Skip to content

Pentagon Cancels Raytheon GPS OCX; Lockheed Wins Work

ArticleBy Factrail

By the · Method: how we source and score · Last reviewed

U.S. Air Force illustration of a GPS Block IIIA satellite

TL;DR

The Pentagon canceled Raytheon's GPS OCX after accepting the system. Lockheed gained replacement work, while taxpayers face unresolved program costs.

Key graph impact

The article explains graph entities; the causal edges belong to the facts and drivers it links to.

Article

US Air Force illustration of a GPS Block IIIA satellite Credit: US Air Force; public domain (US federal government). Wikimedia Commons file (opens in a new tab).

The institutional accountability trail links US Department of Defense (opens in a new tab), Government Accountability Office (opens in a new tab) and defence-industrial production capacity (opens in a new tab). Their roles in the policy and oversight chain should be distinguished from any unverified claim of direct financial benefit.

The Pentagon Accepted Raytheon's GPS System, Then Canceled It. Lockheed Won the Replacement Work.

Investigation | United States | 2 November 2026 |

The decision: an accepted system that never took over GPS

The Pentagon accepted a GPS control system from Raytheon in July 2025 despite documented waivers, then canceled the program on April 17, 2026 after integrated testing exposed serious problems. The Space Force put its replacement strategy on the existing system operated with Lockheed Martin, which announced a $105 million task order in April and roughly $114 million in additional work in September. The record shows a consequential acquisition decision and identifiable commercial beneficiaries, not proof of favoritism: an official $6.27 billion January program-cost figure, a separate GAO account of more than $7 billion spent, and later replacement contract ceilings measure different things. An October audit says fragmented military budgets still impede navigation modernization. The unresolved questions are how much of the canceled system's cost was finally paid, what termination liabilities remain, and whether the replacement can deliver promised capabilities on time.

The political act is not mysterious. The Department of the Air Force's April 20 announcement (opens in a new tab) identifies April 17 as the cancellation date and says the Defense Acquisition Executive acted on a recommendation from the acting service acquisition executive. Independent defense reporting identifies the Pentagon acquisition chief as Michael Duffey and the service official as Thomas Ainsworth. That is a decision within the Trump administration's Defense Department. It is not documentary proof that President Donald Trump personally selected the replacement contractor, issued an instruction to cancel the system, or benefited financially.

The sharper question is why a government could formally accept a multibillion-dollar system and then abandon the program without using it as the principal operational controller of the satellite constellation. The answer requires separating contractual acceptance, system integration and actual operational readiness. These are three different events. It also requires following the next contracts: the termination of Raytheon's OCX effort did not mean that the government stopped buying GPS control software. It shifted the modernization route toward the existing Architecture Evolution Plan, known as AEP, with Lockheed Martin receiving new work.

The Government Accountability Office's October 7 report, GAO-27-107676 (opens in a new tab), provides the crucial audit trail. Its findings are more specific than either a claim that the contractor simply delivered nothing or a claim that the government threw away a fully working system. OCX met many acceptance parameters, but the government had granted waivers and deviations, including a failure to satisfy a time-steering requirement needed for accurate satellite-clock corrections. Officials used the older ground system to perform that function during tests. Cyber survivability was also assessed as a high-risk area. Following acceptance, integrated testing with satellites, antennas and user equipment exposed deficiencies that the acquisition authorities considered too risky to resolve on the required timetable.

A documentary timeline, not a single administration's origin story

2010–2015. Raytheon began developing the Next Generation Operational Control System, or OCX, in 2010. In GAO-15-657 (opens in a new tab), auditors found that the Air Force started development before completing preliminary reviews, made poor acquisition decisions and was slow to recognize problems. The audit described an additional $1.1 billion and four years then required to deliver the system. Those problems arose long before the current Trump administration, and their historical origin must not be retroactively attributed to its 2026 cancellation decision.

