TR-3 delays, aircraft delivered without combat capability, skyrocketing costs: an investigation into Block 4 of the F-35 and what customers are losing.
In summary
The Block 4 standard was supposed to transform the F-35 into an upgradeable platform, capable of integrating new sensors, advanced electronic warfare modes, and an expanded weapons arsenal. But this leap depends first on TR-3, an avionics overhaul that adds a more powerful central processor, more memory, and modernized displays. However, TR-3 has accumulated a series of setbacks: the US administration has had to accept, from the summer of 2024 onwards, aircraft equipped with the new hardware but delivered with software that is still immature and therefore unusable in combat. At the same time, the Block 4 roadmap has been reduced and postponed, with rising costs and development now projected to be completed in the early 2030s. This modernization, intended to secure technological advantage, is becoming an industrial and operational risk for all customers of the program. In 2024, every delivery was delayed, and 174 non-combat-capable TR-3s were received.
The modernization standard that was supposed to take it to the next level
On paper, F-35 Block 4 is the logical continuation of a fighter jet already largely defined by its software. The F-35 is not a “fixed model.” It is a platform that is being developed incrementally, in line with fixes, new threats, and the priorities of its users. The ambition of Block 4 was therefore simple to state: gain digital margins, integrate new sensors more quickly, push data fusion, and expand the range of available weapons.
The problem is that this promise was presented as continuity, when in fact it is an architectural shift. Block 4 is not a “comfort” update. It requires the requalification of part of the IT core, sensors, test benches, and, ultimately, certification procedures. When you “re-platform” a weapons system, you accept a hard rule: every new feature makes the slightest delay more costly, because the industrial chain does not stop on its own.
The promise of a leap in capability, and the debt it creates
The functions targeted by Block 4 can be understood by looking at the trajectory of threats. Modern air defense requires faster processing, more sophisticated jamming, longer-range firing, and shorter decision times. This means more on-board computing, more data to merge, and more demanding algorithms.
This standard aggregates developments in electronic warfare, communications, navigation, and weapons integration. Some projects are visible (weapons, links), while others are less so: the ability to absorb more frequent updates, improve signal analysis, automatically prioritize tracks, or manage denser sensor modes. In this type of program, it is often the invisible developments that are the most expensive, because they affect the foundations: software architecture, testing tools, simulation laboratories, and compatibility with a global fleet that is already flying.
This is where the debt arises. The more functions you add, the more you increase the volume of testing, the more you multiply the interfaces, the more fragile you make the integration. And when a program already has a history of delays, each new ambition acts as a risk multiplier.
The hardware base that turned a construction site into a bottleneck
The key to success is called Technology Refresh 3 (TR-3). The idea is classic: when the software wants to do more, you upgrade the mission computer. Except that in this case, the scale was underestimated. TR-3 is a hardware and software package designed to restore computing, memory, and throughput margins, while preparing for the arrival of Block 4 capabilities.
In concrete terms, TR-3 relies on a new central computer and modernized avionics.
The most reliable descriptions agree: a more powerful computing core, more memory, and better support for displays and sensors to support the growth of radar, infrared, electronic warfare, and communications processing. This is not a simple “processor upgrade.” It is a redesign of the way the aircraft absorbs and redistributes information.
The harsh reality of an aircraft at the limits of its margins
In a modern fighter jet, margins are not just a matter of lines of code. They are physical. More computing power means more power consumption, more heat to extract, more tests to run, and more interfaces to secure. The F-35 comes at a time when each new module reduces the system’s tolerance. When a program announces dozens of improvements, it must prove that the platform can support them without compromising other areas (stability, availability, maintenance time, vulnerabilities).
This is precisely what makes TR-3 so critical. Ambitious Block 4 capabilities cannot be deployed on a computing core that has no margin left. Conversely, a new computing core cannot be delivered with confidence until it is stabilized. This contradiction has become the program’s center of gravity.
The technological refresh that put production on hold
TR-3 slipped, then contaminated everything else. The problems were both hardware-related (component maturity, quality, and volumes delivered by the supply chain) and software-related (integration, testing, stability). The effect on production was immediate: the line continued to roll out aircraft, but the buyer was reluctant to accept them if the software was not qualified.
In 2024, the result was spectacular: 110 aircraft delivered, all late, with an average delay of 238 days. This statistic tells a simple story: the program experienced a year in which on-time delivery became the exception, then ceased to exist.
The processor that derailed the clock
Among the most sensitive elements, the program pointed to the integrated core processor. The component suffered from late maturation and production difficulties at the supplier, with late deliveries and quality issues. U.S. authorities indicated that by the summer of 2025, the processor was finally considered “stabilized”; but in the meantime, the momentum of delays had already snowballed.
This point is central to understanding the mechanics of the “nightmare.” When the computing core arrives late, testing arrives late, fixes arrive late, and qualification arrives late. Meanwhile, production continues. Each month of delay is then converted into stock, rework, or degraded deliveries. And each aircraft that “leaves the factory” but is not accepted becomes a political issue: who pays, where to park it, who is responsible for it, and what is the risk in case of damage?
Delivery of aircraft that are accepted but not combat-ready
Faced with the risk of seeing more than 100 aircraft accumulate at the manufacturer’s facility, the U.S. administration has chosen a middle ground starting in July 2024: to accept aircraft equipped with TR-3 hardware but delivered with non-combat-capable software. These aircraft are used for training, ramping up units, and training mechanics. But they are not supposed to be used in operations.
