The Ford Class Reckoning: Can CVN-NEXT Find the Right Answer

CVN-NEXT and the Lessons
NAVSEA Keeps Not Learning

Proceedings U.S. Naval Institute  |  Analysis & Commentary

Surface Warfare / Acquisition

The Navy is reviewing the design of its next two carriers before contracts are awarded — an admission that the Ford class as built is not what the Navy needs going forward. History suggests the review will produce incremental fixes where architectural reform is required.

Stephen L. Pendergast, LT USNR, IEEE Sr. Member (Ret.)  |  Senior Engineer Scientist May 2026 Sources: USNI, CRS, GAO, DOT&E, NAVSEA

BLUF — Bottom Line Up Front

The Navy's April 2026 announcement of a design review for CVN-82 and CVN-83 — the fifth and sixth Ford-class carriers, not yet under contract — is a tacit acknowledgment that a program delivered $2.5 billion over budget with 23 unproven technologies, persistent EMALS and AAG unreliability, habitability failures, and a carrier fleet falling below eleven hulls for the first time in years, requires more than incremental correction. The danger now is that the review produces a CVN-65-style one-off or small sub-class — expensive to sustain precisely because of its uniqueness — rather than the clean-sheet integrated power architecture the DDG-1000 Zumwalt class demonstrated was achievable when properly tested ashore before ship installation. The core failure of the Ford program was not the ambition of its technologies but the abandonment of disciplined land-based testing before fleet introduction. That lesson, embedded in naval engineering practice since the steam-to-diesel transition, was available. NAVSEA chose not to apply it. The question now is whether the CVN-82 review will correct that error or repeat it in a smaller hull form.

The aircraft carrier USS Gerald R. Ford has been in the news for all the right operational reasons and all the wrong programmatic ones simultaneously. In early 2026, executing Operations Southern Spear, Absolute Resolve, and Epic Fury — including sustained operations against Iran in the Persian Gulf during Operation Epic Fury — she completed more than 10,500 fixed-wing launch and recovery evolutions, and Navy leadership described her sortie generation rate as "eye-watering."¹ The ship works. At the same time, the Director of Operational Test and Evaluation's most recent annual report confirms that EMALS and AAG reliability "continues to adversely affect sortie generation and flight operations" and "remains the greatest risk to demonstrating operational effectiveness and suitability in IOT&E."² The ship works — until it doesn't, and when it doesn't, off-ship technical support must be flown in to fix systems that operational crews cannot maintain independently. That is not an acceptable posture for the Navy's primary strike platform.

Into this context arrives the announcement from Secretary of the Navy John Phelan at Sea-Air-Space 2026: the Navy is conducting a comprehensive review of the Ford-class design for CVN-82 (USS William J. Clinton) and CVN-83 (USS George W. Bush), to be completed by the end of May 2026. "We are looking at 82 and 83 to review the costs, the designs, the systems, to make sure that they make sense, and they have all the systems and requirements that we want going forward," Phelan told reporters. "It's a prudent and practical thing for us to do, given the costs of them as a percentage of the budget."³ Neither ship is yet under contract, which creates genuine design flexibility. What the review will produce, and whether it will reflect the actual lessons of the Ford program rather than political and budgetary convenience, is the central question for the next generation of American sea power.

The Programmatic Record: A Precise Accounting

Before addressing what should be done, the record of what was done deserves precise statement, because acquisition communities have a consistent tendency to let the passage of time soften their institutional memory of program failures.

The USS Gerald R. Ford (CVN-78) was commissioned July 22, 2017, after delivery $2.5 billion over budget and years behind schedule.⁴ She did not achieve Initial Operational Capability until December 2021 — more than four years after commissioning — and her first deployment did not begin until October 2022, five years after commissioning. At delivery, 9 of 11 Advanced Weapons Elevators were non-operational.⁵ EMALS experienced a failure rate of approximately 1 in 75 catapult strokes during early operations, compared to a requirement that was substantially more reliable.⁶ The Dual-Band Radar required replacement on subsequent ships. Habitability was found lacking for both ship's company and the embarked carrier air wing, a finding the DOT&E has maintained across multiple reporting periods; during the 2025-2026 deployment, toilet systems clogged and a laundry room fire temporarily denied sailors the ability to clean their uniforms.⁷

