Exclusive: Pentagon Withholds Internal Report About Flawed $2.7 Billion Intel Program
Why won't senior officials show Congress evidence of a cheaper, off-the-shelf alternative to the military's Afghan battlefield needs?
The Army has spent years defending a multibillion-dollar intelligence system that critics say costs too much and does too little. A new internal report has found that there’s a simple, relatively inexpensive program that could handle many of the same jobs at a fraction of the cost. For the past eight months, though, the Pentagon has kept the report hidden away.
Members of Congress have been asking Defense Department officials to send them the assessment, a copy of which was obtained by Foreign Policy, but the Pentagon has yet to do so. At issue is the Army’s Distributed Common Ground System, expected to cost nearly $11 billion over 30 years and built by a consortium of major Beltway contractors, including Raytheon, Northrop Grumman, Lockheed Martin, and General Dynamics. The system is meant to give troops on the ground an easy way to collect intelligence about terrorists and enemy fighters, and then create detailed reports and maps that they can share with each other to plan and conduct operations. But critics — and even some troops — have long complained that the system doesn’t actually work. They say it’s too slow and hard to use, and that it has left them searching for alternatives in the war zone.
The system’s high cost and technical failings prompted a search for other options. Palantir Technologies, a fast-growing Silicon Valley firm, told the Pentagon that its off-the-shelf systems could accomplish most of the same tasks but cost far less — millions, rather than billions. The Marine Corps, Special Operations forces, the CIA, and a host of other government agencies already use it. Army officials, though, said Palantir wasn’t up to the job. Now, a 57-page report by the Pentagon’s acquisitions arm basically says the Army was wrong to dismiss the Palantir system. The study instead gives Palantir high marks on most of the Army’s 20 key requirements for the intelligence system, including the ability to analyze large amounts of information, including critical data about terrorist networks and the locations of explosive devices, and synchronize it in a way that helps troops on the ground combat their enemies more effectively.
Palantir "can be utilized to partially meet DCGS-A requirements," the report concludes, using the acronym for the Distributed Common Ground System.
The report is likely to sharpen concerns about the Distributed Common Ground System, which has been facing mounting criticism on Capitol Hill. Rep. Jim Moran (D-Va.), one of many long-time detractors, had asked the Pentagon for its findings as recently as last month.
"It’s a scandal that commercially available, battlefield-proven technology is ready to go at a fraction of the billions of dollars the Pentagon is spending to build a similar analysis tool in-house," Moran said in a statement to FP. "I appreciate [Under Secretary of Defense for Acquisition, Technology and Logistics] Frank Kendall taking this issue seriously, and look forward to hopefully resolving it once and for all when the long overdue report’s findings are finally released."
The report, commissioned roughly one year ago, won’t deal a fatal blow to the controversial Army program. But it raises new questions about why the service is wedded to its own system and why officials have been so quick to dismiss Palantir’s capabilities, especially at a time when the Pentagon’s budget is shrinking and Congress is pressing Defense Department officials to find ways of saving money.
An Army official referred a query from FP to the Pentagon’s acquisition arm. Maureen Schumann, a spokeswoman, acknowledged that the Defense Department’s acquisition officials had commissioned a "top level analysis" last summer and presented its findings to officials internally. This spring, Schumann said, Kendall will respond to queries from members of Congress on the findings of that analysis. But it remains unclear if the assessment itself will ever be released. Officials at Palantir declined to comment for this story.
The report that was obtained by FP is steeped in bureaucratese, but at its heart, it says the Palantir system it assesses could play a key role in Afghanistan or future warzones. The assessment, completed last summer, was first commissioned by Kendall after Moran and others asked more pointed questions about what the Palantir system could and could not do for the Army. Ultimately, the hope is to obtain an effective intelligence system that will help troops hunt terrorists, predict insurgent attacks, and plan safe missions. With the wars of the future likely to be guerilla wars like Iraq and Afghanistan, the Army intelligence program will only grow more important in the years to come.
The report consists of a detailed set of charts, graphs, and analyses. It doesn’t say that Palantir could replace the Army system, and its authors didn’t conduct a head-to-head comparison of the two. But it concludes that Palantir, which has collected legions of fans in national security circles and has contracts with the CIA and FBI, performs "very good" or "excellent" on most key requirements, including some that Army brass had long insisted the company was unable to fulfill.
Palantir has "a rich suite of [applications] applicable to" the Army system and has "robust" capabilities to collect many different kinds of information that can be used to create intelligence reports and allow troops to share information with each other, the review finds.
Army officials have long complained that Palantir cannot be used with other applications that are already incorporated into the Army system. An Army spokesman compared the problem last year to being able to download and read a document, but not make changes or be able to share the new version. That lack of "interoperability" has been a key reason why the Army has said that it couldn’t move to Palantir’s cheaper system.
Palantir is just one application, and "is not interoperable with all the other apps right now, so that’s the problem," Army Chief of Staff Gen. Raymond Odierno said in May 2013. In July 2012, Lynn Schnurr, the chief information officer for the Army’s intelligence office, said, "Palantir addresses a segment of the capabilities" of DCGS-A and "does not interoperate" with certain command and control systems that the Army has to use. And in December of that year, Maj. Gen. Harold Greene, a top Army acquisitions official, said the lack of interoperability in any system was a "red line."
But according to the Pentagon report, dated July 2013, Palantir received a score of three, indicating "good," on a one-to-five scale measuring interoperability. And its front-end, or user interface, allows people to collaborate on documents at the same time, the report concluded.
That’s not the same as saying Palantir is completely interoperable with the Army DCGS-A, but it undermines the assertions by Odierno, Schnurr, and others who say Palantir’s alleged inability to properly work with the Army system is a key weakness that prevents the military from switching to it.
Palantir’s overall report card consisted of "very good" and "excellent" on 11 of the 20 criteria. It received a rating of "good" in three areas, and "minimal" in four others, including the ability to synchronize surveillance and reconnaissance information, such as video feeds from drones. The report didn’t rank Palantir in two of the 20 areas, including its ability to handle signals intelligence, or intercepted communications of the kind captured by the National Security Agency.
Rep. Duncan Hunter, a California Republican, got into a public spat with Odierno last year over the program that helped put the controversy into broader public view. In a video that quickly went viral, Hunter told Odierno and Army Secretary John McHugh that he wanted to cut through Army bureaucracy to help troops who were complaining the intelligence system the Army provides was insufficient. "What we want is the best for the warfighter in the most economic way possible, the most efficient and least bureaucratic way forward to get into the warfighter’s hand what they need," Hunter said during a House hearing in April 2013.
But Odierno, a passionate commander who can get emotional when it comes to issues of protecting his own troops, didn’t take the implicit criticism lightly. "I object to this," he told Hunter in a departure from the typically tame back-and-forth of hearings among public officials. "I’m tired of somebody telling me I don’t care about our soldiers and that we don’t respond."
Pointedly referring to the locations in Afghanistan where the system is already in use, Odierno pointedly added that he "could go to 30 places that tell me it’s working tremendously. Is it perfect? No. Will we have iterative processes that can inject more technology? Absolutely."
Hunter told FP in a statement that the Army’s system still has "wide gaps in capability" that can easily be filled with the software that is available commercially, like Palantir. "That’s something that the Army has been stubbornly resistant to acknowledge, even though Palantir as a plug-in would solve the Army’s problems and deliver soldiers a whole new set of capabilities that they have yet to acquire," he said. "From day one, the problems with DCGS have been apparent but so too has the solution."