2016–2019. The program exceeded important cost and schedule thresholds. A 2019 GAO schedule assessment (opens in a new tab) found that OCX was already roughly five years late and had experienced about $2.5 billion in cost growth since program start. The report questioned the remaining schedule margin and warned of more delay. Congress and successive Pentagon leaders therefore had years of warning that contractual progress was not the same as a fielded capability.

July 2025. The Space Force formally accepted OCX from Raytheon. Acceptance transferred ownership under the applicable contract process, but it was not a declaration that OCX was commanding the GPS constellation. The October 2026 GAO audit records the waivers and the unfinished testing. In March 2026 reporting by Ars Technica (opens in a new tab), RTX said it had delivered a mission-capable system and was working with the government on post-delivery issues. That is an important contractor response, not an independent certification that the system met every operational requirement.

November 2025–April 2026. Defense officials compared completing OCX with upgrading the existing Operational Control Segment–Architecture Evolution Plan, or OCS-AEP. The October GAO audit says the analysis judged OCX the higher-risk option because the remaining number and severity of defects were uncertain. On April 17 the Defense Acquisition Executive canceled OCX. The April 20 Air Force announcement (opens in a new tab) presented incremental improvement of the existing system as the alternative.

April and July 2026. The Space Force issued contracts to modify the existing control system for future GPS IIIF satellites, according to GAO. On April 16, Lockheed Martin announced a contract worth up to $105 million for launch, early-orbit and disposal functions; Breaking Defense's April 17 account (opens in a new tab) said the award had been made April 8, before the formal cancellation. This sequence matters: the replacement path was being prepared before the final termination announcement. It does not establish that the award was improper or that the acquisition decision was rigged.

September–October 2026. Lockheed Martin's September 10 disclosure (opens in a new tab) reported approximately $114 million of additional GPS ground-segment task orders: $96 million for a GPS IIIF contingency operations system and $17.5 million for the first phase of an AEP hardware/software refresh. The two specified values sum to $113.5 million, rounded in the announcement. On October 7, GAO issued its wider navigation-system audit, warning that the military services' separate budgets and programs remain difficult to align. Neither date establishes that the replacement had already achieved full operational readiness.

Who had the power to stop it?

The formal decision was not a procurement referendum. The Defense Acquisition Executive had authority to cancel the acquisition, acting on a recommendation through the Air Force's space acquisition chain. The Space Force operated the GPS constellation, provided analysis of the two options and would have to integrate the replacement with the satellites and user equipment. Thomas Ainsworth publicly defended a shift from large, all-or-nothing deliveries to incremental capability. Michael Duffey was identified in contemporary Via Satellite reporting (opens in a new tab) as the acquisition official who made the cancellation decision.

The cancellation did not erase congressional authority over defense appropriations or oversight. The October GAO report was sent to the congressional defense and appropriations committees, including members of both parties. Congress can demand cost, schedule and performance records, alter appropriations and legislate acquisition governance. It did not personally run the technical acceptance tests. The report says the Under Secretary of Defense for Acquisition and Sustainment became the milestone decision authority for the new AEP Modernization effort and required twice-yearly progress updates from the Air Force acquisition leadership. Those are identifiable control points against which future claims can be tested.

A second, less visible veto structure remains. Congress established a Positioning, Navigation and Timing Oversight Council to coordinate the military's navigation enterprise. But GAO-27-107676 (opens in a new tab) found that this council can recommend priorities while the Army, Navy and Air Force retain their own budgetary authority. A council may identify a critical timing dependency without being able to compel the service holding the funds to pay for it. The report discusses a House version of a fiscal 2027 defense bill that would create a single responsible official, but explicitly notes that the proposal would not necessarily give that official direct budget control. A proposal is not enacted law; it cannot be credited as an already functioning remedy.

This division of authority is politically relevant to Defense Secretary Pete Hegseth and congressional appropriators, but it should not be turned into a fictional personal order. The evidence assigns the OCX cancellation to the acquisition chain and the broader coordination challenge to department-wide governance. A credible accountability test would request the signed decision memorandum, the alternative analysis and the scheduled AEP-M reports, rather than guessing what a cabinet member privately intended.

