For the long of its history, power grid relied on large electrical generation plants with stable and predictable output. The integration of renewable energy changed this. With wind and solar injecting intermittent power from thousands of distributed nodes increasing the variables that the grid could manage. Consequently, advanced algorithms such as model predictive control and neural network based control become necessary to maintain stability and performance under such conditions. In this context, this paper explores efficient computing solutions for implementing these control algorithms into Field Programmable Gate Arrays (FPGAs). FPGAs are reconfigurable integrated circuits capable of offering high-performance parallel processing at the lower cost of energy consumption. Three hardware architectures for MVM are compared, with focus placed on the trade-off between processing latency and silicon area. Asymmetric quantization is also studied to assess its effect on hardware resource usage and model accuracy. Afterward, the suitability of each architecture for renewable energy integration is discussed. The broader goal of this study is to promote the adoption of FPGA-based implementations as a practical path to low-latency, intelligent control at the grid edge, increasing the computational and energy efficiency associated with modern renewable energy.