The root cause, identified consistently by the Government Accountability Office across multiple annual assessments, is unambiguous: the program simultaneously introduced 23 unproven technologies into a lead ship without completing required land-based testing for critical systems before they were installed in the hull. As GAO documented, "the decision to move equipment into ship construction before land-based testing was completed forced retrofits across the ship" — retrofits that drove both cost growth and schedule delay on every subsequent hull.⁸ CVN-79 (USS John F. Kennedy), supposed to deliver July 2025, will not deliver until March 2027 at the earliest, still waiting on AAG certification and Advanced Weapons Elevator completion. CVN-80's delivery has slipped to July 2030. The total estimated procurement costs — $13.2 billion for Kennedy, $14.25 billion for Enterprise, $15.2 billion for Doris Miller — represent a cost escalation trajectory that has no credible end in sight under the current design and production approach.⁹

The CVN-65 Warning

The Navy has been here before. USS Enterprise (CVN-65), commissioned in 1961, was a technological marvel — eight A2W reactors producing unprecedented power for an aircraft carrier. She was also a programmatic anomaly: a single ship so technically unique that her maintenance requirements, spare parts supply chain, and operational support structure could never achieve the economies of a true class. Her complex reactor plant required specialized expertise unavailable in the broader Navy maintenance system and drove disproportionate lifecycle costs throughout her 51-year service life. The Nimitz class that followed standardized on two A4W reactors per ship precisely to avoid repeating this maintenance burden. Any CVN-82 redesign that produces a one-off or two-ship sub-class with unique systems risks recreating the Enterprise trap in a $14 billion hull.

What DDG-1000 Got Right — and Wrong

The Zumwalt-class destroyer program is simultaneously the Navy's best recent example of correct systems engineering methodology and its most vivid demonstration of what happens when that methodology is applied to a program the institution does not fully support.

The DDG-1000 Integrated Power System — 78 megawatts generated by two Rolls-Royce MT30 Main Turbine Generator Sets and two RR4500 Auxiliary Turbine Generator Sets, distributed to propulsion, combat systems, and ship's services from a common bus — is the correct architecture for a modern warship designed to accommodate directed-energy weapons, advanced sensors, and future high-power loads.¹⁰ It is a true all-electric ship: there is no mechanical propulsion drive, no reduction gear, no shaft-speed constraint on the generators. Power routing to any ship system — including future railguns or high-energy lasers — is a software and switching problem, not a mechanical redesign. The power management system, combat management system, and platform management system share the Total Ship Computing Environment Infrastructure, a unified open-architecture computing environment that eliminates the legacy interface seams that have caused integration problems throughout the Ford program.¹¹

The critical differentiator from Ford was testing discipline. The IPS was demonstrated at the Land Based Test Site at Naval Surface Warfare Center Carderock Division in Philadelphia across a three-year integration and risk-reduction program — including normal operations and deliberate fault conditions — before a single IPS component was installed in a ship's hull.¹² This is exactly the methodology the Ford program abandoned for EMALS, AAG, and the Advanced Weapons Elevators. It is not a new or controversial principle. The Navy's own acquisition regulations require it. The Ford program received waivers and exceptions that allowed hardware to enter ship construction before testing was complete, and the program is still paying for those waivers a decade later.

Where DDG-1000 failed was not in its engineering but in its institutional and operational concept. Originally planned for 32 hulls, the class was cut to seven, then three, primarily because the primary mission — land attack with the Advanced Gun System and Long Range Land Attack Projectile — lost its requirements basis when the cost of LRLAP munitions rose to approximately $800,000 per round, making naval gunfire an implausibly expensive alternative to air-launched precision weapons.¹³ A class of three ships cannot achieve the logistics, training, and maintenance economies that justify the upfront investment in new technology development. The three Zumwalts are now being converted to carry Conventional Prompt Strike hypersonic weapons — a valid and important mission, but one that was not the design driver. The ship is excellent. The program is a cautionary tale about requirements stability and class size.

The Ford program abandoned testing discipline that has governed naval engineering since the transition from steam to diesel propulsion. DDG-1000 demonstrated that discipline is achievable. The question for CVN-82 is whether the institution has the memory to apply it. — Author's assessment based on GAO/DOT&E program documentation

The Architecture Question: What CVN-82 Should Be

The Phelan review stated objective — "to further increase lethality, enhance survivability, and improve producibility, while also simplifying the design and potentially leading to decreased cost" — is the correct set of goals stated at precisely the right moment, since no contracts have been awarded for CVN-82 or CVN-83.¹⁴ The danger is that the review will produce incremental modifications to a fundamentally flawed architecture rather than the clean architectural decision the program requires.