The money ledger: three figures that cannot be casually added

The Air Force's April 20 release (opens in a new tab) gave approximately $6.27 billion as the program's cost as of January 2026, including Raytheon funding to date and government costs such as testing and program-office support. This is a reported program-cost figure with a defined date and scope. It is not a verified calculation of Raytheon's profit, a final termination settlement or a complete post-cancellation cash ledger.

A separate 2026 GAO acquisition review (opens in a new tab) described the OCX cancellation after more than thirteen years of development and more than $7 billion spent. Ars Technica's March reporting (opens in a new tab) discussed a roughly $7.6 billion official cost for the ground system plus a projected additional augmentation costing more than $400 million, producing a headline near $8 billion. The reports do not present an identical accounting denominator, date or distinction between incurred cost, projected completion cost and follow-on scope. Treating $6.27 billion, more than $7 billion and approximately $8 billion as interchangeable proof of cash paid to one company would be false.

The replacement figures are more precise but narrower. Lockheed's April up-to-$105 million award and September $96 million plus $17.5 million task orders are identifiable contractual opportunities. Taken at their announced ceilings or stated values, they amount to about $218.5 million. They are not an audited sum of disbursements. The September awards cover two different activities, and the April award covers launch-related support; the records do not show that the three contracts constitute the complete eventual cost of AEP-M. Nor do they establish that a dollar cut from Raytheon's old program was literally transferred to Lockheed.

Lockheed also manufactures GPS IIIF satellites under separate work. That creates a commercially significant relationship across the space and ground segments, but satellite manufacturing contracts must not be folded into the ground-control replacement figure. The Defense Department's fiscal 2026 weapon-systems budget material (opens in a new tab) names Raytheon, Lockheed Martin, BAE Systems and L3Harris in different GPS-related roles. An ecosystem of contractors is not a single recipient of every navigation dollar.

Who benefits, and what does benefit actually mean?

Lockheed Martin is the clearest named recipient of new replacement work in the available record. It received the April award and publicly announced the September task orders. Its existing AEP sustainment position may reduce transition complexity, since the older system already commands the GPS constellation. It is reasonable to describe Lockheed as a commercial beneficiary of continued modernization of that system, while stopping short of claiming a measured net profit or an improper political reward.

Raytheon, part of RTX, was the original OCX contractor. The government accepted its system in 2025, but the larger program was canceled in 2026. The record does not provide a final contractor-by-contractor breakdown of money paid, obligations terminated, claims resolved, reuse rights, warranties or intellectual property. RTX's public position, reported by Ars in March, was that it had delivered a mission-capable system and was collaborating with the government on issues discovered after delivery. Reporting that position fairly does not require endorsing it over the government test results.

Taxpayers and military users have different stakes. The public paid for a program that did not enter its intended full operational role, while the Space Force argues that spending more on unresolved defects would have created unacceptable risk. A sunk cost cannot be recovered merely by forcing deployment of a risky system. Conversely, cancellation is not costless if new contracts, retesting, cybersecurity work and schedule adjustments are needed. The actual comparison is the future cost and capability of the replacement against the remaining cost and risk of finishing OCX, not the emotional size of money already spent.

The available documents establish no personal financial interest of Duffey, Ainsworth, Hegseth or Trump in either supplier's GPS awards. They do not establish a quid pro quo, a bribery scheme, a prohibited conflict of interest or a procurement-law violation. Political contributions, lobbying contacts and meeting records could become relevant in a separate investigation, but none has been tied by the reviewed evidence to this specific cancellation decision. Company size and a favorable award are not substitutes for such proof.

What the replacement might not deliver on time

The technical consequence is not that civilian GPS immediately stopped working. The existing ground control system continued operating the constellation. The risk concerns how quickly newer satellites and stronger signals can be supported. The October GAO audit says the new GPS IIIF satellites were designed around capabilities from OCX and that moving to OCS-AEP requires additional software. The first ten GPS IIIF deliveries were, as of September 2026, averaging nine months later than a revised December 2024 schedule, although GAO notes those projected dates remained ahead of the program's formal acquisition baseline.

