Intelligent reflecting surfaces (IRS) are believed to be the pledging technology for efficient signal propagation in future networks. The intelligent reflecting surfaces are also believed to enhance physical layer security in the future wireless and wired networks. One of the major drawbacks of energy-efficient and secure networks is the incorporation of a greater number of RF chains. To overcome this an IRS can prove to be a viable solution. In this paper, IRS-inspired generalized frequency division multiplexing-non-orthogonal frequency division multiplexing (GFDM-NOMA) is proposed to obtain an effective solution to increase the coverage of communication and energy efficiency (EE). In the proposed work the main focus is laid on the IRS-assisted GFDM-NOMA system and an algorithm for EE is proposed to achieve a fine trade-off between minimization of total power consumption and sum data rate maximization. Hybrid optimization of the beamforming vis-a-vis transmit at the BS and the reflecting beamforming at IRS is proposed to exploit the EE of the system. The theoretical bit error rate (BER) expression for downlink IRS-assisted GFDM-NOMA systems is derived mathematically and is authenticated through the simulation outcomes. Joint performance improvements which include achievable sum-rate and out of band radiation (OOB) are as well obtained. Simulation outcomes reveal proposed algorithm yields better performance in comparison to the traditional IRS, NOMA, and GFDM systems.