How Much Power Supply Do You Actually Need for an RTX 50 Build?

Every time a new GeForce generation lands, the same question floods my inbox: "Is my old power supply good enough?" With the RTX 50 series, that question matters more than it has in years. These cards draw serious power, they spike hard, and they all rely on a single 12V-2x6 connector that has zero tolerance for sloppy installation. Get the PSU right and you'll never think about it again. Get it wrong and you're staring at random crashes — or worse, a melted connector. In this guide I'll walk you through exactly how much wattage each RTX 50 card actually needs, why ATX 3.1 has become the standard worth holding out for, and which units I'd happily drop into a build today.

The Short Answer: Wattage by GPU

Let's not bury the lead. If you just want a number to plug into your shopping basket, here's where the RTX 50 lineup sits. NVIDIA publishes minimum recommendations, and in my experience those figures are sensible starting points rather than aggressive overestimates — but I'd nearly always add a little headroom on top.

The headline jump is the RTX 5090. With a 575W total graphics power figure, NVIDIA recommends a 1,000W power supply as the floor. That's a meaningful step beyond what most people had in their machines a couple of years ago. The RTX 5080, by contrast, is far more reasonable at a 360W TGP and an 850W recommendation — comfortably within reach of plenty of existing builds, provided the connector situation is sorted (more on that shortly).

Step down to the RTX 5070 and a 700W unit covers your GPU plus the CPU, motherboard, drives and peripherals. The RTX 5060 and RTX 5050 are the modest members of the family — a quality 650–750W ATX 3.x unit handles them with ease. The pattern here is clear: this is a generation where buying slightly more PSU than the bare minimum is the smart, stress-free move.

Why Headroom Matters More Than Ever

Here's the bit that trips up even experienced builders. The wattage on the box is the sustained rating, but modern GPUs don't draw power in a smooth, polite line. They spike. The RTX 5090 in particular can pull around 400–550W during gaming, with brief excursions higher than that. In AI and rendering workloads it often sits near its maximum draw for extended periods, which is a very different stress profile from a quick gaming spike.

This is exactly why I tell people to leave breathing room. The rule of thumb I've used for years — and which still holds for RTX 50 — is to aim for roughly 20–30% of wattage headroom above your peak power requirement. That margin keeps the system stable, lets the PSU run cooler and quieter, and extends its lifespan because it isn't constantly working flat out.

Notice the gap between typical draw and the recommended PSU — that space is your transient buffer.

The danger scenario is the one that catches people out at 2am: your machine runs perfectly stable for hours, then crashes the instant the GPU spikes hard or the CPU and GPU boost simultaneously. An underspecced or older PSU can momentarily fail to deliver under those combined surges, and the result is an abrupt shutdown that looks like a software fault but absolutely isn't. I've chased that ghost in other people's builds more times than I'd like to admit, and the fix is almost always a beefier, more modern power supply.

Transient spikes are invisible in everyday use until the moment everything aligns — then a weak PSU buckles.

ATX 3.1: The Standard Worth Holding Out For

This is the part of the conversation that's genuinely changed how I recommend power supplies. ATX 3.1 isn't just a marketing sticker — it's a meaningfully better fit for cards that spike the way RTX 50 GPUs do. These units are explicitly designed to handle GPU power excursions, whereas older supplies simply weren't engineered with that behaviour in mind.

The numbers behind it are worth knowing. The ATX 3.0 specification already required a PSU to handle power excursions of up to 200% of its rated power for short durations, and to cope with 100-microsecond power draws at three times its maximum sustained output. In plain English: a compliant 1,000W unit is built to ride out brief surges that would knock an older design flat. ATX 3.1 refines this further, with a hold-up time of 12ms at full load (relaxed from the older 17ms requirement), and a recommendation of 17ms at 80% load.

The Connector Story: 12V-2x6 vs 12VHPWR

If you remember the RTX 4090 melting-connector saga, you're already braced for this section. The good news is that the RTX 50 series moves things forward. These cards use a single 12V-2x6 socket rather than the older 12VHPWR, and crucially they need just one such cable — not the multiple connections that were speculated about early on.

