The chapter explores the issue of creating distributed multi-agent systems based on the Mixture of Experts (MoE) architecture in the context of sixth-generation (6G) wireless technologies. The problem lies in ensuring effective coordination and interaction of millions of experts deployed across various hardware platforms using high-speed wireless networks. The general steps to solve this problem include scaling MoE systems to a system of systems, as well as the introduction of intelligent resource management mechanisms and ensuring high performance through the use of digital antenna arrays and digital beamforming to mitigate interference. The article proposes a solution based on combining distributed systems of experts interacting via high-speed wireless technologies to optimize performance and reliability. The main idea of the approach is to use the MoE architecture, which allows the distribution of computational resources depending on the tasks, ensuring high flexibility and scalability of the system. The approach has been checked by analyzing the integration capabilities of artificial intelligence technologies for adaptive network management and using high-speed wireless technologies. Additionally, a modification of Orthogonal Time Frequency Space (OTFS) modulation based on Non-Orthogonal Frequency Division Multiplexing (N-OFDM) has been proposed to increase spectral efficiency and reduce packet duration by using overlapping impulses. The research results demonstrate the possibility of achieving high performance, reliability, and resource optimization, making the proposed approach promising for use in various industries.