The cost of waiting: why modernization funding is not the same as operational resilience
A boardroom test for funding resilience before disruption sets the timetable
ARKHOS | 14 November 2026
A modernization budget appears in the accounts. The exposure created by postponing it is harder to see. That asymmetry can make a delay look financially prudent until a critical service becomes unavailable.
On 13 November, the US Federal Aviation Administration ordered a nationwide ground stop after an air-traffic-control communications disruption. The restriction was lifted after more than four hours. Transportation Secretary Sean Duffy’s initial explanation pointed to overloaded older network equipment and inconsistent data. The investigation remained open. These are reported official statements, not a completed technical root-cause analysis. [1]
We do not attribute the incident to Microsoft, AI, or a specific contractor. Nor do we calculate an economic loss by multiplying a national revenue figure by four hours. Flight cancellations across an entire day are not a clean measure of this incident’s impact, particularly when other disruptions may also be present.
The stronger investment question is already answerable: what evidence should a board require before accepting another year of exposure to systems it knows are difficult to sustain?
A documented modernization problem, not a proven explanation for this outage
In September 2024, the US Government Accountability Office reported the results of FAA’s 2023 assessment: 51 of 138 systems were unsustainable and another 54 potentially unsustainable. Together, 105 systems fell into those two categories; 58 had critical operational effects on safety and efficiency. This is a historical baseline, not a November 2026 inventory, and it does not identify the cause of this week’s incident. [2]
The distinction matters. A system can continue operating while its maintainability deteriorates: replacement parts become harder to obtain, specialist knowledge becomes scarce, and changes become harder to test. “Still working” is a statement about the present. It is not a sufficient forecast of recoverability under stress.
A board should therefore ask for a service-level exposure map, rather than a list ranked only by asset age. Which customer or operational service depends on each component? How long can it operate in degraded mode? What happens when two dependencies fail together? Who can restore it, with which parts and within what demonstrated time?
Funding is only the beginning of execution
GAO’s September 2026 audit provides a more current view. Congress appropriated $12.5 billion in July 2025 for the new air-traffic-control initiative. Phase 1 targeted December 2028. FAA reported replacing 2,560 of 5,170 old copper wires by May 2026. [3]
Yet GAO found 11,389 individual project schedules that were not integrated, and no comprehensive, well-documented lifecycle cost estimate. The audit identified a risk to execution, not a quantified prediction of this outage. Its implementation figures are dated snapshots, not today’s completion status. [3]
The management implication is broader than aviation: procurement activity, installations and budget utilization are intermediate outputs. A service becomes more resilient only when its dependencies work together, operators can use them, and recovery has been demonstrated. An integration milestone deserves its own acceptance test and accountable owner.
That also changes how acceleration should be governed. Replacing everything at once can create transition risk, while indefinite deferral preserves legacy risk. The useful choice is a sequenced migration: isolate critical dependencies, test compatibility and fallback, rehearse cutover, and retire old components only after operational acceptance. Buying newer technology without changing execution discipline can simply move the point of failure.
Modernization is a foundation, not the whole transformation
Replacing a network component restores or improves technical capability. Digital transformation changes how an organization delivers a service, makes decisions and allocates accountability using that capability. Confusing the two creates two errors: declaring transformation complete when equipment has been installed, or buying a new operating model before its essential infrastructure can support it.
The distinction is particularly important in a live, safety-critical service. GAO’s 2026 report describes earlier Newark disruptions in April–May 2025, which FAA attributed to old copper wiring and relocating airspace control without adequately testing the commercial telecommunications lines carrying critical data. That is a separate case, not evidence of the cause of the November ground stop. Its practical lesson is that modernization can itself introduce exposure if transition assurance is weak. [3]
Nor is 11,389 schedules inherently evidence of poor management: a national program may legitimately require thousands of local plans. The material finding is that they were not integrated. A component can be delivered on time while the service remains dependent on an unfinished interface, an unavailable operator or an untested fallback. The unit of success should therefore be an end-to-end service accepted into operation, not an isolated installation.
Put the cost of deferral into the investment case
A credible business case should compare at least three paths over the same period: sustain the existing estate with explicit mitigations; modernize in stages; or replace more extensively. Each needs a complete operating model, transition plan and residual-risk assessment. The cheapest purchase price may not produce the lowest total exposure.
For each path, quantify four distinct ledgers:
1. The cost of keeping the service running. Include maintenance, scarce skills, manual reconciliation, support contracts and compensating controls. Use actual costs and identify where allocations are estimates.
2. The cost of disruption and recovery. Model plausible scenarios using interrupted throughput, lost contribution margin, recovery work, contractual obligations and downstream effects. Separate work that is lost from work that is merely delayed. Avoid counting the same loss at multiple stages of a supply chain.
3. The cost of transition. Include parallel operations, testing, training, integration, data migration and the possibility of a failed cutover. A transformation proposal that excludes these costs is incomplete.
