing roughly sixty percent of the Nation's population." (p.1)

"…62 key U.S. locations representing roughly half of the national population." (p.5, Conclusion)

Disposition: VALIDATE (with an internal-paper inconsistency flagged).

The "AAM concentrates in major metros" thesis is one AirIndex's market-coverage choices independently confirm. Of AirIndex's 25 markets, 21 sit inside Dyment's 34 named MSAs:

• In both AirIndex 25 + Dyment 34 MSAs (21): Atlanta · Boston · Chicago · Cincinnati · Columbus · Dallas · Denver · Houston · Las Vegas · Los Angeles · Miami · Minneapolis · Nashville · New York · Orlando · Phoenix · San Diego · San Francisco · Seattle · Tampa · Washington DC
• AirIndex-tracked, not in Dyment MSAs (4): Austin · Charlotte · Salt Lake City · San Antonio
• Dyment MSAs not AirIndex-tracked (13): Baltimore · Detroit · Philadelphia · Portland · Reno · San Jose · Syracuse · Wichita · Akron · Cleveland · Dayton · New Orleans · Toledo

Two of AirIndex's "extras" (Charlotte, Salt Lake City) are arguably covered by Dyment's 28 state regions (Charlotte Area, Salt Lake Area). Austin and San Antonio are not — Texas is not represented as an MSA cluster or state region in Dyment's appendix.

Internal-paper inconsistency: Page 1 says "sixty percent of the Nation's population"; page 5 says "roughly half." Small, fixable in any forthcoming revision.

Claim 3 — 1,400 remediated heliports anchor the buildout

"The buildout considered here includes 2,570 vertiports coming online by 2045, of which 1,400 are remediated sites, existing heliports converted into vertiports." (p.2)

New-build mix (1,160 sites): 600 unserviced @ $1.5–2.2M · 350 serviced @ $2.8–3.8M · 150 urban multiports @ $17–23M · 60 airport multiports @ $55–68M.

Disposition: REFINE. This is the largest single refinement opportunity in the paper. The 1,400-heliport remediation assumption is the largest line item in the $11.2B ground CAPEX. It rests on FAA dataset quality. AirIndex's substrate has multiple independent observations directly relevant:

1. FAA NASR 5010 active-status population = 5,647 heliports minted as aix:hp:* (status code O or CI). The 1,400 figure is roughly 25% of that active population — proportionate, but the population's accuracy is the question, not the share.
2. Independent cross-reference (LZControl): 14,844 helicopter landing zones, including 4,009 hospital sites versus FAA's 2,675 — roughly 1,334 more hospitals than FAA records. Per Rex Alexander (NFPA 418 Chair): "AirNav, SkyVector, and all similar systems pull directly from ADIP — they are mirrors, not independent sources. If it is wrong in ADIP you can pretty much guarantee it is wrong everywhere else." LZControl is the only independent cross-reference worth citing.
3. Practitioner observation (Rex Alexander): "The average heliport in the U.S. has some degree of obstruction penetration." FAA coordinates are approximately 54% accurate per state audits. Approach-surface integrity (the 8:1 obstruction surface) cannot be verified from ADIP alone — it requires Google Earth review plus on-site validation.
4. Capability Tagging Schema v1.0 (/methodology/capability-tagging): eight attributes that materially differentiate "active heliport" from "vertiport-remediable site" — fire suppression, battery isolation, NFPA 418 compliance status, ground power, secure perimeter, weather infrastructure, bird mitigation, EB 105A geometry. Most aren't tracked in FAA NASR 5010.

What this means for the 1,400 figure: the number of remediation candidates may be defensible — LZControl's count suggests the universe is larger, not smaller. The cost discount implied by "remediated" is the open question. Dyment notes "remediation can reduce costs by leveraging existing aviation sites and infrastructure" but doesn't quantify the discount. NFPA 418–compliant battery thermal suppression, foam capability, approach-surface remediation, and ground-power upgrades on a legacy heliport may compress the remediated-site cost closer to greenfield than the framing implies.

Claim 4 — Per-site CAPEX bands

Ground infrastructure mix and per-site costs (p.2):

• 600 unserviced vertiports @ $1.5–2.2M each
• 350 serviced vertiports @ $2.8–3.8M each
• 150 urban multiports @ $17–23M each
• 60 airport multiports @ $55–68M each

Disposition: REFINE.

