Expanding use of large-scale neural recordings has led to paradigm shifts in how neural computation is understood [1]. Technologies used for collecting these datasets face fundamental limitations ranging from power dissipation, penetration depth, and resolution [2] while facing additional practical challenges related to sensor integration with the subject or patient. While the ultimate goal of simultaneously recording from every neuron in the brain is far out of reach, more modest advancements still need to be made to ensure the proper neural coverage is achieved to construct a sufficient approximation of the latent space [3], underlying dynamics [4], or other form of systems based analysis. Despite the relative advantages of imaging and the successes already achieved [5], it is our view that microelectrode arrays will, in the medium term, surpass the capabilities of functional imaging by enable recordings across the depth and breadth of the brain at high temporal resolution and sufficient spatial resolution. However, this is predicated on advancements in material choices [6] and/or moving beyond the traditional time division multiplexing techniques [7]. In this work we present the theoretical justification for and practical demonstration of mutual capacitive recordings of neural activity using code division multiplexing. We began with a finite element simulation of a Hodgkin-Huxley neuron using the Nernst-Plank-Poisson formalism for both the intra and extracellular ion concentrations coupled with electrodynamics to verify the physical feasibility of the measurement. Following proof of concept experiments, we developed (1) an application specific integrated circuit (ASIC), with a power consumption of 9.7 µW/electrode when sampling at 20 kHz, and (2) a 1,024 electrode nanocapacitor array with a density of 10k electrodes/mm$^2$. In vitro validation confirmed both single unit and low frequency activity can be detected. The main advantages of capacitive transducers are (1) the scalability of the electrodes vs. the interconnects and amplifiers, (2) a reduced impact on sensitivity with miniaturization, (3) and amenability to various fabrication techniques. Though in a nascent stage, the continued development of this and similar technologies will be foundational in further answering how activity is coordinated across the brain to generate behavior.