Sixth generation (6G) and other next-gen networks may greatly profit from non-orthogonal multiple access (NOMA). Promising new technologies that could be used to increase capacity are cognitive radio (CR) and multiple access. With the combination of cognitive radio networks (CRNs) and network-oriented multi-access NOMA capabilities, a communications revolution is expected. The study shows a new way to improve the NOMA power domain’s average capacity in the downlink (DL) when dealing with an uncooperative interweave CRN (UICRN). This occurs when a primary user (PU) is unable to receive data over the dedicated channel due to the presence of noise or interference. The four network topologies use quadrature phase shift keying (QPSK) modulation over 70 MHz bandwidths (BWs). The transmit powers, user distances, and power location coefficients vary. The network architectures include massive multiple-input multiple-output (mMIMO) (128 × 128, 256 × 256, and 512 × 512), among others. In the performance study, channel instability, successive interference cancellation (SIC), and frequency-dependent Rayleigh fading are all looked at. The average output of the given model is found using MATLAB. Adding 128 × 128, 256 × 256, and 512 × 512 mMIMO to DL NOMA boosts average capacity by 60, 65.5, and 66.7% and improves by 70.1.1, 76.4, and 77.6% using the UICR NOMA model for the best user. mMIMO and CRN significantly improve average capacity performance. The produced expressions match the Monte Carlo simulations, validating our findings.