Thoughts on JEDI and DOD's Cloud Budget Estimate

Back at the beginning of May, the Department of Defense’s Chief Management Officer published a report requested by Congress on the upcoming Joint Enterprise Defense Infrastructure (JEDI) procurement. Buried at the tail-end of that report is an estimate of the DOD’s proposed cloud computing budget over the next five fiscal years. Today’s post takes a look at those numbers and compares them to cloud spending by the DOD over the last three fiscal years that Deltek GovWin has been able to verify.

First, a little background. When the DOD announced in 2017 that it would procure a JEDI cloud estimates of the pending contract’s ceiling value hovered in the “multi-billions of dollars” range. This made sense given the scale of the effort that the DOD is proposing. Then, in February of 2018, a more exact figure of $10B appeared in an analyst estimate. This number has been continually repeated ever since despite the fact that the DOD itself never stated it or confirmed it. The department even responded to a question about JEDI’s ceiling value in the draft RFP by saying: “The ceiling will be provided in the final RFP.”

Then there are the proposed cloud budget numbers provided by DOD which show that the department requested $230M for cloud computing in fiscal 2018 and it intends to request just over $1.6B overall in the five years from fiscal 2019 to 2023. None of these figures is anything close to the $10B estimates being talked about, so what is a realistic ceiling value for JEDI?

The DOD’s proposed cloud budget numbers break down as follows:

The JEDI report to Congress states that there is no specific line item budget for JEDI cloud, but that the DOD intends to “reconcile” application budgets to modernize legacy applications, presumably using cloud-based capabilities, so the hint is that DOD will reprogram dollars for the JEDI effort that are not represented in the estimate. The question becomes then just how wide the margin is between the numbers that DOD provides and what might be spent in the future. I have no way of knowing the answer to this question other than to provide two data points that offer insight into DOD’s verifiable cloud spending and suggest what DOD could spend based on an analysis of its Research, Development, Test, and Evaluation (RDT&E) and Procurement documents.

Actual DOD Cloud Spending

The latest data from the Deltek Cloud Database now includes a complete set of spending data for fiscal years 2015 through 2017. These numbers show the following.

Taking into account that this spending data is for commercial contracts Deltek has confirmed are cloud efforts we can see that DOD’s estimated $230M cloud budget for FY 2018 is pretty much in line with its spending so far. The discrepancy widens in fiscal 2019 when DOD breaks out what it intends to spend on commercial clouds versus “Other” clouds, presumably on-premise government clouds.

We can make a couple of observations based on this data. First, the DOD currently underestimates the amount of money it spends on commercial cloud capabilities. Verified commercial cloud spending across DOD already hovers around $200M per year, not the $100M mark proposed in the DOD cloud budget. So, if the DOD intends to spend more on commercial cloud we already need to roughly double the numbers they have already provided. Second, the DOD still spends a great deal of money on its own private clouds. If JEDI proceeds the way I suspect it will (i.e., as largely a tactically-based combat-focused cloud), then it will be interesting to see how DOD classifies that spending – either as commercial or as “Other.” The JEDI congressional report suggests the former.

Cloud in DOD’s FY 2019 RDT&E and Procurement Budget Requests

Now we need to compare the numbers suggested by DOD to those derived from the DOD’s FY 2019 budget request. The figures below are from DOD programs that intend to use cloud for one purpose or another. Breaking out spending specifically on cloud is impossible so keep in mind that these budget estimates include all program-related expenditures.

The best we can say at the moment, based on the differences between the DOD’s current cloud spending data, its proposed budgets, and the cloud-related RDT&E/Procurement program budget data, is that JEDI’s annual value is likely to fall between $200M and $1B per year. If, however, a goal of JEDI is for the DOD to achieve efficiencies that allow it to reduce costs DOD’s IT spending must go down for JEDI to be considered a success. We also need to keep in mind that spending will continue to go to milCloud 2.0 and to the other infrastructure efforts that the military departments currently have underway. Lastly, Deputy Secretary of Defense Patrick Shanahan stated last April that JEDI would account for 10-20% of the DOD's cloud budget. If we use the numbers provided by DOD in the Future Years Defense Estimates table above this means JEDI will be worth between $160M and $320M through FY 2023. Double those numbers for a 10 year period of work and the most the contract could be worth is $640M!

Big ceiling values are often announced for cloud contracts, but actual spending against them falls far short of the advertised totals. My best guess, therefore, is that JEDI’s ultimate ceiling value will come in between $5B and $7B, but spending on it will run between $200M and $400M per year. Either way, I doubt the total value of the JEDI contract, if Congress allows it to be awarded, will ever be $10B.