Here's a riddle with a real answer. Two gaming laptops both say "RTX 5060" on the box. One plays games noticeably better than the other. Neither is broken, and neither is lying about the chip. The difference is a number most listings don't show you: how much power the laptop lets its graphics chip use.
That number is the GPU's power limit, usually called TGP (total graphics power). On a desktop graphics card it's fixed. In a laptop, it's a dial the manufacturer sets, and where they set it can matter as much as the name on the sticker. So the short answer to the headline is this: a laptop RTX 5070 runs on a fraction of the desktop card's power, and it can sit anywhere in a wide range depending on which laptop it lives in.
Same name, very different budget
Start with the power gap. When Nvidia announced its first RTX 50 laptop GPUs in January 2025, Tom's Hardware reported a 50 to 100 W range for the RTX 5070 Laptop GPU, and referred to the desktop RTX 5070 as a 250 W card. Even at the top of its range, the laptop chip gets less than half that.
The chips aren't identical either. Nvidia's RTX 50 Series laptops page lists the RTX 5070 Laptop GPU with 4,608 CUDA cores, and Tom's listed it with 8GB of memory at launch. Same family name, its own spec sheet.
What TGP actually controls
Think of TGP as the graphics chip's allowance. Tom's Hardware put it plainly: TGP "is paramount for GPU performance," and manufacturers can configure the RTX 5090 Laptop GPU anywhere between 95 and 150 W "for every form factor and chassis."
Those ranges are wide across the whole line-up. Tom's launch report gives 80 to 150 W for the RTX 5080 Laptop GPU and 60 to 115 W for the 5070 Ti. PC Gamer's table adds the cheaper chips: 45 to 100 W for the RTX 5060 and 50 to 100 W for the RTX 5050.
Nvidia's own spec sheet doesn't print those power ranges, but it doesn't give a single boost clock either. It lists a range for each laptop GPU: 1,425 to 2,347 MHz for the RTX 5070, and 1,455 to 2,497 MHz for the RTX 5060. And the performance chart on that page carries its own power footnote, "150W GPU TGP."
Why does the dial move? Cooling. As PC Gamer's Dave James explains, the power limits on a laptop GPU "are entirely driven by the level of power and cooling they have at their disposal and that can change model-by-model." So two laptops can share a chip and still get very different allowances.
There's one more wrinkle: Dynamic Boost. When PCWorld tested the feature in 2020, it described Nvidia's aim as using the "unused" thermal budget in a laptop, taking power the CPU isn't using and giving it to the GPU while gaming. Those numbers come from older hardware, an RTX 2080 Super Max-Q laptop rather than an RTX 50 machine: Nvidia said the feature added 15 watts to the GPU's budget, PCWorld reported, and in its tests the GPU went from 80 to 95 watts, for gains of roughly 5 to 9 percent. The feature still turns up in today's figures: PC Gamer's RTX 5050 example below comes with "(plus dynamic boost power)" attached.
When the cheaper chip wins
Here's the twist that makes this matter. Because the power ranges overlap, a lower-tier chip with a generous limit can beat a higher-tier one on a tight leash. PC Gamer's example: "An RTX 5050 machine running at peak 100 W (plus dynamic boost power) will regularly best an RTX 5060 running at just 55 W in pretty much any game."
That's not a hypothetical cheap trick, either. In the same piece, PC Gamer points to real RTX 5060 laptops spread across that range: an MSI Cyborg at 55 W, an Asus V16 at 65 W, an HP Victus at 75 W and a Lenovo LOQ that the deal write-up describes as having a 100 W GPU (its key-specs line says 115 W). Same chip name, very different allowance. PC Gamer's advice was to skip the 55 W MSI and lean towards the Lenovo.
The catch is finding the number. PC Gamer says MSI is one of the few makers that lists peak power limits on its product pages, while the Asus V16's 65 W limit appeared in neither the product listing nor the retailer listing. Lenovo does publish its figures, but through what James calls an opaque parts site, and for HP he had to contact customer service. Reviews, the piece says, are often the only reliable way to find out.
How to check before you buy
The wattage hunt is annoying, but it's doable:
- Read the manufacturer's product page, not just the retailer listing. PC Gamer found MSI lists peak GPU power there; look for a TGP or wattage next to the GPU name.
- Compare like with like. Two laptops with the same GPU name and very different TGPs are not the same performance tier, whatever the price tags suggest.
- Search reviews of that exact model. PC Gamer says it always notes TGP in its gaming laptop reviews.
- Ask the manufacturer, then double-check. PC Gamer went to HP's customer service for a figure, and James wasn't entirely sure the answer he got was right. Treat what you're told as a lead until a spec sheet or review backs it up.
- Don't assume from the price. In PC Gamer's examples, the cheaper HP had a higher limit than the Asus.
The riddle from the top, solved: the GPU name tells you which chip is inside, and the wattage tells you how hard it's allowed to work. A laptop "RTX 5070" isn't a desktop RTX 5070, and it isn't even always the same as another laptop RTX 5070. Whenever you shop, put the watts next to the name. And if a listing won't tell you, a review of that exact model is the next place to look.
Reporting notes