The most important trade-off is explicit in the government comparison. Space Force officials told GAO that either OCX or the modified older system might provide most positioning, navigation and timing capabilities around 2028–2029. They also said some cybersecurity upgrades and modernized civilian signal capabilities on the older-system path could arrive in 2030, later than under the OCX plan. This is a forecast, not a promise that either path will hit those dates. It exposes the opportunity cost of the safer-looking replacement: avoiding unknown OCX defects may defer some capabilities even if the basic GPS service continues.

The broader enterprise is not a single computer program. GPS modernization involves satellites transmitting the more jam-resistant M-code, ground systems that control them and receivers integrated into ships, aircraft, vehicles and weapons. GAO found that Army and Navy ground equipment had made progress while aviation receiver development still faced integration, funding and certification obstacles. The Navy was pursuing an interim aviation solution. Military services also pursue non-GPS sources of navigation and timing because satellite signals can be jammed or spoofed. The audit specifically cites disruption of GPS signals in the war in Ukraine, a concrete operational reason not to treat these delays as abstract technology accounting.

A replacement ground system cannot by itself make a receiver ready for a combat aircraft. Nor can an improved receiver substitute for satellites or ground software that have not delivered their planned functions. That is why GAO's October recommendations concern consolidating and coordinating portfolios across the services, rather than only demanding a better single contract.

The strongest counter-case for the cancellation

The strongest defense of the April decision is grounded in the same audit that documents the failure. The government had accepted OCX, then found serious integrated-system issues. A system that passes factory tests but cannot reliably work with real satellites, antennas, timing and operational networks can be dangerous to deploy. The government was not obligated to keep spending indefinitely because it had already invested billions. AEP had been incrementally upgraded for years and already performed essential constellation control.

Officials also did not simply choose a replacement without examining alternatives. GAO describes studies completed in November 2025 and a comparison of projected capability dates and technical risks. The Air Force announcement presents the decision as an effort to deliver usable capability faster and reduce reliance on a single complex delivery. Lockheed's existing role may have made it a practical supplier of upgrades. Those are legitimate procurement considerations even if the original program's outcome is unacceptable.

The counter-case is not a blanket acquittal of program management. GAO had warned in 2015 and 2019 about flawed early acquisition choices, contractor problems and optimistic schedules. The July 2025 acceptance with waivers deserves scrutiny because public descriptions of acceptance can sound like success when operational control has not been achieved. And the replacement plan's own schedule, total cost and cybersecurity performance need independent validation. Ending a troubled contract may be prudent; it does not prove that the preceding oversight system worked.

Nor should RTX's perspective be erased. It said the delivered system was mission-capable and that it was addressing post-delivery concerns with the government. Without the complete technical deficiency register, contractual acceptance criteria and termination correspondence, a reporter cannot adjudicate every disputed cause between the contractor and the government. The published GAO findings do, however, establish that the government granted a time-steering waiver and identified high cybersecurity risk at acceptance.

What remains unproved and the records that could settle it

The public record reviewed here does not establish a final, comparable all-in accounting for OCX, AEP-M and their future operating costs. It does not show the exact amount of Raytheon's realized profit, any final termination settlement, the full price of all three phases of AEP modernization, or the final amount paid on Lockheed's new task orders. It does not establish that a politician personally chose Lockheed or that political money caused the cancellation. It also cannot establish that the replacement's projected 2028–2030 capabilities will arrive on schedule.

A useful next document request would seek the signed April 17 acquisition decision memorandum and November 2025 options study, redacted as necessary for security; the July 2025 acceptance waivers; the unresolved integrated-test defect inventory; OCX obligations, disbursements and termination liabilities by contract line; and the April, July and September AEP-related task-order documents. Those records would allow an apples-to-apples comparison of cost to complete, performance risk and legal liability. Public versions of the twice-yearly AEP-M progress reports would then show whether the replacement actually meets its milestones.