The figures are unusual on this scale. In mid-2025, the US buyer reported having accepted 174 aircraft in this configuration, precisely because an accumulation of aircraft in long-term parking posed an industrial and contractual risk.
The financial lever that betrays uncertainty
This choice is accompanied by a payment retention mechanism: approximately $5 million retained per aircraft accepted in this incomplete version, with gradual release when technical milestones are reached (weapons validation, component stabilization, etc.). This resembles a warranty clause applied at the fleet level.
This is not a normal procedure for an aircraft that is supposed to be “operational” upon delivery. It is a crisis management measure. It also reveals a political tension: maintaining the flow to the forces, without fully accepting that part of the value delivered is conditional.
The hidden cost lies elsewhere. Accepting early means accepting retrofitting. You have to plan for time in the hangar to inject the right software, revalidate configurations, and absorb test feedback. All of this consumes maintenance time and training slots. In an air force, these slots are rare. They compete with training, operational preparation, and alert readiness.
Above all, configuration management becomes an ongoing problem. A “TR-3 material” aircraft is not necessarily “TR-3 operational.” Crews, mechanics, and planners need to know exactly what software is loaded, what weapons are authorized, what sensor modes are reliable, and what limitations apply. When a fleet becomes fragmented, availability declines, even if each aircraft, taken individually, is state-of-the-art.
The reduction of promises that transforms a standard into a moving target
The second shock, less visible but more lasting, concerns the roadmap itself. In 2018, the program announced a set of 66 capabilities to be delivered by 2026. Since then, the trajectory has been pushed back, redrawn, and the content revised. The US authorities have confirmed that a sub-program dedicated to Block 4 will consist of a subset of the capabilities initially planned, and that others will be postponed or withdrawn.
The most frequently cited end date for this subset is now 2031 at the earliest. An internal review even estimated that many capabilities would not join the fleet until the mid-2030s, partly for reasons of technical feasibility. We have therefore moved from a “planned” standard to a “renegotiated” modernization.
Thermal constraints and propulsion: the wall that cannot be circumvented
Why reduce the content? Because certain capabilities, particularly those that draw heavily on energy and cooling, depend on parallel modernization of the engine and thermal management. This point is often underestimated. More powerful avionics, more efficient sensors, and more demanding electronic warfare do not come free: they increase the thermal load and electrical demand. Without additional margins, the architecture becomes blocked.
The program therefore finds itself making a pragmatic choice: first deliver what is compatible with the existing architecture, and postpone what requires a leap in power and cooling. This is rational. But it is also an admission: the “complete” Block 4, as it was imagined, no longer corresponds to what the platform can absorb at the promised pace. And when a standard becomes dependent on another project (engine, thermal), it ceases to be a standard and becomes a permanent negotiation between sub-programs.
The knock-on effects on customers and credibility
The delays and fluctuating content have three concrete consequences.
The first is industrial. In 2025, the manufacturer posted a record number of deliveries (191 aircraft), largely fueled by the catch-up of stock accumulated during the lockdown period. A record year can mask a less comfortable truth: a “normal” pace has been replaced by fits and starts, which complicate workforce planning and partner visibility.
The second is operational. A fleet fragmented by software configurations requires strict management discipline: which aircraft can pull what, with which version, and with what limitations? This consumes maintenance hours, training cycles, and hangar slots. The cost does not always appear in a single budget line, but it is paid for in availability.
The third is political. The F-35 is the largest Western fighter aircraft program, with a U.S. target of 2,470 aircraft over an announced service life of 77 years. Its credibility rests as much on the promise of evolution as on current performance. When modernization turns into an uncertain trajectory, the conversation changes: customers no longer ask “when is the next capability coming?”, they ask “which one is going to be abandoned, and when?”
Governance that must choose between transparency and narrative
The program is attempting to regain control by isolating Block 4 and TR-3 in a sub-program with dedicated metrics. The intention is sound: to separate production (which is running smoothly) from development (which is slipping) and make trade-offs more visible. But we must be clear-headed: reclassifying a project does not, on its own, repair the accumulated debt.
The most troubling point is this: we built an airplane like software, then discovered that software cannot run faster than physics, nor faster than test benches. Block 4 is not an “accident.” It is the result of continuous ambition, added year after year, on a platform whose actual margins were narrower than the narrative.
A way out of the crisis is possible, but not painless
There is a way out. It involves more testing, earlier, with shorter, more iterative cycles. It also involves a discipline of prioritization: stop promising lists, and deliver measurable, validated, and sustainable capabilities. Finally, it requires consistency between avionics, sensors, weapons, and propulsion, instead of making modernization dependent on aligning schedules.
The thorny question remains: how much will it cost, and how long will it actually take? Past estimates already put the cost of Block 4 at $16.5 billion, up more than 50% from the initial baseline. Until the new costing and final list of capabilities are finalized, Block 4 will remain a paradoxical standard: indispensable from a military standpoint, but too uncertain to be sold as a simple upgrade.
Sources
U.S. Government Accountability Office (GAO), GAO-25-107632, F-35 Joint Strike Fighter: Actions Needed to Address Late Deliveries and Improve Future Development, September 2025.
Congressional Research Service (CRS), R48304, F-35 Lightning II: Background and Issues for Congress, December 11, 2024.
Defense News, Key tests for latest F-35s will begin in 2026, February 5, 2025.
Lockheed Martin, F-35 Breaks Delivery Record, Continues Combat Success in 2025, January 7, 2026.
Breaking Defense, F-35 Block 4 upgrade delayed until at least 2031: GAO, September 3, 2025.
Live a unique fighter jet experience