NMCI: when the Navy handed its shore computers to a con...

In October 2000, the Department of the Navy did something almost no organization its size had tried. It signed over the day-to-day running of its shore-based computers, its desktops, its network, its email, and its help desk, to a single private company. The contract was the Navy Marine Corps Intranet, NMCI, and the company was Electronic Data Systems, EDS. It became one of the largest information technology outsourcing arrangements ever attempted, in government or industry.

This is not a story about a weapons system that never worked. NMCI was a services deal: the government paid a contractor to provide a running network, priced roughly by the "seat," the way a business might pay for managed IT. That distinction matters, because the honest critique of NMCI is not "the technology failed." The technology largely worked, and by the end hundreds of thousands of people relied on it every day. The honest critique is about cost growth, a rollout that ran years behind schedule, chronic user dissatisfaction, and a structural worry that ran underneath all of it: by outsourcing the whole thing, had the Navy given up control of its own network?

Both the critique and the defense are documented, and both belong in the record. This post lays out what the primary sources establish, keeps the several dollar figures distinct, and lets the two verdicts sit side by side.

What NMCI was, and why the Navy wanted it

Before NMCI, the Department of the Navy did not run one network. It ran thousands of them. A Naval Postgraduate School study archived at DTIC describes the pre-NMCI environment as a sprawl of independent legacy networks, built piecemeal by individual commands, with inconsistent security, no unified management, and no enterprise-wide standard. In an era when network security was becoming a warfighting concern rather than an administrative one, that fragmentation was a genuine liability, not a paperwork problem.

NMCI was the answer: consolidate that sprawl into a single, centrally managed, more secure enterprise network covering Navy and Marine Corps shore installations. Rather than build and staff that network with government personnel and government-owned equipment, the Navy chose to buy it as a service. EDS would own and operate the infrastructure and deliver seats to users. The government would pay per seat and hold the contractor to performance targets.

The Government Accountability Office, in report GAO-07-51, confirms the shape of the deal: a single service provider delivering desktop, network, email, and help-desk services across the shore establishment. It was, in structure, an outsourcing contract from the start.

The timeline

October 6, 2000: The Navy awards the NMCI contract to EDS. Government Executive's contemporaneous coverage described it as one of the largest IT contracts ever signed.
October 31, 2002: GAO issues GAO-03-33, "Information Technology: Issues Affecting Cost Impact of Navy Marine Corps Intranet Need to Be Resolved," flagging unresolved cost-impact questions on working-capital-funded activities such as shipyards and depots.
First quarter 2003: EDS recognizes a $334 million pre-tax loss on the NMCI contract, a sign of how hard the deal was for the contractor.
March 2006: A roughly $3 billion, three-year option extends the contract through September 2010.
June 2006: About 303,000 seats are operational, well short of the original schedule.
December 8, 2006: GAO issues GAO-07-51, the central oversight document, finding the program had "yet to meet expectations."
August 26, 2008: Hewlett-Packard completes its acquisition of EDS, and the NMCI work moves to HP Enterprise Services.
September 30, 2010: The original NMCI contract ends. A Continuity of Services Contract with HP bridges the gap while the Navy stands up the successor.
2013: HP wins the Next Generation Enterprise Network, NGEN, contract, worth up to roughly $3.5 billion over five years.

The money, kept straight

NMCI is a case where the single most important discipline is not getting a fact wrong but keeping several correct facts from blurring into each other. There are three or four dollar figures in play, they are all real, and they mean different things. Here they are, each with its year and its meaning.

The original 2000 value: up to about $6.9 billion, including options. According to GAO-07-51, the October 2000 contract was a five-year term with a $4.1 billion minimum value, plus a three-year option worth about $2.8 billion. Add the base and the option and you get a potential total of roughly $6.9 billion. Press coverage at the time often called it a "five-year, $6.9 billion" deal, but that shorthand is slightly loose: the $6.9 billion already folds in the option. The firm five-year minimum was $4.1 billion.

The restructured value: about $9.3 billion over 10 years. After the contract was restructured and the Navy exercised the three-year option, GAO reported in 2006 that the current structure had become 10 years and about $9.3 billion in minimum value. This is a different, larger number than the original $6.9 billion potential, and it is the figure that best represents the NMCI contract as it actually ran.