The bands are reasonable on their face, and Dyment correctly notes that "local conditions of each facility (land prices and construction costs vary widely)" drive variance. The refinement opportunity is what's not explicit in the bands:

• NFPA 418 compliance status (battery thermal suppression, foam capability) — costed in or not?
• Battery isolation capability for eVTOL specifically — a category that didn't exist for legacy heliports
• Ground-power infrastructure for charging (kW-class draws on local distribution)
• Permitting-friction cost in Tier III states (see Claim 7)

Dyment's appendix says aircraft cost inputs draw on industry interviews; the per-site infrastructure cost methodology isn't decomposed in this paper. AirIndex's Capability Tagging Schema v1.0 provides the per-attribute axis on which to interrogate whether the bands assume baseline-spec, NFPA-418-compliant, or fully eVTOL-equipped facilities.

Claim 5 — eIPP regions overlap substantially with the 62 metros

"Many of the metropolitan regions participating in eIPP fall within the 62 markets identified in this infrastructure analysis." (p.4)

Disposition: VALIDATE.

AirIndex has minted aix:prog:* identifiers for federal AAM programs including DOT-EIPP. Multiple AirIndex markets are eIPP-designated; Dyment's Appendix 2 marks eIPP states (Utah, Maine, North Carolina, Virginia, Oklahoma) and eIPP-designated MSAs throughout the city table. The overlap is structural — both UAM Geomatics and AirIndex selected major metros where AAM economic activity concentrates and where eIPP is testing operational concepts.

Phoenix specifically: eIPP-designated, AirIndex MODERATE tier at 65/100 following its May 10 tier-flip on advancing state legislation (aix:mkt:F6MH5R2K; SB1819 aix:bill:6W2SA6W2; SB1457 aix:bill:MSMWIMSM).

Claim 6 — Three pillars (Aircraft / Operational / Infrastructure)

"AAM Deployment Requires Three Parallel Pillars… If infrastructure development lags behind the other two pillars, certified aircraft will enter markets that lack the network facilities required for meaningful passenger and cargo volumes…" (p.2)

Disposition: VALIDATE — and adopted as positioning frame.

This is the structural gift the paper hands the field. Pillar 3 — Infrastructure Deployment — is the layer AirIndex measures live. Adopting the three-pillar language across methodology and editorial is the natural move: the paper has named the architecture, and operational-readiness measurement is where Pillar 3 becomes legible day to day.

Claim 7 — Capital is the constraint; state regulatory recognition is implicit

The paper frames the gap as a capital deployment problem — $16.6B needed, PPPs are the path, federal loan guarantees and Private Activity Bonds enable it (p.5).

Disposition: REFRAME.

Capital is a constraint. It is not the binding constraint in every market. AirIndex's State Regulatory Spectrum (v0.5, May 14 2026, 28 states placed across the Dyment-NEXA 34-MSA footprint) identifies a four-tier framework where the binding constraint differs by tier:

For Tier IV states, capital alone doesn't solve the problem — TxDOT operates as a planning and funding body with no licensing authority over private-use heliports. Dyment's bottom-up CAPEX bands implicitly assume permitting and recognition are solved. In Tier IV states, they aren't. Dallas and Houston — two of Dyment's 34 named MSAs — sit in anchor-confidence Tier IV. The deployment risk for capital in those metros is qualitatively different from capital in Tier I metros, and the per-site CAPEX bands don't visibly reflect that variance.

Florida specifically (Tier III): FS §330.30(1), FAC Chapter 14-60. Site approval required before construction; 45-day approval validity; 2-year permit term; practitioner-reported long permitting cycles. Per-site CAPEX in Florida should reflect material schedule risk and carrying cost; Dyment's bands don't decompose by state regulatory burden.

Claim 8 — Demand-driver methodology (ten city-specific inputs)

"Passenger demand serves as the core driver… modeled using ten city-specific inputs, including airport origin-destination traffic, mobility alternatives, per capita GDP, congestion and travel distance, human capital, population density, livability, corporate presence, business aviation activity, and existing heliports." (Appendix 1, p.7)

Disposition: VALIDATE structure; note overlap.

Substantial overlap with AirIndex's seven-factor readiness scoring — population density, business aviation activity, existing heliports, and corporate presence are explicit AirIndex factors. The axes are different: AirIndex measures receiving-market readiness; Dyment measures demand to be served. The two methodologies sit on adjacent axes — one forecasts where demand will form, the other measures whether the markets that demand will route into are operationally ready to receive it.

Claim 9 — Industry inflection point after 2035

"A central assumption is an industry inflection point after 2035, when increased aircraft automation, improved airspace access, and greater interoperability reduce costs and accelerate adoption…" (Appendix 1, p.7)

Disposition: COVERAGE GAP.

