What I think?
To be real with you, I think that for most of the gaming CPUs an AIO is just a pure aesthetic upgrade. Many Air Coolers like the Thermalright Peerless Assassin 120 run equally or just very slightly louder than an AIO. Further, since there are fewer moving parts , air coolers last longer and a fan fail means you replace the fan unlike an AIO where buying a new one often beats "repairing your pump". But let's look at data just to make sure I am not blabbering some random BS.
| Category | CPU Model | Air Cooler Temp (°C) | AIO Cooler Temp (°C) | Temperature Delta (Δ) | Multi-Core Performance / Score Impact |
|---|---|---|---|---|---|
| Ryzen 5 | Ryzen 5 7600X | 88°C | 74°C (240mm) | -14 °C | +1.5% Cinebench score |
| Ryzen 5 9600X | 71°C | 64°C (240mm) | -7 °C | <1.0% difference | |
| Ryzen 5 7600X3D | 77°C | 69°C (240mm) | -8 °C | 0.0% gaming impact | |
| Ryzen 5 5600X | 84°C (Stock) | 61°C (240mm) | -23 °C | +2.0% sustained boost | |
| Ryzen 5 5600X3D | 81°C | 72°C (240mm) | -9 °C | 0.0% gaming impact | |
| Ryzen 7 | Ryzen 7 9800X3D | 80°C | 72°C (360mm) | -8 °C | +1.2% Cinebench score |
| Ryzen 7 7800X3D | 82°C | 74°C (360mm) | -8 °C | 0.0% gaming impact | |
| Ryzen 7 7700X | 93°C | 81°C (360mm) | -12 °C | +3.5% multi-core boost | |
| Ryzen 7 9700X | 72°C | 65°C (240mm) | -7 °C | <1.0% difference | |
| Ryzen 7 5800X3D | 83°C | 74°C (240mm) | -9 °C | 0.0% gaming impact | |
| Ryzen 9 | Ryzen 9 9950X | 93°C | 84°C (360mm) | -9 °C | +4.2% Cinebench score |
| Ryzen 9 7950X | 95°C (Limit) | 87°C (360mm) | -8 °C | +4.8% Cinebench score | |
| Ryzen 9 7950X3D | 84°C | 76°C (360mm) | -8 °C | +1.5% multi-core score | |
| Ryzen 9 9900X | 79°C | 71°C (360mm) | -8 °C | +2.0% multi-core score | |
| Ryzen 9 7900X | 94°C (Limit) | 85°C (360mm) | -9 °C | +4.0% multi-core score | |
| Core i5 | Core Ultra 5 245K | 73°C | 64°C (240mm) | -9 °C | +1.0% Cinebench score |
| Core i5-14600K | 84°C | 74°C (240mm) | -10 °C | +2.5% Cinebench score | |
| Core i5-13600K | 83°C | 73°C (240mm) | -10 °C | +2.2% Cinebench score | |
| Core i5-12600K | 76°C | 67°C (240mm) | -9 °C | +1.0% boost stability | |
| Core i5-14400 | 78°C (Stock) | 59°C (240mm) | -19 °C | 0.0% performance impact | |
| Core i7 | Core Ultra 7 265K | 85°C | 75°C (360mm) | -10 °C | +2.8% Cinebench score |
| Core i7-14700K | 96°C (Throttles) | 84°C (360mm) | -12 °C | +5.5% Cinebench score | |
| Core i7-13700K | 95°C (Throttles) | 83°C (360mm) | -12 °C | +5.0% Cinebench score | |
| Core i7-12700K | 84°C | 74°C (360mm) | -10 °C | +2.0% boost headroom | |
| Core i7-14700 | 88°C | 77°C (240mm) | -11 °C | +3.0% sustained boost | |
| Core i9 | Core Ultra 9 285K | 86°C | 76°C (360mm) | -10 °C | +3.0% Cinebench score |
| Core i9-14900K | 100°C (Throttles) | 89°C (360mm) | -11 °C | +7.5% Cinebench score | |
| Core i9-14900KS | 100°C (Severe) | 92°C (360/420) | -8 °C+ | +9.0% multi-core score | |
| Core i9-13900K | 100°C (Throttles) | 90°C (360mm) | -10 °C | +6.8% Cinebench score | |
| Core i9-12900K | 94°C | 82°C (360mm) | -12 °C | +4.5% Cinebench score |
This clearly states that lower power consuming chips like the Ryzen 5 basically have zero benefit from an AIO. However, it is worth noting that there is a significant impact with an AIO for overclocking chips that consume a lot of power (175W+) like the Intel Core Ultra chips, which can consume absurd amounts of power (200W+) under full load—and we clearly see an improvement in the Core i7-14700K. For most AMD Ryzen 9s and almost all i9s, you should consider an AIO, because they keep the temps generally lower, which prevents them from thermal throttling.
Now here comes the exception. When you add "workstation" to the equation, there is always a tax you have to pay. When you use server-grade Intel Xeon(s), Intel Core-X(s), AMD Threadrippers, or AMD Epyc-grade CPUs, that's when Water Cooling becomes more of a necessity. Now yes, you absolutely could use air cooling for high-end CPUs, but for those week-long continuous 24/7 workloads, Water Cooling (and some AIOs) becomes more attractive because they are less likely to throttle and sustain temps better under all-core workloads. This difference is not just limited to workstation CPU(s), but is just how "workstation" tasks work. In all fairness, gaming is still using your CPU for a sprint: you play for 3–5 hours, or 8+ hours if you're crazy or on a weekend where your CPU works. But a workstation load is like a marathon. The difference is, even in a marathon, the CPU is forced to run at 100% utilization for who-knows-how long. That's where Liquid Cooling shines.