To resolve today's spectrum scarcity problem, cognitive radio network is widely expected to usher in the next wave in wireless communications. This paper has developed a high-speed sphere decoder for multiple-inputs multiple-outputs (MIMO) transceivers used in cognitive radio networks. First, a novel sphere decoding algorithm is proposed which increases the decoding speed by 30% without any noticeable performance degradation. Based on the proposed algorithm, single universal sphere decoder is designed for MIMO transceivers with adaptive modulation and variable number of antennas. Finally, a novel parallel structure is developed which increases the decoding speed n-fold by using n sphere processors. The implementation results into field programmable gate array (FPGA) device indicated that a 20-parallel decoder attained a constant speed of 500 Mbps, even if the number of antennas or the order of modulation are changed in cognitive radio networks. The de signed sphere decoder can be used in various wireless networks with MIMO antenna processing and fast adaptive modulation.