The program-lifetime bundle: more than $10 billion, when NMCI is combined with its successor. Trade press, including Nextgov, has reported an NMCI program value around $9.9 billion and described NMCI plus the follow-on NGEN growing past $10 billion. That "more than $10 billion" figure is best understood as a combined program lifetime that bundles NMCI with later NGEN spending, not as a single NMCI contract line. It is worth stating, but it belongs with a label attached.

A number that is not an NMCI figure at all: HP's $13.9 billion purchase of EDS. When HP completed its acquisition of EDS on August 26, 2008, the enterprise value of that deal was roughly $13.9 billion, at $25.00 per share. That is the price of buying the whole EDS company, per HP's own SEC filing. It is not the value of the NMCI contract, and the two should never be conflated. HP bought EDS the corporation; the NMCI work simply came along as part of the corporation's business.

One more figure sits off to the side: the $334 million pre-tax loss EDS recognized on NMCI in the first quarter of 2003. That was a loss to the contractor, not a cost charged to taxpayers. It illustrates how difficult the deal turned out to be for EDS, which had also taken Navy-related charges earlier, in late 2002. It is not a government overrun and should not be presented as one.

What GAO found

The spine of the oversight record is GAO-07-51, published December 8, 2006. Its findings are specific, and they are a snapshot of that moment, not a final verdict on the program's whole life.

By GAO's account, after investing about six years and $3.7 billion in NMCI, the Navy had "yet to meet the program's two strategic goals" and had met only 3 of 20 performance targets, 15 percent, that the Navy itself had mapped to those goals. That is the single most-cited number from the program, and it deserves its caveats: the $3.7 billion, the six years, and the 3-of-20 count are all as of December 2006. They are not lifetime totals, and the program improved afterward.

The rollout tells a similar story of ambition outrunning delivery. GAO reported that by June 2006, about 303,000 seats were operational at roughly 550 sites, against an original goal of somewhere between 412,000 and 416,000 seats operational by fiscal year 2004. In other words, two years past the original deadline, the network had reached only about three-quarters of the seat count it was supposed to hit years earlier. Seat counts did keep climbing: by 2008 the network served more than 700,000 users across more than 600 locations.

Satisfaction lagged too. The program carried an 85 percent satisfaction target. GAO reported Navy-measured end-user satisfaction around 80 percent in September 2005, below target, and a broader history in which satisfaction varied by survey date, dipping as low as the mid-50s in Navy-reported figures from late 2002. Among leadership and operators the picture was worse: GAO reported an average commander satisfaction score of 0.76 on a 0-to-3 scale, and network-operator scores as low as 0.00 to 0.33 as of March 2006. These are Navy-reported measures and shift with the survey, but the direction is consistent: for years, the people using and running the network were not satisfied with it.

The earlier report, GAO-03-33 from October 2002, is narrower. It focused on cost-impact management, specifically unresolved questions about how NMCI's costs would land on working-capital-funded activities such as shipyards and depots, rather than on any single headline overrun. Its recommendations were later closed as implemented.

The honest critique

Set against its own goals, NMCI underperformed for a long stretch, and GAO said so plainly. The initial deal, worth up to about $6.9 billion in 2000, was restructured to a 10-year, roughly $9.3 billion minimum by 2006, and the NMCI-plus-NGEN program lifetime topped $10 billion. The rollout ran years behind schedule. Six years and $3.7 billion in, the Navy had met 3 of 20 of its own performance targets. Users, commanders, and network operators reported dissatisfaction that stayed below target for years.

Underneath the cost and schedule numbers was a structural concern that outlasted them. Because NMCI was contractor-owned and contractor-operated, EDS, and later HP, owned the infrastructure and ran the network. Critics warned that this left the Navy dependent on a single vendor and, in a real sense, not in control of its own network. That was not an abstract worry. It was serious enough that when the Navy designed the successor, it deliberately changed the model.

The honest defense

The other side of the ledger is equally documented. The need NMCI was built to meet was real. The DTIC-archived Naval Postgraduate School study describes a genuine operational and cybersecurity liability: thousands of disconnected legacy networks, inconsistent security, no unified management. Consolidating that into one managed, standardized, more secure enterprise was a legitimate warfighting-support requirement, not make-work.