The most important architectural question is whether CVN-82 should incorporate a genuine Integrated Power System on the DDG-1000 model — nuclear-powered, but with all-electric propulsion and a common high-voltage bus serving all ship loads. The Ford-class uses nuclear reactors to drive steam turbines mechanically coupled through reduction gears to propulsion shafts, with separate Ship's Service Turbine Generators providing electrical power. EMALS was bolted onto this mechanical propulsion architecture through rotating flywheel energy storage — a technically clever but inherently reliability-challenged solution to the problem of delivering pulsed power for catapult strokes from generators designed for continuous electrical loads.

A nuclear IPS architecture would eliminate this seam. The reactor plant drives generators directly. EMALS energy storage charges from the common high-voltage bus between launches and discharges through controlled power electronics during the launch stroke. The power management system allocates generation capacity across propulsion demand, EMALS charging, combat system loads, and ship's services in real time through software, not mechanical configuration. When future directed-energy weapons require hundreds of megawatts of pulsed power, the architecture accommodates them without hull redesign.

Design Dimension	Ford Class (CVN-78 through CVN-81)	DDG-1000 IPS Model Applied to CVN-82	Assessment
Propulsion architecture	Nuclear steam → mechanical drive → reduction gear → shaft. Separate SSTGs for ship's service electrical.	Nuclear → generator → common HV AC bus → electric drive motors. Single integrated plant.	IPS superior: eliminates mechanical interface; power routing by software
EMALS power delivery	Pulsed power from rotating flywheel energy storage rings; complex control interface to ship's service plant	Direct charge/discharge from common HV bus; integrated with ship power management system	IPS superior: eliminates M-G set reliability issues; common control architecture
Future high-power weapons	Requires significant modification to accommodate railguns or directed-energy weapons above installed EMALS load	Software power allocation to new loads; no hull modification required within generating capacity	IPS superior: designed for load growth from inception
Total ship computing	Multiple legacy interfaces between platform management, combat systems, and power management	TSCEI open architecture; unified computing environment; defined interfaces	TSCEI superior: eliminates integration seams that caused Ford AWE/EMALS integration failures
Manning	Designed for reduced crew; automation goals partially achieved; habitability shortfalls	DDG-1000 reduced crew via automation; habitability designed to standard from outset	Mixed: Ford automation goals achievable; habitability requires deliberate design allocation
Industrial base risk	Newport News Shipbuilding sole source; established Ford production line	New design disrupts established production rhythm; development cost significant	IPS transition costly: estimated 5-7 year delay to first IPS CVN vs. continuing Ford design
Testing discipline	23 unproven technologies, inadequate LBTS before hull installation; retroactive fixes on every hull	Requires committed LBTS program; DDG-1000 demonstrated 3-year LBTS is achievable and necessary	Correctable: but requires institutional commitment and schedule realism the Ford program lacked

The CVN-65 Risk: Avoiding the One-Off Trap

The most dangerous outcome of the Phelan review would be a decision that CVN-82 and CVN-83 receive significant design modifications that make them meaningfully different from CVN-78 through CVN-81 while remaining nominally within the Ford class — a two-ship sub-class with unique maintenance requirements, unique supply chains, and unique training demands that can never achieve the support economics of a genuine class.

This is precisely what CVN-65 (USS Enterprise) became. Commissioned in 1961 with eight A2W nuclear reactors — versus two per ship for the Nimitz class that followed — she was operationally magnificent and logistically burdensome throughout her service life. The reactor plant required specialized maintenance expertise that never existed at sufficient scale within the Navy's nuclear maintenance establishment, because it was never needed for any other hull. Spare parts were procured in quantities too small for competitive pricing. When she decommissioned in 2012 after 51 years, the cost of her reactor deactivation was elevated by the complexity of her unique plant. The lesson was supposed to be: don't build one-off ships with unique systems. The Nimitz class demonstrated that lesson was learned. The Ford program suggests the institutional memory lasted one design cycle.