Congress also has a governance choice. It can ask whether the PNT Oversight Council's recommendations are meaningful when each service can retain separate budget priorities, and whether the proposed single-manager approach would have enough authority to fix that. GAO's three October recommendations call for the Army, Navy and Air Force to assess consolidation and develop plans where it would improve alignment; the department concurred. Agreement to consider recommendations is not proof of implementation.

The accountability verdict

The documented outcome is substantial without embellishment: after years of warnings and billions in program costs, the Pentagon accepted Raytheon's new GPS controller with technical waivers, found additional integration problems and canceled the program. Under the replacement route, Lockheed Martin obtained identifiable new ground-system work while the Pentagon retained responsibility for a technically interdependent and institutionally fragmented modernization effort. The cancellation can be a defensible effort to stop escalating risk and still be a case study in long-running procurement failure.

The question now is not whether one corporation secretly won the entire GPS system. It is whether defense leaders and Congress will publish enough comparable cost, acceptance and delivery evidence to show that the next architecture is better—and whether the military and civilian capabilities deferred by the change actually arrive. That is the test of institutional accountability, not an unproven accusation of corruption.

Source trail and evidentiary boundaries

  1. GAO, Precision Navigation Systems: Challenges Remain with Fielding New Capabilities, GAO-27-107676 (opens in a new tab), October 7, 2026 — acceptance waivers, cancellation chain, replacement timelines, GPS IIIF risks, governance and DOD responses.
  2. Secretary of the Air Force Public Affairs, OCX program cancellation (opens in a new tab), April 20, 2026 — formal decision date, January cost scope and official explanation.
  3. GAO, Weapon Systems Acquisition: Beyond Business as Usual, GAO-26-109135 (opens in a new tab), 2026 — distinct broader cost characterization and program outcome.
  4. GAO, GPS Ground System Development Problems, GAO-15-657 (opens in a new tab), September 9, 2015 — original acquisition decisions and early cost growth.
  5. GAO, GPS Schedule Assessment, GAO-19-250 (opens in a new tab), May 21, 2019 — further delays and cost growth.
  6. Lockheed Martin, GPS Ground Segment Task Orders (opens in a new tab), September 10, 2026 — supplier statement of $96 million and $17.5 million task orders.
  7. Breaking Defense, Lockheed's GPS Ground System Award (opens in a new tab), April 17, 2026 — timing and scope of the up-to-$105 million contract.
  8. Ars Technica, OCX Testing and RTX Response (opens in a new tab), March 2026 — independent pre-cancellation reporting, contractor response and cost-estimate caveat.
  9. Via Satellite, Pentagon Chooses AEP Upgrades (opens in a new tab), April 21, 2026 — named acquisition decision officials and competing control paths.
  10. Defense Department, FY2026 Weapon Systems Acquisition Budget (opens in a new tab), 2025 — contractor roles and pre-cancellation budget assumptions; not a post-cancellation actual-spending ledger.
  11. Air & Space Forces Magazine, April GPS Ground Award (opens in a new tab), April 9, 2026 — independent confirmation of the April 8 award before formal termination.
  12. GPS World, Space Force Terminates OCX (opens in a new tab), April 2026 — industry reporting of the formal announcement; largely dependent on the official release.

Limitations: Sources 2 and 8 were corroborated through indexed reporting when direct page retrieval was restricted. Company awards are not verified disbursements. No right-of-reply requests were sent in this research run. The related Factrail entities and the cover's Commons source are now identified for publisher verification. This manuscript is an author-researched draft, not an independently approved or published investigation.

Evidence

The sources this article cites, claim-bearing material first.

  1. gao.gov · gao.gov/products/gao-27-107676
  2. gao.gov · gao.gov/products/gao-15-657
  3. gao.gov · gao.gov/products/gao-19-250
  4. files.gao.gov · files.gao.gov/reports/GAO-26-109135/index.html
Open sources page →

More analysis

Ranked by shared facts, drivers and people — never by publication date alone.