And NMCI delivered on that core purpose. It did standardize and secure Navy shore IT, and it eventually carried more than 700,000 sailors, Marines, and civilians across more than 600 locations on a single managed network. The GAO snapshots that look so harsh in 2006 captured a program still climbing, and it kept climbing after the report. This was not a system that never worked. It was a large, difficult outsourcing experiment that took longer and cost more than promised but ultimately produced the consolidated enterprise it set out to build.

Even the control problem became a lesson learned rather than an abandoned idea. The original NMCI contract ended September 30, 2010, with a Continuity of Services Contract from HP bridging the gap. The successor, the Next Generation Enterprise Network, kept the enterprise model but changed the ownership structure: NGEN shifted from a contractor-owned, contractor-operated network to a government-owned, contractor-operated one. HP won the NGEN contract in 2013, worth up to roughly $3.5 billion over five years. The Navy did not conclude that consolidating its networks was a mistake. It concluded that owning the network while contracting for its operation was the better arrangement, and it built that judgment into what came next.

That is the fair reading of NMCI. Judged as a schedule-and-satisfaction story through 2006, it is a cautionary tale that GAO documented in detail. Judged as a decade-long effort to consolidate and secure a fragmented enterprise, it largely succeeded, and its hardest lesson, about ownership and control, was carried forward rather than thrown away.

Fact-check notes and sources

NMCI outsourced Navy and Marine Corps shore-based IT (desktop, network, email, help desk) to a single contractor, EDS, awarded October 6, 2000: U.S. Government Accountability Office, GAO-07-51 full text; original award coverage, Government Executive (Oct. 2000).
Original 2000 contract: five-year, $4.1 billion minimum, plus a three-year option worth about $2.8 billion, for a potential total of about $6.9 billion including options: U.S. Government Accountability Office, GAO-07-51.
Restructured contract became 10 years and about $9.3 billion in minimum value (as reported by GAO in 2006), distinct from the initial $6.9 billion potential: GAO-07-51.
After about six years and $3.7 billion invested, the Navy had met only 3 of 20 performance targets (15 percent) and had "yet to meet" its two strategic goals; figures are a December 2006 snapshot: GAO-07-51.
By June 2006, about 303,000 seats were operational at about 550 sites, versus an original goal of 412,000 to 416,000 seats by fiscal year 2004: GAO-07-51.
85 percent satisfaction target with about 80 percent end-user satisfaction (September 2005), an average commander score of 0.76, and network-operator scores of 0.00 to 0.33 on a 0-to-3 scale (March 2006); Navy-reported figures that varied by survey date: GAO-07-51.
Earlier oversight, "Information Technology: Issues Affecting Cost Impact of Navy Marine Corps Intranet Need to Be Resolved" (Oct. 31, 2002), on cost-impact issues for working-capital-funded activities: U.S. Government Accountability Office, GAO-03-33.
EDS recognized a $334 million pre-tax loss on NMCI in the first quarter of 2003; a contractor loss, not a cost to the government: Electronic Data Systems Corp., SEC Form 8-K exhibit (2003).
HP completed its acquisition of EDS on August 26, 2008, at roughly $13.9 billion enterprise value ($25.00 per share); this is the whole-company price, not the NMCI contract value: Hewlett-Packard Co., SEC Form 10-Q (FY2008).
Reported NMCI program value around $9.9 billion, with NMCI plus successor NGEN topping $10 billion as a combined program lifetime; a March 2006 option (about $3 billion) extended the contract through September 2010: Nextgov/FCW (2009). Treat the "more than $10 billion" figure as a combined NMCI-plus-NGEN lifetime, not a single NMCI contract line.
NMCI's stated purpose was consolidating thousands of fragmented legacy networks into one managed, more secure enterprise: Naval Postgraduate School / DTIC, "The NMCI Experience and Lessons Learned".
The NMCI contract ended September 30, 2010; a Continuity of Services Contract with HP bridged the gap to NGEN, which shifted to a government-owned, contractor-operated model; HP won the NGEN contract (up to about $3.5 billion over five years) in 2013: Federal News Network (2013); GAO-07-51 landing page.

NMCI: when the Navy handed its shore computers to a contractor in one of the biggest IT deals ever