The U.S. Space Force is going back to the drawing board on the software that flies GPS. SpaceNews reported on Sept. 5 that the service is starting work on a new ground-control architecture for its future GPS constellation, and that the effort explicitly applies lessons from the Next Generation Operational Control System, or OCX, the $6.27 billion program the Pentagon canceled earlier this year after 15 years of development.

That single sentence carries a lot of weight. OCX was supposed to be the modern brain of the world's most-used navigation system. Instead its cancellation in April left the constellation dependent on the existing control system, which is now being extended to fly the next-generation satellites rolling off the line. The Sept. 5 report signals that a replacement is now underway. Most of the SpaceNews story sits behind a paywall, so what follows leans on the April cancellation record and on the Government Accountability Office's earlier warnings to lay out what the new program will have to do, and why "lessons learned" is doing so much work in that sentence.

How OCX Died

The formal end came on April 17, 2026, announced three days later. According to Breaking Defense, the Space Force had spent $6.27 billion on OCX as of January 2026, against an earlier projection of $3.7 billion. The system was originally supposed to be complete in 2016. The fiscal 2027 budget request still carried roughly $332 million for it.

The official rationale was blunt for a government program termination. OCX, the Space Force said, "was unable to deliver needed capabilities on an operationally relevant timeline at an acceptable level of risk." Col. Stephen Hobbs, who commands Mission Delta 31, said extensive system issues had arisen during integrated testing of OCX with the broader GPS enterprise, with problems across a broad range of capability areas. He put the conclusion in plain terms: "The challenges of onboarding the system in an operationally relevant timeline proved insurmountable."

Tom Ainsworth, the Space Force's acting service acquisition executive, framed the decision as a process lesson rather than just a program failure. "It's important we refine and update acquisition processes to prioritize rapid, incremental capability delivery," he said. That phrase, incremental delivery, is the clearest hint available of what the post-OCX architecture is meant to look like: something fielded in pieces that can be tested and used along the way, rather than a single monolithic system that has to work end to end before operators can touch it.

RTX, the contractor, has a different framing. The company said it delivered the system in 2025 and would "work closely with the government on the next steps." Delivery and operational acceptance are not the same thing, and the gap between them is where OCX spent its final years.

What the Watchdog Saw Coming

None of this arrived without warning. In a June 2023 report, the GAO noted that the Space Force had postponed OCX delivery to at least December 2023, citing development challenges, and that the service itself acknowledged "remaining risks could lead to additional delays." That was a program already seven years past its original completion date, still pushing its own milestones to the right.

The GAO report also laid out why the ground segment mattered so much beyond simple schedule embarrassment. The military M-code signal, a more jam-resistant, military-specific signal that first launched on a satellite in 2005, depends on all three legs of the system: ground, space, and user equipment. The satellites were in place, but the ground segment was delayed and the receiver side was unsettled. Satellites broadcasting M-code do little good if the control segment cannot fully manage the signal and troops do not have receivers to use it. On that last point, the GAO found the Space Force "lacks a major committed customer" for a new handheld M-code receiver.

The same report flagged a structural question about the constellation itself. The Space Force met its approved requirement for 24 M-code-capable satellites on orbit, but determined it needs at least three more, 27 in total, to meet certain user requirements for accuracy, and the GAO found it "not likely" that 27 would be consistently available over the next decade. The Defense Department concurred with both of the GAO's recommendations. Those constellation-level questions are separate from OCX, but they shape the environment any new ground system will inherit: a fleet meeting its baseline requirement while short of what some users need, with the ground segment lagging the satellites.

The Stopgap Now Flying the Fleet

With OCX gone, GPS is flown on the Architecture Evolution Plan, or AEP, the current control system managed by Lockheed Martin. The Space Force said in April it would continue enhancing that system, and Lockheed was awarded $105 million to do so, building on more than a decade of prior AEP upgrades.

The contract scope, as reported by The Defense Post, reads like a list of everything OCX was supposed to handle. It covers launch, early-orbit operations, and disposal support for GPS IIIF satellites, along with continued improvements to the resilience and performance of the wider GPS enterprise and strengthened M-code capability for contested environments where signals may be disrupted or jammed. Lockheed is building the GPS IIIF spacecraft in Denver, having finished the GPS III production line, and the IIIF design offers higher anti-jam performance and improved accuracy than its predecessors.

That is the awkward position the Space Force is in. It is launching a more capable generation of satellites while relying on a control system that predates them, patched to keep pace. AEP can be upgraded, and has been repeatedly, but $105 million buys sustainment and incremental improvement, not a clean-sheet replacement. The stopgap works precisely because it is a known quantity, which is also why it cannot be the long-term answer.

What the Next System Has to Do

The SpaceNews lede supports two claims about the new effort: work is starting, and lessons from OCX are being applied. Everything beyond that is inference from the record, but the record is fairly explicit about the requirements.

First, the new architecture has to close the M-code ground-segment gap the GAO has been documenting for years, on a system that can actually be certified and operated rather than delivered and shelved. Second, it has to fly GPS IIIF, which AEP is now being extended to do in the meantime. Third, given the emphasis on contested environments in the reporting on the AEP contract, it has to be built for a threat environment that did not exist when OCX was conceived. And fourth, per Ainsworth's own words, it has to arrive in increments, so that the next program does not spend 15 years and run roughly 70 percent over its original projection before anyone can say whether it works.

Why It Matters

GPS is not just a military asset; it is civil infrastructure that banking, transportation, and utilities all quietly depend on. The ground segment is the part of the system nobody sees, and it is the part that just failed at a cost of $6.27 billion. The Space Force is now attempting to do in the shadow of that failure what it could not do in 15 years of trying: field a modern control system that keeps up with the satellites it is launching.

The stakes of getting the second attempt right are higher than the first, because the fallback is already in use. AEP is a decade-plus-old architecture being asked to support a new generation of spacecraft and a signal whose first satellite launched in 2005. If the post-OCX program repeats the pattern of slipping milestones and unresolved test issues, there is no further stopgap behind the stopgap. The phrase "lessons learned" will be measured against a simple test: whether the next ground system reaches operators before the satellites it is meant to fly are themselves due for replacement.

Sources