I'm putting together my first complete PC build. I'm mainly a casual gamer who plays indie games and occasional AAA titles with friends, but I also want the system for experimenting with local LLMs and agentic AI workflows. I'm a developer and will be pursuing a master's degree in AI, so I'm specifically interested in having local hardware rather than relying entirely on hosted AI services.
The current build is intended to be powerful and reasonably long-lived without going overboard, especially since memory prices are high. I'd appreciate advice on where I could get better value, whether 64 GB of RAM is appropriate to start with, and whether I could safely use a lower-wattage power supply. I'd also like suggestions for a less expensive or more compact case. I'm not completely new to PC hardware, but this will be my first time assembling an entire system myself. I'm hoping to avoid parts that will feel obsolete or require replacement within the next year, while keeping the build balanced for both gaming and local AI experimentation.
3 Answers
For your use case, 64 GB is a reasonable starting point rather than something to remove outright. Local models can consume a lot of system memory, although GPU VRAM will usually be the bigger limitation for running larger models. You can always add more RAM later if your workloads actually need it.
The CPU choice depends on your priority. A gaming-focused X3D chip would generally be better for frame rates, while a higher-core-count processor makes more sense for compiling, multitasking, and CPU-based workloads. Also, don’t count on any build being truly future-proof—focus on a current platform with a good upgrade path and the features you need now.
Double-check the prices before ordering. The listed 64 GB memory kit and 4 TB SSD appear dramatically overpriced, which could distort the entire budget. A 2x32 GB DDR5-6000 CL30 kit from a reputable brand should normally cost far less, and there are several good 4 TB PCIe 4.0 NVMe drives that offer similar real-world performance for much less.
Those savings could go toward a better GPU, additional storage, or simply stay in your budget. For gaming, the GPU will have the largest impact, while local AI performance depends heavily on GPU VRAM, so make sure the graphics card’s memory capacity matches the models you actually plan to run.
You probably don’t need a 1300 W power supply for a system with this class of CPU and a 5070 Ti. A high-quality 850 W unit should be enough for the current parts, and a good 1000 W model gives you extra headroom for future upgrades. I’d prioritize a modern ATX 3.x unit with a native 12V-2x6 cable and strong protections over an unnecessarily high wattage rating.
For the case, the North XL is spacious but larger than necessary unless you specifically want lots of room for expansion or a very large cooler. A regular Fractal North, Meshify 2, Lancool 216, or Phanteks XT Pro Ultra would generally provide plenty of airflow at a lower price. Just confirm that the case supports your GPU length and chosen cooler or radiator.
That’s helpful. I’m trying to balance gaming with development and AI experimentation, so I’m comfortable giving up a little gaming performance for more workstation capability. My main goal is avoiding a platform that’s already at the end of its useful upgrade path.