A CVN-82 that restores steam catapults — President Trump's stated preference, and reportedly under active consideration in the review — while retaining AAG, Advanced Weapons Elevators, and the new island configuration would produce exactly a CVN-65-style one-off: a mix of restored older technology and newer technology that fits neither the Ford production baseline nor a clean new design. Newport News Shipbuilding would have to reconstitute steam catapult design and manufacturing capabilities that have atrophied significantly since the last Nimitz-class units. The C-13 steam catapult production infrastructure, workforce expertise, and design documentation exist primarily in archived form. The cost and schedule to reconstitute them for a two-ship buy would likely exceed the cost of fixing EMALS to the required reliability standard — which remains the correct engineering path.

The EMALS Reliability Path

The fundamental EMALS engineering problem is not the Linear Induction Motor technology — it works — but the power electronics and control system that manage energy extraction from the rotating storage rings and delivery to the LIM during the launch stroke. These are software-intensive systems that improve with operational data and iterative software releases, exactly as combat system software improves over a ship's service life. The Navy has 8,725 EMALS shots of operational data from Ford's FY2024 deployment alone. Continuing to improve EMALS control software while Ford, Kennedy, Enterprise, and Doris Miller accumulate operational experience is the correct engineering path. Reverting to steam catapults for CVN-82 would sacrifice the genuine long-term advantages of EMALS — lower structural stress on airframes, programmable end-speed control for different aircraft types and weights, accommodation of future unmanned aircraft — for a politically convenient solution to a reliability problem that is being progressively resolved.

The Unmanned Aviation Imperative

Any CVN-82 design review conducted without explicit requirements derived from unmanned carrier aviation is architecturally incomplete. The MQ-25 Stingray tanker is already part of the carrier air wing, and its catapult launch and arrested landing requirements are different from manned strike fighters. Future unmanned combat air vehicles — heavier, potentially operating at higher end-speeds, and potentially requiring different launch and recovery profiles than current aircraft — will increasingly define what the carrier's launch and recovery systems must be capable of.

EMALS's programmable end-speed control is not a minor advantage in this context. A steam catapult delivers a fixed energy profile optimized for a narrow range of aircraft weights and end speeds. EMALS can be programmed for any end-speed within its power envelope, allowing the same catapult to safely launch an MQ-25 at its optimal profile and an F/A-18 at its optimal profile without mechanical adjustment. As unmanned aircraft diversify in size, weight, and mission profile, this programmability becomes increasingly operationally critical. A reversion to steam catapults for CVN-82 would require designing future unmanned aircraft around the catapult's fixed capability rather than designing the catapult capability around the aircraft's requirements. That is the wrong direction for a ship designed to serve the fleet through the 2080s.

What a Proper CVN-82 Design Review Would Examine

A design review conducted for engineering reasons rather than political ones would address the following questions, in this order of priority:

First: Land-Based Testing commitments. Before any technology decision for CVN-82, the review should establish whether and how land-based test and evaluation will be conducted for each modified or new system. The Naval Surface Warfare Center's Land Based Test Sites exist precisely for this purpose. DDG-1000's IPS LBTS program at Philadelphia demonstrates the methodology. No technology that has not completed adequate LBTS should be authorized for installation in CVN-82's hull. This is not a new requirement. It is existing acquisition policy that the Ford program was allowed to waive. The CVN-82 review should make LBTS completion a hard contractual prerequisite.

Second: EMALS reliability disposition. The review should commission an engineering assessment of the specific failure modes driving EMALS unreliability — the power electronics, control software, and energy storage system components that have degraded mean-cycles-between-failure — and determine whether they are correctable through redesign, software update, or component replacement within the current architecture. This assessment should be conducted by an independent technical authority, not by General Atomics as the developer. DOT&E has the independence and the technical baseline data to conduct or commission this assessment. Its conclusions should drive the technology decision, not the President's stated preference for steam.

Third: IPS architecture feasibility for CVN-82. The review should assess whether CVN-82 can incorporate an all-electric nuclear IPS — nuclear plant driving generators into a common high-voltage bus, with electric propulsion motors and all combat system loads including EMALS drawing from that common bus. The A1B reactor plant already produces substantially more electrical power than the A4W. The engineering question is whether the existing A1B plant can be reconfigured to drive generators directly rather than through the steam turbine/mechanical drive arrangement, or whether a new reactor plant design is required. Newport News Shipbuilding and NAVSEA have the design capability to answer this question. If the answer is achievable within a reasonable development timeline, it should be the baseline for CVN-82. If not, the review should specify the development program required to make it achievable for CVN-83 or a true CVN-NEXT class.

