The recent demonstration of computational speedup achieved by a noisy intermediate-scale quantum circuit, compared to a classical supercomputer, has accelerated even more the pursuit for the implementation of a universal quantum computer. The results achieved by the quantum hardware and algorithm communities have moved the industry-academia alliance one step closer to quantum computing’s “Hello World” era. Among the viable systems for the physical implementation of a quantum computer, silicon spin qubits are at the forefront of quantum research today, partly due to their exceedingly long coherence times and a reduced on-chip footprint. Furthermore, massive parallel fabrication of spin qubit hardware in CMOS foundries shows the potential for large-scale production. In this thesis, I analyse and compare two different silicon spin qubit platforms; one is university-fabricated while the other is based on quantum dots produced in a CMOS foundry. I study their behavior in the few-electron regime using advanced radio-frequency techniques, enabling fast charge sensing and gatebased dispersive readout, the latter being a strong candidate for achieving compact readout and wiring of a large-scale quantum computer. Using the university-fabricated device, I show the ability to read out the singleelectron spin configuration in 24 μs, a fundamental requirement for the implementation of a spin qubit. Nevertheless, the reproducibility of these devices is low and their fabrication is complicated for a typical academic cleanroom. Motivated by the recent demonstration of a CMOS spin qubit, I then analyse the performance of a foundry fabricated two-dimensional array of CMOS quantum dots. Developing a hybrid dispersive charge-sensing technique, I demonstrate for these devices all the major functionalities of state-of-the-art university fabricated quantum dot devices. By harnessing the two-dimensionality of the system I report in the time domain deterministic single-electron movement and charge swaps within the array.