I have used liquid cooling-for thermoelectric projects. I even made my own copper water blocks. I'm able to keep temps about 1.5°C above the inlet water temps with a ~100W heat load. Using tap water in winter, I can get "hot" side temps of around 5°C. With thermoelectrics, every degree you lower hot side temps gives you about 0.70° lower cold side temps with a single stage, about 0.5° with two stages. That's how I got my temperature test chamber to approach -60°F. I may build a new version when I have the time using vacuum insulation panels and three stages. This might get me under -100°F, which would let me test stuff which is used in Antarctica.
Don't much see the point for CPUs. I could use an old A/C evaporator and pump to recirculate the water, maybe get a temperature rise of 2°C for every 100 watts, but that's overkill. The goal is to keep the CPU under maybe 60°C. Air cooling generally works for that. The setup I described might be useful for a 2 kW CPU. Thankfully we haven't reached that yet, but at the rate Intel is going you never know.