Fourth: Habitability and crew systems. The DOT&E has cited habitability shortfalls in every annual report since Ford's commissioning. Plumbing failures, laundry system fires, and inadequate berthing for the carrier air wing are not acceptable on a $13 billion warship. These are not technology failures — they are design failures that reflect a requirements process that prioritized combat capability metrics over the lived experience of the 5,000 people who operate the ship. CVN-82's design review should include a human systems integration review with the rigor applied to combat systems.

The Industrial Base Constraint

Any discussion of CVN-82 design alternatives must acknowledge the binding constraint of the U.S. carrier shipbuilding industrial base. Newport News Shipbuilding — a division of Huntington Ingalls Industries — is the sole facility capable of building nuclear-powered aircraft carriers. Its workforce, tooling, and production infrastructure represent decades of accumulated investment. The Ford-class production line, whatever its problems, is the only carrier production line the United States has. A design change significant enough to break the current production rhythm — new reactor plant, new propulsion architecture, new hull form — would introduce years of delay at a moment when the fleet is already falling to ten carriers due to the Kennedy delivery slip and Nimitz retirement. The practical result of the carrier fleet dipping below eleven hulls, as the CRS has documented, is gaps in global carrier presence that adversaries notice and exploit.¹⁵

This constraint argues for a phased approach: CVN-82 and CVN-83 as improved but recognizable Ford derivatives, with EMALS reliability corrections, habitability improvements, and maximum use of LBTS before hull installation for any modified systems; followed by a genuine CVN-NEXT design program, starting now, that develops the all-electric nuclear IPS architecture with the testing rigor the DDG-1000 applied to its IPS — not rushed to the first hull until the land-based testing program is complete. The lead time for nuclear carrier design is measured in decades. A CVN-NEXT design program starting in fiscal year 2027 would produce a ship ready for procurement no earlier than the late 2030s — which is exactly when the Nimitz-class retirements create the strategic opportunity for a genuine architectural transition.

NAVSEA always seems to need to learn things the hardest most expensive way. — Observation from a retired senior engineer scientist with direct experience in naval computing systems development, 2026

The Institutional Reform Imperative

The deeper problem underlying the Ford class program is not technical — it is institutional. The incentive structures governing NAVSEA program offices, major defense contractors, and the congressional oversight process are not aligned with the patient, disciplined, test-before-you-build methodology that produces reliable warships. Cost-plus contracts insulate contractors from the financial consequences of test failures that require design changes. Congressional district interests in maintaining production at specific facilities create pressure to start construction before design is mature. Program offices face career incentives that reward delivering ships on schedule over delivering ships that work as specified.

DDG-1000's IPS LBTS program succeeded partly because it was led by personnel at NSWCDD-SSES who had the technical authority and institutional standing to insist on testing completion before hull installation — and because the program office supported that insistence. The Ford program's failure to apply the same discipline to EMALS, AAG, and the AWE reflects a program office culture and a contractor relationship that prioritized schedule over test completion. No design review of CVN-82 will be successful if it does not also address the institutional conditions that produced the Ford program's failures.

The Navy's acquisition community has produced important internal documentation of lessons learned from the Ford program. The question is whether those lessons will influence CVN-82's design and contracting structure, or whether they will be filed alongside the lessons-learned documents from the A-12 Avenger, the LCS, and every other major acquisition program whose failures were thoroughly documented and then partially repeated in the next generation.

Conclusion

The Phelan review of CVN-82 and CVN-83 is the right action at the right moment. No contracts have been awarded. The design is not locked. The Ford class's operational performance during Epic Fury — 10,500 sorties, three combat theaters, sustained high-tempo operations — demonstrates that the basic concept is sound. What is not sound is the production and integration methodology that delivered the class years late, billions over budget, and with reliability problems that DOT&E identifies as the greatest risk to demonstrated operational effectiveness nearly a decade after commissioning.

The minimum acceptable outcome of the review is a commitment that no technology enters CVN-82's hull without completing an LBTS program equivalent to what DDG-1000 applied to its IPS. This single change, applied consistently, would have prevented the majority of the Ford program's cost and schedule problems.

