How can I bring this high-end PC build down to about $3,000?
Asked by
MellowOrbit42
I'm planning my first major PC build or upgrade since 2019. My current system has a Ryzen FX-8350, Radeon RX 6700 XT, and 32GB of DDR4, and it's starting to struggle with newer games. Gaming makes up roughly 80% of my usage, but I also run AI models for work, so I'd like to keep 64GB of memory. I'm aiming for a powerful system that can last around seven years without needing another major upgrade, but I'd prefer to avoid spending more than about $3,000. I already have two 4K monitors and all my other peripherals, and I'm not concerned about an expensive case or flashy aesthetics. I may sell my current PC to offset the cost. Which components would you change or remove to lower the price without sacrificing too much performance?
Answers
The biggest question is whether your AI workloads need NVIDIA’s CUDA ecosystem. If they do, sticking with an NVIDIA GPU makes sense even if the price-to-performance ratio is worse. If your AI software works well with AMD, a Radeon RX 9070 XT could save several hundred dollars while providing very strong gaming performance. It’s worth checking software compatibility before switching.
Reply to this answer
You can save a meaningful amount on cooling. A Thermalright Phantom Spirit-class air cooler performs surprisingly well, even with high-end processors, and costs far less than a premium liquid cooler. You could also look for a less expensive AM5 motherboard with strong VRMs instead of paying extra for features you won’t use. Keep the power supply reputable, but compare current sales because there are often cheaper units with similar quality.
Reply to this answer
Memory pricing has been unusually rough lately, so this is an expensive time to buy a large DDR5 kit. I’d compare several 64GB kits and avoid paying a premium for extreme speeds or tight timings. For gaming at 4K, the GPU matters much more than small differences in memory speed, and a slightly cheaper kit should perform nearly the same.
Reply to this answer
If 64GB is genuinely needed for your work, keep it. Cutting to 32GB is the easiest way to reduce the total, but it doesn’t make sense if your AI models regularly use more memory. A better compromise may be choosing a less expensive CPU or GPU model rather than sacrificing system RAM.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Related Discussions