PC Power Supply Calculator

To calculate your required PSU wattage: (1) Sum the TDP of your CPU and TBP of your GPU. (2) Add ~50W for RAM, storage, fans, and motherboard overhead. (3) Apply a transient multiplier — typically 1.35× for mid-range GPUs or 1.57× for halo-tier GPUs like RTX 5090. (4) Add a 20–35% safety buffer depending on ATX version. PSUCheck automates all of these steps using real Cybenetics-sourced data.

The RTX 5090 (575W TBP) can spike to over 900W transiently. NVIDIA officially requires a 1000W PSU minimum. With a high-end CPU like the i9-14900K (253W sustained), the total transient peak can exceed 1100W. We recommend a 1200W ATX 3.1 PSU with a native 12V-2x6 port — not an 8-pin adapter cable.

The RTX 4090 (450W TBP) spikes to ~707W transiently (1.57× multiplier). With a 125W+ CPU, total peak exceeds 830W+. NVIDIA recommends 850W minimum. For a high-end pairing like i9-14900K + RTX 4090, use a 1000W ATX 3.1 PSU to ensure safe margins at all transient loads.

850W is sufficient for systems with RTX 5080 (360W TBP → 540W transient) paired with a mid-to-high CPU. However, 850W is borderline for RTX 4090 + high-TDP Intel CPUs, and insufficient for RTX 5090 configurations. Use PSUCheck to calculate your exact transient peak before buying a PSU.

A GPU transient power spike is a sub-millisecond burst of power demand — typically 135–200% of the GPU's rated TBP — that occurs during workload transitions (e.g., starting a game, loading a shader). Oscilloscope measurements from TechPowerUp and Igor's Lab show the RTX 5090 reaching 900W+ in these spikes. ATX 3.1 PSUs are designed to tolerate 200% excursions for 100 microseconds.

Not significantly. PSUs only draw the power your components need, regardless of rated capacity. However, PSUs run at peak efficiency at 40–60% load. A 1600W PSU powering a 400W system runs at only 25% load — a less-optimal efficiency point. Oversizing by more than 3× can reduce efficiency by 2–5%, costing a few extra dollars per year in electricity.

According to the Intel ATX 3.1 Design Guide (2024), compliant PSUs must withstand power excursions up to 200% of rated capacity for durations under 100 microseconds, and hold output voltages within ±5% during those excursions. This is the standard that allows 850W ATX 3.1 PSUs to safely power RTX 5080 systems with 540W transient peaks.

For gaming PCs running 4–6 hours/day, Gold (87–90% efficiency) vs Platinum (89–92%) saves roughly $8–15/year at US electricity rates. The $40–80 price premium for Platinum takes 3–7 years to break even. For 24/7 systems or mining rigs, Platinum or Titanium is clearly worth it. For casual gamers, Gold provides excellent value.

Never daisy-chain PCIe 8-pin connectors for high-power GPUs (above 225W). Each connector should run as a dedicated cable directly from the PSU. Daisy-chaining multiplies resistance and can cause connector melting or fire at the transient current levels of modern GPUs. This is especially critical for RTX 4080, 4090, 5080, and 5090.

Cybenetics is an independent PSU testing laboratory that measures efficiency across more load points (10%, 20%, 40%, 60%, 80%, 100%, and 110%) than the basic 80 PLUS test (only 20%, 50%, 100%). Their ETA (efficiency) and LAMBDA (noise) ratings are considered more rigorous and real-world accurate. ETA-A++ corresponds roughly to Titanium efficiency tier.

A 500W gaming PC running 4 hours/day at $0.13/kWh (US average) costs roughly $95/year in electricity. A 750W system costs ~$143/year. High-end builds with RTX 5090 running 8 hours/day can approach $400–500/year. Use PSUCheck's 5-Year TCO Calculator to compare electricity costs by PSU efficiency tier for your specific usage profile.