The military's next-generation GPS ground control system, OCX, has consumed $8 billion over 16 years and still isn't operational. Program officials admit they're "still considering how to move forward" with what has become one of the most troubled defense software projects in recent memory.

This is a case study in why "just throw more money at it" doesn't work in software. $8 billion buys you a lot of talent, but it can't fix fundamental architecture mistakes.

Someone needs to ask why we're "still considering how to move forward" after 16 years.

The Next Generation Operational Control System (OCX) was supposed to modernize GPS ground control, replacing aging systems from the 1990s. It was meant to improve security, add new capabilities for military GPS satellites, and provide better jam resistance.

The contract was awarded in 2010. The original completion date was 2013. It's now 2026, and OCX still isn't in full operational use.

According to Ars Technica, the program has been plagued by software complexity, changing requirements, cybersecurity failures, and contractor performance issues. The prime contractor, Raytheon (now part of RTX Corporation), has consistently missed deadlines and delivered software that failed testing.

At one point, cybersecurity auditors found such severe vulnerabilities in the OCX code that the entire system would have been immediately compromised if deployed. That was years ago, and subsequent fixes have led to more delays and cost overruns.

From a software engineering perspective, this is what happens when you try to build a massively complex system without iterative delivery and continuous integration. Defense contracting traditionally uses a "waterfall" model: define all requirements upfront, spend years building, deliver at the end.

But GPS requirements kept changing. New satellites were developed with capabilities OCX needed to support. Cybersecurity standards evolved. Enemy jamming techniques advanced. By the time Raytheon delivered code, it was already obsolete.

Meanwhile, the cost kept climbing. $8 billion is an almost incomprehensible amount of money for a software system. For comparison, SpaceX developed the entire Falcon 9 rocket and Dragon spacecraft for less than $1 billion.

The money isn't buying results—it's paying for a process that doesn't work.

Defense officials have tried various interventions. They brought in outside software experts. They implemented agile development practices (sort of). They restructured the contract. Nothing has fundamentally fixed the underlying problem: OCX was architected wrong from the start, and trying to retrofit fixes into a broken foundation is nearly impossible.

The phrase "still considering how to move forward" is particularly telling. After 16 years and $8 billion, program leadership doesn't have a clear path to completion. That's not a technology problem—it's a management and procurement problem.

The military has backup plans. The old GPS ground control system, while outdated, still works. And the Space Force is exploring alternative approaches, including incremental upgrades and potentially starting over with a new system design.

But that raises the obvious question: if we're going to start over anyway, what was the $8 billion for?

Defense contractors have little incentive to deliver on time and under budget when contracts are structured as cost-plus. Raytheon gets paid for the work, whether or not it results in a functioning system. Delays and overruns mean more billable hours.

From a technical perspective, modern software development has solved many of these problems. Companies routinely build complex distributed systems with millions of lines of code and deploy them reliably. But they do it with rapid iteration, automated testing, and real accountability for results.

Defense procurement is still stuck in a model where contracts are awarded based on lowest bid and connections, not technical competence. Where requirements are frozen upfront, and changes require years of renegotiation. Where software is treated like hardware—something you build once and deploy.

That doesn't work for software. It especially doesn't work for mission-critical systems that need to evolve with threats.

The technology to build a modern GPS ground control system exists. The expertise is available. What's missing is a procurement process that actually delivers results instead of just consuming budget.

Sixteen years. Eight billion dollars. Still doesn't work.

At some point, you have to ask whether the system is designed to fail.