U.S. IT-industry is the world leader in AI development. However, because of the negative experience in funding this activities there have been two «AI winters», and Federal Government hadn’t participated in such R&D since the late 1980th. Nevertheless, with the rapid AI technology development the White House have heard specialists’ opinions in the end. Creation of 2016 National Artificial Intelligence Research and Development Strategic Plan symbolized that the States have come back to this S&T area. However, in spite of U.S. leadership in the AI and more generally in information technology sphere U.S. state R&D system is in focus of debates about its efficiency. There are questions about the system structure, its institutes and funding rules. All of them have roots in circumstances of their origin. Of course, for the members of Congress the main point is the rules of participating of organizations in R&D because it’s a money question. It seems that nobody asks if the main issue is not in particular rules but in the more fundamental aspects such as consequences of the U.S. institute of lobbyism existence and lack of effective long-term planning of R&D. The article considers U.S. state R&D system, which intended to provide coordination in such investigations and keeping U.S. technology leadership.