October 2018
Case
HBS Case Collection

Shield AI

By: Mitchell Weiss and A.J. Steinlage

Format:Print
| Language:English
| Pages:21

Abstract

Shield AI’s quadcopter – with no pilot and no flight plan – could clear a building and outpace human warfighters by almost five minutes. This was not to say that it was better than the warfighters or would replace their jobs, but it was evidence that autonomous robots deployed in this fashion could keep up with the pace of operators, serve as a force multiplier, and help protect civilian and service member lives. But was it evidence that the Shield AI team should ask their newest potential customer for a contract 50-60 times more than their only two contracts to date? Would asking for what a system of coordinated, exploring robots really cost - $50 to $60 million - scare off their U.S. government customer? No one would blink if Lockheed Martin, Boeing, or other large defense companies proposed a $60 million effort to do the same, but would it make Shield AI, barely two years past founding, look arrogant, ignorant, or both?