4. The value of future capacity. Assess opportunities that the current architecture constrains, such as faster product changes or greater throughput. Label these as forecasts, test the demand assumptions, and do not treat theoretical time savings as cash unless there is a mechanism to realize them.
ARKHOS’s proposed decision discipline is to show an expected-cost range alongside a severe-but-plausible stress case. Where incident probabilities are weak, present sensitivity analysis instead of an apparently precise expected loss. Safety, legal and continuity requirements remain constraints; they cannot be waived simply because a spreadsheet shows an attractive return.
Make the decision incremental, not rhetorical
The relevant comparison is not “modernization costs money; downtime is expensive.” It is the present-value difference between credible options over the same horizon. A staged program is financially preferable when its incremental investment and transition costs are outweighed by avoided sustainment costs, reduced expected disruption, and realizable capacity benefits, after allowing for residual risk. This is a decision rule, not a calculation of FAA’s losses.
For a proposed one-year deferral, management should show what is genuinely saved or postponed, which operating costs continue, how exposure changes during that year, and whether waiting improves information or avoids a risky cutover. Postponing capital spending is not the same as permanently saving it. Conversely, buying early is not automatically value-creating if the chosen solution is immature or the organization cannot implement it safely.
Three checks make the model usable. First, calculate the break-even reduction in disruption needed to justify each option, rather than assume an attractive return. Second, test common-cause failures: two suppliers or two data centers do not provide independent resilience if they share a critical network, control plane or operational dependency. Third, separate technical restoration from business recovery. A system restart does not instantly clear a backlog, restore displaced capacity or complete customer remediation. Those are distinct recovery workloads, and their costs must not be counted twice.
Where safety or mandatory continuity thresholds cannot be met, the issue is not which positive ROI wins. Management must first identify options that satisfy the constraint, then compare their cost and deliverability. An explicit deferral decision should name a risk owner, temporary mitigations, a review date and observable triggers that end the deferral.
The Gulf’s growth agenda makes sequencing more important
For Saudi and Gulf organizations expanding aviation, logistics, tourism or industrial capacity, the relevance is not that their infrastructure has the same weaknesses. That would require local evidence. The transferable issue is the growing number of dependencies across an end-to-end service.
A new terminal, warehouse or customer platform can expand demand on older identity, communications, settlement or operational systems. Evaluating the new asset in isolation can miss the capacity and recovery requirements it creates elsewhere. Investment approval should include an explicit dependency budget and evidence that the weakest critical link can support the intended scale.
This is also why AI investment and foundational modernization belong in the same portfolio discussion. Additional automation can increase the speed and volume of decisions. It does not by itself improve the availability of the systems that execute them. The investment committee should ask which foundations must be strengthened first, and which capabilities can safely be introduced in parallel.
Five measures worth taking to the board
ARKHOS proposes a concise operating scorecard:
- Critical-service coverage: the share of critical services with an identified owner, dependency map and approved recovery target.
- Demonstrated recovery: the latest tested technical restoration time and time to recover business service, compared with separate service requirements; identify test scope, failures and exclusions.
- Unsupported exposure: critical dependencies without reliable support, replacement capacity or a funded mitigation plan.
- Migration readiness: compatibility, load, fallback and operator-readiness tests passed before each release.
- Risk retired: critical legacy dependencies actually removed or demonstrably mitigated after acceptance, with residual failure scenarios stated. Do not equate the share of components installed with the share of risk eliminated.
Targets should be set by operational consequence, not copied from another industry. Review exceptions with the same discipline as capital overruns, and require a dated plan when management requests deferral.
The lesson is not to approve every modernization proposal after a dramatic outage. It is to stop treating delay as the zero-cost alternative. If the board postpones the investment, what exposure has it accepted, who owns it, and when will that decision be tested again?
Sources and evidence boundaries
[1] ABC News, live coverage updated 13 November 2026, 21:57 EST, reporting FAA and US government statements. Event facts are provisional and may develop.
https://abcnews.com/US/live-updates/faa-ground-stop-live-updates-hundreds-flights-canceled/?id=138125140
[2] US GAO, Air Traffic Control: FAA Actions Are Urgently Needed to Modernize Aging Systems, GAO-24-107001, 23 September 2024. Historical 2023 assessment; not current system status.
https://www.gao.gov/products/gao-24-107001
[3] US GAO, Air Traffic Control Systems: Ambitious New Modernization Effort Needs to Improve Cost and Schedule Planning, GAO-26-107992, 14 September 2026. Implementation and planning data have their own stated observation dates.
https://files.gao.gov/reports/GAO-26-107992/index.html
The evaluation framework and Gulf implications are ARKHOS analysis, not findings about the undisclosed decisions of any particular board. No verified monetary loss for the November incident is asserted.
© 2026 ARKHOS. All rights reserved. Rights in cited sources remain with their respective owners.