AirIndex measures live present-state readiness, not 10–20-year forecasts. The 2035 inflection assumption is exogenous to AirIndex's scope. Worth noting as a fundamental forecast-vs-measurement axis difference, not a refinement target.

Claim 10 — Comparable transportation benchmarks

JFK ~$19B · LaGuardia ~$8B · Denver International >$10B · Boston Big Dig $14.6B · California HSR $88–128B+ · Brightline West ~$12B. "$16.6 billion represents a distributed, system-wide investment enabling an entirely new transportation layer…" (p.5)

Disposition: VALIDATE (framing).

Not an AirIndex data point. The framing is sound — distributed national infrastructure layer rather than single-asset megaproject.

Claim 11 — Operator revenue opportunity $210B+ over 20 years (five use cases)

"…the operator revenue opportunity for five use cases (regional air mobility, airport shuttle, on-demand air taxi, business aviation, and medical) is estimated by UAM Geomatics in the range of $210+ billion over the next 20+ years." (p.1)

Disposition: COVERAGE GAP. AirIndex doesn't model operator revenue forecasts.

Claim 12 — Public-Private Partnership financing path

Federal loan guarantees, tax-assisted instruments, Private Activity Bonds (p.5).

Disposition: COVERAGE GAP. AirIndex doesn't track financing instruments.

Where AirIndex has nothing to say (named honestly)

• 20-year capital forecasts — AirIndex measures live, not multi-decadal.
• Operator revenue opportunity sizing — not in scope.
• Financing instruments (PPPs, Private Activity Bonds, federal loan guarantees) — not tracked.
• Aircraft cost / ticket pricing models — not tracked.
• The remaining 22 states beyond the 28-state v0.5 placement in the State Regulatory Spectrum — v0.5 covers the Dyment-NEXA 34-MSA footprint; the path to v1.0 is full 50-state coverage with broader primary-source verification.
• Per-attribute capability data on facilities — the Capability Tagging Schema is reserved, not populated. Population is buyer-signal gated. The schema gives the per-site CAPEX bands a decomposition axis; it doesn't yet produce a counter-figure.

What the data tells us

The paper is structurally compatible with AirIndex's positioning. The three-pillar framing names AirIndex's pillar in print. The 62-metro footprint substantially overlaps AirIndex's 25 markets. The demand-driver methodology shares roughly four of ten factors with AirIndex's readiness scoring.

The refinements aren't contradictions. They're resolution improvements on three specific axes where AirIndex has data the macro forecast doesn't decompose:

1. Heliport remediation realism — FAA dataset integrity (LZControl's independent 14,844 LZs vs FAA's 5,647), positional accuracy (~54% per state audits), and capability-tagging attributes (NFPA 418, battery isolation, obstruction-surface remediation) that may compress the remediated-site cost discount closer to greenfield than the paper implies.
2. State regulatory deployability — Tier IV (Texas) and Tier III (Florida) states change the deployment math even when capital is available. Dallas and Houston sit in anchor-confidence Tier IV; Miami and Orlando in Tier III.
3. Per-site CAPEX decomposition — the Capability Tagging Schema gives the per-site bands an attribute axis the paper's methodology doesn't expose.

The angle the data is not suggesting:

• "Dyment is wrong about $16.6B" — we have no competing aggregate.
• "AirIndex covers more metros than NEXA" — Dyment covers 62, AirIndex covers 25.
• "AAM economic concentration is different than Dyment claims" — our market choices independently confirm it.

The angle the data is suggesting:

• The macro forecast is sound. The three places it needs operational resolution are above, and the substrate has data on each.
• Pillar 3 isn't just a capital problem in several of NEXA's 34 named MSAs — it's a state recognition problem first.
• The heliport remediation discount is the largest forecast variable; the data has something to say about it.

Sources

Primary paper: Michael J. Dyment (Managing Partner, NEXA Advisors LLC) and Benjamin Zevin (Director, UAM Geomatics Inc.), As eIPP Launches, AAM Aircraft are Nearly Ready. The $16.6 Billion Infrastructure Gap Remains (Paper 26-1.5, NEXA Advisors / UAM Geomatics, March 2026).

AirIndex substrate cited: AIX-ID System; Capability Tagging Schema v1.0; State Regulatory Spectrum v0.5 (May 14 2026, 28 states placed); FAA NASR 5010 active-status heliport population (5,647 minted as aix:hp:*); LZControl independent landing-zone reference (14,844 LZs, 4,009 hospital sites).