AI and Cyber in the Administration’s FY 2020 Research and Development Budget Priorities

Published: August 08, 2018

Federal Market AnalysisBig DataBudgetCybersecurityInformation TechnologyOFFICE OF SCIENCE AND TECHNOLOGY POLICY (EXECUTIVE OFFICE OF THE PRESIDENT)Research and Development

Expect billions of dollars to flow into artificial intelligence and cyber solutions.

At the end of July 2018, the White House Office of Science and Technology Policy, released its list of research and development priorities for fiscal year 2020. This list can help industry understand technology areas in which the administration wants agencies to invest. The memo notes that the U.S. “requires prioritized investment in artificial intelligence, autonomous systems, hypersonics, a modernized nuclear deterrent, advanced microelectronics, computing, and cyber capabilities. Advanced cyber capabilities at scale require investment in new computing and technology paradigms, including adaptive and automated defensive measures.” Given that AI/autonomy and cyber will be big priorities in the future I’ve tried below to provide some numbers concerning what the federal government could be spending on these technology areas in the years to come.

Artificial Intelligence/Autonomous Systems

Although autonomy, AI, and machine learning are unique in their own ways, when it comes to parsing the market the data for all three is wrapped up together. Deltek’s Federal Market Analysis team estimates, based on an analysis of requirements it can identify as “big data,” that federal agencies spent at least $816M on big data solutions, including services, in FY 2016. Analysis of FY 2017 data is underway for publication in October.

In terms of AI specifically detailed data for planned investments is only available for the Department of Defense and even that must be considered in context. The following chart, re-posted from a piece of analysis published in June, shows those numbers. Multiple organizations across the DOD invest in projects that intend to use AI and machine learning in one way or another, typically to develop a kind of autonomous capability for unmanned platforms.

The thing to remember is that these numbers show the budgets for programs that intend to use AI or machine learning as part of the effort. Not all dollars identified will go to AI, but a good portion will, so think of these numbers as the top end for FY 2019. When the FY 2020 budget request comes out we’ll see what the figures look like then. For the time being consider these numbers alongside Pentagon estimates that it plans to spend $1.7B through FY 2023 standing up the new Joint Artificial Intelligence Center.

Civilian agency spending on AI remains a wild card with the best available data being in the annual Networking and Information Technology Research and Development (NITRD) budget request. At present documentation is only available for FY 2018, for which NITRD agencies (including DOD) requested $2.3B for programs with a big data/AI/High Performance Computing component.

Lastly, Deltek FMA is forecasting that federal agencies will spend $2.6B on big data solutions, including AI and machine learning, in FY 2020, rising to $3B by FY 2022.

Cyber Security/Warfare

Definitely the area gathering the most attention in recent years, Deltek FMA calculates that as of FY 2016 identifiable federal agency prime obligations for cyber were running at just above $4B per year, with the majority of those dollars going to services vs. products. Spending on cyber has been advancing on the order of an additional $100M to $300M per year as agencies try to get a handle on emerging and ever changing threats.

In terms of forecasting, spending on cyber will continue to be one of the fastest technology growth areas, rising from $12.2B in FY 2020 to $12.9B in FY 2022. The gap between identifiable spending vs. forecast contractor-addressable spending is accounted for by the fact that not all of the government’s spending with contractors can be identified in reported spending data. A tremendous volume of spending on cyber goes through contracts that are not explicitly cyber-related in their statements of work. The best FMA can do is estimate what that spending is.

Of course, only a portion of addressable dollars will go to R&D-type efforts. The important thing to keep in mind is that improving cyber remains a key goal for the Administration so dollars will continue flowing into related work as time goes on.