RUNLOCALAIv38
->Will it run?Best GPUCompareTroubleshootStartLearnPulseModelsHardwareToolsBench
Run check
RUNLOCALAI

Independently operated catalog for local-AI hardware and software. Hand-written verdicts. Source-cited claims. Reproducible commands when we have them.

OP·Eruo Fredoline
DIR
  • Models
  • Hardware
  • Tools
  • Benchmarks
TOOLS
  • Will it run?
  • Compare hardware
  • Cost vs cloud
  • Choose my GPU
  • Prompting kits
  • Quick answers
REF
  • All buyer guides
  • Learn local AI
  • Methodology
  • Glossary
  • Errors KB
  • Trust
EDITOR
  • About
  • Author
  • How we make money
  • Editorial policy
  • Contact
LEGAL
  • Privacy
  • Terms
  • Sitemap
MAIL · MONTHLY DIGEST
Get monthly local AI changes
Monthly recap. No spam.
DISCLOSURE

Some links on this site are affiliate links (Amazon Associates and other first-class retailers). When you buy through them, we earn a small commission at no extra cost to you. Affiliate links do not influence our verdicts — there are cards we rate highly that we don't have affiliate relationships with, and cards that sell well that we refuse to recommend. Read more →

© 2026 runlocalai.coIndependently operated
RUNLOCALAI · v38
  1. >
  2. Home
  3. /Learn
  4. /Courses
  5. /Hardware Planning for Local AI
  6. /Ch. 10
Hardware Planning for Local AI

10. Motherboard and PSU

Chapter 10 of 20 · 15 min
KEY INSIGHT

A single high-end GPU typically needs a quality 850W PSU, while dual-GPU configurations require 1400W+—never cheap out on power delivery. ```bash # Check current system power draw when idle nvidia-smi -q -d POWER | grep "Power Draw" # Monitor power under inference load python3 -c " import torch import time # Sustained inference loads the GPU model = torch.nn.Linear(4096, 4096).cuda() data = torch.randn(512, 4096).cuda() for i in range(1000): result = model(data) torch.cuda.synchronize() if i % 100 == 0: print(f'Iteration {i}') " ```

GPU and CPU selection means nothing without proper supporting components. Motherboards and power supplies require careful consideration for stable AI workloads.

Motherboard Requirements

Essential motherboard features for AI workstations:

Feature Minimum Recommended Notes
PCIe slots PCIe 4.0 x16 PCIe 4.0 x16 GPU bandwidth
PCIe lanes (CPU) 16 20+ Resolved by CPU model
RAM slots 4 4 Upgrade path
RAM capacity 64GB 128GB Model loading
M.2 slots 2 3+ Fast model storage
USB 3.2 Gen 2 4.0 Peripherals

CPU and Platform Considerations

PCIe lane allocation varies by CPU:

CPU Total PCIe Lanes GPU Usage Remaining
Ryzen 5 7600X 24 16 8 (NVMe, USB)
Ryzen 9 7950X 24 16 8
Core i9-14900K 20 16 4
Threadripper 3960X 64 64 64 remaining

Single-GPU users have flexibility. Multi-GPU requires more lanes.

PSU Calculations

Calculate total system draw:

Component Typical Load
RTX 4090 450W peak
RTX 4080 320W peak
RTX 3090 350W peak
CPU (AM5) 125W
CPU (LGA1700) 125-250W
RAM (4 sticks) 20W
Storage, fans 30W
Motherboard 50W

Single RTX 4090 build: 450 + 150 + 20 + 50 = 670W minimum Dual RTX 4090 build: 920 + 150 + 20 + 50 = 1140W minimum

Recommendation: 1000W for single high-end GPU, 1600W+ for dual configuration.

Quality Considerations

PSU quality matters more than wattage:

  • 80+ Gold minimum for efficiency
  • 80+ Platinum for dual-GPU systems
  • Brand matters: Seasonic, Corsair HX, EVGA SuperNOVA are reliable
  • Avoid cheap units—GPU load transients can exceed average draw

Case and Cooling

High-end GPUs require:

  • Minimum 3-slot clearance for RTX 3090/4090
  • 120mm+ intake fans
  • Positive air pressure (more intake than exhaust)
  • ambient temperature under 30°C for GPU longevity
EXERCISE

Calculate the total power budget for a system with Ryzen 9 7950X, RTX 4080, 64GB RAM, 2 NVMe drives, and base system. Recommend a PSU wattage with 20% headroom.

← Chapter 9
System RAM Benefits
Chapter 11 →
External GPU Enclosures