They are as DACs go, like chalk and cheese. They use different reconstruction filter techniques and very different architectures.
Digital to Analogue conversion as a consequence of the sample theory is an inexact process… it is ultimately lossy. The design engineers focus on specific area they wish to address… this results in different performances upon critical examination … you effectively choose the method and consequently DAC that works with how your brain best listens to audio.
The NDAC and ND555 use traditional Texas Instruments ladder variant DAC chips (PCM1704K) with IIR low pass reconstruction filtering… and focus on the electronics to get the best sonic performance from those chips.
The Chord Hugo TT2 uses a proprietary Pulse Array converter, using FIR low pass reconstruction filtering and using very simple output and current to voltage electronics to reduce colouration. (But at possible cost of more device variation)
Each reconstruction filtering method has pros and cons… especially to do with processing noise and demands on power supplies, phase distortion (transient smearing) and timing resolution.
Chord focus on a bespoke FIR windowing algorithm (WTA) using an ultra low power FPGA processor whereas Naim who do things quite differently focus on implementing its recursive IIR filter in an Analog Devices DSP processor engine (SHARC)
This all results on different renderings upon critical listening, whilst with casual background listening one might notice no difference (other than possible tonal presentation which is not really related to the DAC method)
Only you can can decide which rendering method you prefer as it will almost certainly vary from person to person, as well as resolution of down stream system and speakers.