The United States Department of Defence is a huge organisation. With its vast workforce and unparalleled military budget, the organisation not only supports countless service members and civilian employees but also acquires and maintains an impressive array of equipment that is beyond the scope of a single paragraph.
Coordinating all of this can be quite overwhelming. The different agencies of the DOD confront the challenge of making strategic decisions, managing resources, and adhering to budget constraints for potentially contested events, such as natural disasters or hostile entities.
Defcon AI and AIR Mobility Command
After two years of incubation, Defcon AI, a startup based in Virginia, has successfully raised a $44 million seed round to tackle this challenging problem.
Look at the Air Mobility Command, which is a command within the U.S. Air Force. When operators plan airlifts, they need to take into account a wide range of factors. These include the availability of aircraft, the number of crews needed, suitable rest areas for the crews, refuelling options, appropriate airfields, and locations for cargo handling.
Paul Selva, Defcon AI’s co-founder, chief strategy officer, and retired U.S. Air Force general, reportedly told Eltrys that the company has developed a software system that enables operators to simply set parameters and then let the software do its work. The software effectively works with the given parameters or inputs to generate an optimal plan, which includes cost tables, resource requirements, and a schedule.
Planning under normal circumstances can be challenging, but during a crisis, defence operators face the added pressure of allocating resources without any time to spare. Defcon AI is here to help.
“I’ve had experience in all the roles that have a significant impact,” Selva said. Throughout his extensive military tenure, Selva assumed various prestigious positions, including commander of the Air Mobility Command, responsible for supervising a significant portion of the Air Force’s airlift aircraft fleet. Afterwards, he rose to the position of commander of the U.S. Transportation Command, responsible for overseeing transportation missions worldwide, encompassing various modes of transportation such as ships, trucks, trains, and more. Prior to his retirement in 2019, he received a nomination from President Barack Obama to serve as the Vice Chairman of the Joint Chiefs of Staff.
In 2022, he played a key role in the establishment of Defcon AI alongside Yisroel Brumer and Grant Verstandig, who are also founding partners of Red Cell Partners. Red Cell has a fascinating model in which they not only make internal investments but also incubate companies like Defcon AI. They have a knack for spotting talented entrepreneurs who have the potential to lead these ventures. At times, entrepreneurs come to Red Cell even before establishing their own company, and the firm takes care of tasks such as board building, legal matters, HR, and finance as the company expands.
Silva explains that Air Mobility Command expressed a mission need that the industry was not addressing, prompting the establishment of Defcon AI. Our initial belief was that we could tackle this problem using mathematical and software approaches, but we needed to approach it from a different angle.
Both Brumer and Verstandig have impressive backgrounds. Prior to joining Red Cell, Brumer held a significant position at the Pentagon as the acting director of OSD/Cape (Office of the Secretary of Defence, Cost Assessment, and Programme Evaluation). In this role, he served as the key figure responsible for analysing data and overseeing the budget submission process for the Department of Defence. Verstandig is a successful entrepreneur who has successfully incubated or grown businesses such as Rally Health and defence startup Epirus through Red Cell.
Brumer asserts that Defcon AI concentrates on addressing a highly complex problem. Utilising a variety of algorithms, such as machine learning and mathematical optimisation, the startup’s system effectively simulates scenarios and produces optimal logistical solutions. During the early phases of product development, Defcon AI employed reinforcement learning algorithms that do not rely on data. However, the company has since started incorporating a growing amount of DOD data to enhance the software’s capabilities. Operators have the option to simulate potential disruptions to their operations and can instruct the system to prioritise different factors, such as speed or cost effectiveness.
Within just two years of its establishment, the company achieved an impressive milestone by securing approximately $15 million in government contracts and successfully deploying a production version for a real-world operation with Air Mobility Command. The company is currently working on certifying the software to handle classified and secret information. This will not only expand its use in the DOD but also allow it to process a larger amount of data. In addition to its planning and simulation software, the company is now incorporating trucks, trains, and ships into its offerings.
Defcon AI shows no signs of slowing down. The company is seeing a significant increase in applications within the DOD, allowing its software to have a tangible impact on operations. Brumer mentioned that there is also a high demand for the product from the private sector. In general, the company emphasises the importance of collaborating with end users to achieve a superior product and gain a real advantage in competitive scenarios.
“Operational planners are actively evaluating risk for their commanders,” Selva stated. “Working closely with a discerning audience, such as decision support tool users, can greatly enhance the final result.” 
