Quick Answer

A local AI and Stable Diffusion rig in Hermanus is sized by GPU VRAM. The RTX 4060 Ti 16 GB (~R10,500) runs SDXL and Flux.1 well; for serious throughput an RTX 4090 24 GB clears a 1024px SDXL image in under 4 seconds. Pair the GPU with a Ryzen 5 7600 and 32 GB DDR5. The GPU's VRAM, not CPU speed, sets what models you can run and how fast.

Why VRAM is the deciding spec

Stable Diffusion loads the model into GPU memory, so VRAM caps the workload:

  • 8 GB: SD 1.5 only
  • 12 GB: SDXL comfortable, Flux tight
  • 16 GB (RTX 4060 Ti 16 GB): SDXL + Flux + ComfyUI batches
  • 24 GB (RTX 4090): fastest output, handles video and large batches

For a minimum spec, target a 16 GB card; the RTX 4090 is the performance tier if your work demands speed.

The supporting build and Overberg delivery

  • CPU: Ryzen 5 7600 (~R4,200) - just feeds the GPU
  • RAM: 32 GB DDR5-6000 (~R2,000)
  • Storage: 1 TB NVMe (~R1,200) - checkpoints are 2-7 GB each
  • PSU: 650 W Gold for the 4060 Ti, 850 W+ for a 4090

Hermanus is on the Overberg coast, so Evetech deliveries take 3-4 business days; keep the PC ventilated against salt air and power-test before signing.

FAQ

What's the minimum GPU VRAM for SDXL?

12 GB runs SDXL, but 16 GB (RTX 4060 Ti 16 GB) is the practical minimum for comfortable SDXL and Flux work, avoiding the out-of-memory errors that 8 GB cards hit at 1024px and above.

Is an RTX 4090 worth it for Stable Diffusion?

For heavy or professional use, yes - its 24 GB VRAM clears a 1024px SDXL image in under 4 seconds and handles video models and large batches. For hobby use, the RTX 4060 Ti 16 GB is far better value.

Does Evetech deliver to Hermanus?

Yes, Evetech ships to Hermanus on the Overberg coast in 3-4 business days. Power-test the GPU and confirm it's detected before signing.

TIP

high-VRAM GPU with a clean 850 W+ 80+ Gold PSU if you choose a 4090 - Stable Diffusion holds the card at full power for sustained periods, so a quality supply keeps long batches stable.