Quick Answer
A Witbank Stable Diffusion build under R20,000 must prioritise VRAM. An RTX 4060 Ti 16GB (around R10,500) with a Ryzen 5 7600 and 32GB DDR5 renders SDXL images in roughly 8-15 seconds each and keeps the whole build under R20,000.
VRAM Is The Make-Or-Break Spec
For local AI image generation, the graphics card's video memory decides which models run. SD 1.5 fits in 8GB, but SDXL and FLUX want 12-16GB to avoid out-of-memory errors and slow CPU fallback. Under R20,000, the RTX 4060 Ti 16GB is the clear pick over an 8GB card - the extra VRAM is what unlocks SDXL and FLUX locally.
Building Around The GPU
Pair the 4060 Ti 16GB with a Ryzen 5 7600 (around R4,500), 32GB DDR5 (around R3,000) and a 1TB NVMe SSD (around R1,100) to hold checkpoints - each SDXL model is 6-7GB. A 650W 80+ Gold PSU covers the card. Install Forge or Automatic1111 first for simplicity, then ComfyUI for node-based workflows once you want fine control.
Witbank Notes
Evetech ships assembled AI rigs to Witbank and across Mpumalanga by courier. Inference runs offline, so after downloading models you generate without a connection. The warm inland climate suits a mesh case with two intake fans to keep the GPU under 75C during long generation queues.
FAQ
Can I run SDXL under R20,000?
Yes - the key is choosing the RTX 4060 Ti 16GB. Its 16GB VRAM runs SDXL and FLUX locally, which 8GB cards cannot do reliably.
What software should I install?
Start with Forge or Automatic1111 for ease, then ComfyUI for advanced node-based workflows. All run on this build.
How fast are renders on this build?
An SDXL image takes roughly 8-15 seconds on the RTX 4060 Ti 16GB, depending on resolution and steps - fast enough for everyday creative work.
000, choose the 16GB RTX 4060 Ti over any 8GB card - VRAM, not raw speed, is what lets your Witbank rig run SDXL and FLUX locally.