So what actually changed? Physically the two connectors are nearly identical and remain interchangeable, but the 12V-2x6 revision extends the power pins by 0.25mm. That tiny tweak improves connection reliability and reduces operating temperatures. Just as importantly, the sense pins are shorter, which means if the connection is compromised the GPU can power down immediately rather than continuing to draw current through a poorly seated plug.

There's a clever safety layer in the ATX 3.1 spec worth understanding. The four sense pins (SENSE0 and SENSE1) communicate the PSU's maximum power capability to the GPU. If those pins aren't properly plugged in, power simply won't flow to the GPU at all. It's a deliberate "won't work unless it's right" mechanism designed to force proper seating and prevent the overheating that caused so much grief last generation. I genuinely appreciate this approach — it turns a silent, dangerous fault into an obvious "it doesn't turn on" problem that anyone can diagnose.

The 12V-2x6 connector looks almost identical to its predecessor, but the extended power pins and shorter sense pins are the difference that matters.

Native cable or adapter?

Whenever possible, use the native 12V-2x6 cable that ships with an ATX 3.1 power supply rather than an adapter. A native cable provides a single, clean, direct path for power delivery, with no extra junctions to introduce resistance or seating problems. Adapters can work, but they add complexity exactly where you least want it. My standing advice is to pair an RTX 50 card with a modern ATX 3.0 — or preferably ATX 3.1 — PSU that has the native 16-pin connector built in.

Efficiency Ratings: Don't Skimp Here

Wattage gets all the attention, but the efficiency rating quietly determines how pleasant your PSU is to live with. The 80 PLUS scheme runs through Bronze, Gold, Platinum and Titanium tiers, and for an RTX 50 build I'd advise buying a Gold or Platinum certified unit at minimum.

Why does it matter beyond the energy bill? Titanium- or Platinum-rated supplies waste less power, which means they run cooler and quieter and — importantly for these spiky GPUs — deliver more stable voltage under heavy GPU load. When a 5090 is hammering the rails, that stability is exactly what keeps your system rock solid rather than flirting with instability.

My honest take: for anything 5080 and above, the small price premium for Platinum is money well spent. For a 5070 or below, a good Gold unit is perfectly sufficient and I wouldn't feel you were cutting a dangerous corner.

My Recommended PSUs for RTX 50 Builds

Enough theory — let's talk specific units. These are supplies that pair sensibly with RTX 50 cards based on their ratings, certifications and connector support. I've grouped them by where they slot into the lineup.

For the RTX 5090 flagship

See Corsair HX1500i on Amazon UK
£355.83price at 30 Jun, may change

For the RTX 5080 and below

See MSI MAG A1000GL on Amazon UK
£179.28price at 30 Jun, may change

Cross-generation all-rounders (RTX 40 / 50 compatible)

See Seasonic Vertex GX-1200 on Amazon UK
Check price & availability on Amazon

A native 12V-2x6 cable and an ATX 3.1 rating are the two boxes I most want ticked when matching a PSU to an RTX 50 GPU.

ATX 3.1 Native vs Older PSU + Adapter

One of the most common dilemmas I hear is: "I've got a perfectly good older unit — can I just use an adapter and save my money?" Here's how the two approaches really stack up.

Which PSU Is Right for You?

Power supplies aren't a one-size-fits-all purchase, so here's how I'd match a unit to the kind of builder you are.

Quick Reference: PSU Tiers at a Glance

To make the shopping decision painless, here's a tidy ladder from the modest end of the range up to the flagship. Treat these as the wattage targets I'd actually buy, with sensible headroom baked in.

How I Rate the RTX 50 PSU Situation

Stepping back, how do I feel about the state of power supplies for this generation? Honestly, fairly positive. The connector revision addresses last generation's biggest pain point head-on, the transient-handling requirements are mature, and there's a healthy spread of suitable units across price points. The only real friction is the higher wattage demand at the very top of the stack — but that's the price of a 575W flagship.

The score reflects a generation where doing it properly is straightforward, the safety mechanisms are genuinely improved, and the main cost is simply buying a bigger, better PSU than you might have a few years ago. That's a fair trade in my book.

Frequently Asked Questions

Pick the right wattage, insist on a native 12V-2x6 cable and an ATX 3.x rating, and your RTX 50 build will just work — quietly and reliably.