I sat through a briefing where an F-35 commander sighed and tapped an empty checklist. You could feel the certainty leave the room: the jet on the tarmac had hardware that promised one future and software that argued for another. If you’ve ever watched a device slow as updates stop, you know the sinking feeling.
I’ll tell you what the documents — the one Bloomberg reviewed and years of reporting from Defense One and the F-35 program office — actually show. You don’t need to be an engineer to grasp the risk: millions of lines of code and years of investment can still be undone by a software rollout that doesn’t work. I’m going to walk you through the failures, the players, and why a jet that cost about $80,000,000 (€74,000,000) to build is flying missions with software that didn’t land its promised upgrades.
The jets flew with untested software
The Pentagon’s draft review bluntly reports that TR-3 “was predominately unusable” for most of 2025. That’s an observation that reads like a mission brief: the plane had new code meant to multiply computing power and capability, but the field results were unstable and incomplete.
TR-3 was supposed to replace Technology Refresh 2 (TR-2) and deliver 20 to 25 times more computing power, greater memory, and a panoramic cockpit display. Lockheed Martin claimed deliveries of TR-3–enabled F-35s began in July 2024 and tied those jets to Block 4 capabilities: more missiles, enhanced electronic warfare, and smarter targeting. But the Pentagon’s review — as reported by Bloomberg — says no additional combat capabilities were delivered in 2025 because the version of software in the fleet hadn’t been fully tested.
Why are F-35 software upgrades delayed?
You can trace delays to stability failures during integration and a strict acceptance policy from the Pentagon: if a jet arrives with dysfunctional software, the Defense Department won’t accept it. Russ Goemaere, the Pentagon’s F-35 program spokesman, insisted TR-3 must at least match TR-2 equivalency before operational acceptance. In practice that halted deliveries and forced Lockheed Martin and the Pentagon to iterate on code that kept revealing defects under flight conditions.
Field failures made new deliveries conditional
Flight squadrons reported that installing TR-3 sometimes made jets unreliable — a real-world operational failure that altered procurement behavior overnight.
When test squadrons started trying to wring out TR-3 problems, the Pentagon stopped taking F-35s unless the software truly worked. That’s a blunt corrective gesture, but it also slowed upgrades across the fleet. You can hold Lockheed to account for development and delivery, and you can hold program managers at the Pentagon to account for acceptance criteria — but when millions of lines of code collide with hardware in the sky, the results are messy and public.
Can F-35s operate without software updates?
Yes — to an extent. The jets can still fly and carry out missions on legacy TR-2 software, but they won’t have Block 4 capabilities that expand missile carriage, electronic warfare, and target recognition. That leave commanders balancing risk: fly current jets with older software or field newer jets that may have stability gaps. Either option is a trade-off in contested environments.
Money, politics, and reputations are riding on code
The program’s price tag is decades-long and personal: taxpayers, contractors, and defense leaders all have skin in the game.
The F-35’s overhead has been debated publicly for years. In 2016, then-president Donald Trump called costs “out of control,” and later pushed for fewer jets and lower per-unit prices. He eventually helped drive a roughly 25% reduction in the per-plane cost by 2018, but total program spending remained enormous and politically fraught. When software hiccups stall capability deliveries, critics point to cost overruns; program defenders point to incremental improvements and mission utility — both make sense, but neither solves the immediate software gap.
Who is responsible for F-35 software issues?
Responsibility is shared. Lockheed Martin develops and integrates software, the Pentagon sets acceptance standards and operational requirements, and test squadrons expose failures. Outlets like Defense One and Bloomberg have chronicled the dispute between contractor deliveries and Pentagon acceptance. Public program spokespeople, such as Russ Goemaere, provide the Pentagon line; Lockheed’s public statements promote Block 4 capabilities. When public statements meet classified test reports, the result is often a narrative gap visible to anyone paying attention.
Operational pressure hasn’t paused combat use
Despite the software stall, F-35s have been positioned and used in recent tensions around Iran.
The U.S. moved F-35s into theater ahead of fights, and there are reported instances of F-35s being used to shoot down a drone in the past week. Israel has engaged with its own F-35s in air-to-air fights over Iran, too. That means jets without the promised 2025 combat upgrades are still flying missions. It’s an uncomfortable reality: you’re asking pilots to trust aircraft that may not have the latest validated software — an imperfect solution when the alternative is leaving air superiority gaps.
The technical story is simple but stark: complex software rollouts fail in ways hardware doesn’t, and the F-35 program has been tripped up by that mismatch. I’ve watched deliveries paused, acceptance criteria tightened, and PR statements traded for operational patches. The program is expensive and consequential — and sometimes it behaves like a smartphone chained to the past.
What happens next will hinge on three levers: rigorous testing that actually mirrors combat conditions, clearer public accounting of failures and fixes, and political will to demand fixes without starving readiness. The program could also become a high-priced Swiss watch missing key gears if those fixes don’t arrive fast enough.
Lockheed Martin did not respond to a request for comment on this report; the Pentagon offered public statements through program officials, and outlets from Bloomberg to Defense One have reported the underlying documents and test histories. If you want to follow the thread: the F-35 official site, congressional budget hearings, and independent reporting remain the clearest windows into progress or regress. Who is going to accept responsibility when a fleet’s software fails over a theater of war?