The correct outcome is a CVN-82 that incorporates EMALS reliability corrections developed and validated ashore, addresses documented habitability failures as a first-class requirement, and commissions a serious CVN-NEXT IPS architecture study — nuclear all-electric drive on the DDG-1000 model — as the design basis for the next true class of American carrier. That class should not enter production until its IPS has been tested at full power in a land-based facility for at least three years. If that means waiting until the 2040s, the fleet structure gap should be addressed by life extensions and readiness improvements on existing hulls rather than by rushing an immature design into production.

The CVN-65 Enterprise taught the Navy not to build unique one-off ships. The Nimitz class applied that lesson for forty years. The Ford program forgot it. The DDG-1000 program relearned the testing discipline that prevents such forgetting. CVN-82 is the opportunity to remember both lessons at once — before committing $14 billion to another hull that the fleet will still be fixing in 2040.

Notes and Sources

1. Naval News. (May 12, 2026). "U.S. Navy Reveals Future Plans for Its Aircraft Carrier Fleet." [Ford 10,500 sorties; Operations Southern Spear, Absolute Resolve, Epic Fury.] navalnews.com
2. Director, Operational Test and Evaluation. (February 2025). FY 2024 Annual Report: CVN-78 Gerald R. Ford Class Nuclear Aircraft Carrier. Office of the Secretary of Defense. [EMALS/AAG "greatest risk" language; deployment data; habitability findings.] dote.osd.mil
3. USNI News. (April 21, 2026). "Navy Reviewing Ford-class Carrier Design Ahead of Future Contract Awards." Shelbourne, M. [Phelan Sea-Air-Space quotes; CVN-82 procurement acceleration to FY2029.] news.usni.org
4. National Security Journal. (November 2025). "The Navy's USS Gerald R. Ford Aircraft Carrier Almost Became an Unfixable Problem." [CVN-78 $2.5B over budget; delivery timeline; 23 unproven technologies.] nationalsecurityjournal.org
5. Ibid. [9 of 11 weapons elevators non-operational at delivery; EMALS 1-in-75 failure rate.]
6. Congressional Research Service. (December 4, 2025). Navy Ford (CVN-78) Class Aircraft Carrier Program: Background and Issues for Congress. Report RS20643. O'Rourke, R. [Comprehensive program cost, schedule, and technology status.] congress.gov
7. The National Interest. (May 2026). "The United States Navy's newest and largest aircraft carrier USS Gerald R. Ford...toilets have clogged, and a fire broke out in the laundry room." [Habitability problems during 2025-2026 deployment.] nationalinterest.org
8. Government Accountability Office. (June 2025). Defense Acquisitions Annual Assessment: Drive to Deliver Capabilities Faster Elevates Risk, GAO-25-107569. [CVN-78 program unit cost increase; CVN-79 AWE construction; land-based testing deficiencies.] gao.gov
9. CRS Report RS20643, op. cit. [CVN-79 $13.2B; CVN-80 $14.25B; CVN-81 $15.2B estimated procurement costs.]
10. Marine Log. (September 26, 2014). "DDG 1000 Marks Another Milestone." [Rolls-Royce MT30 35.4 MW; RR4500 3.8 MW; total 78 MW ship's service power; IPS generator light-off.] marinelog.com
11. Shipshub. (2025). "Zumwalt-class Destroyer." [TSCEI; open architecture; Raytheon combat system integration; 6 million lines of software.] shipshub.com
12. Defense Media Network. "DDG 1000's Integrated Power System Software, Hardware Come Together in Successful Test." [NSWCDD-SSES Philadelphia LBTS; 3-year integration and risk-reduction program; normal and fault condition testing.] defensemedianetwork.com
13. The War Zone. (February 5, 2025). "USS Gerald R. Ford Was Still Struggling With Its Dual Band Radar Prior to Deployment." [AAG replacement parts shortfalls; off-ship technical support requirement; fourth AAG engine consideration.] twz.com
14. USNI News, op. cit. [Phelan review rationale: "increase lethality, enhance survivability, improve producibility, simplify design, decrease cost."]
15. Congressional Research Service. (May 2026). Navy Force Structure and Shipbuilding Plans: Background and Issues for Congress. O'Rourke, R. [Carrier fleet gap: 10 carriers during Kennedy delivery delay; USS Nimitz retirement May 2026.] congress.gov