- Nvidia has announced its RTX 5060 arrives on May 19
- The GPU has an MSRP of $299 and comes with 8GB of VRAM (which remains controversial)
- With AMD rumored to reveal its rival RX 9060 XT on May 21, the week after next could be a huge one for gamers seeking a more affordable GPU
Nvidia has announced that its RTX 5060 desktop graphics card will go on sale on May 19, as will the laptop version of this GPU.
VideoCardz caught wind of the date being announced on X (formerly Twitter). It was a launch day that had been previously rumored, with the price tag on the desktop GPU being $299, and the specs already revealed by Nvidia previously (when its sibling, the RTX 5060 Ti, was launched).
Starting May 19 at 9 a.m. Pacific Time, GeForce RTX 5060 graphics cards, desktops, and GeForce RTX 5060 Laptop GPUs will be available from our partners and retailers worldwide.#GeForceRTX50 pic.twitter.com/VLA9bZUWRWMay 6, 2025
The RTX 5060 Ti arrived in both 16GB and 8GB flavors, but the RTX 5060 will only debut with the latter loadout of video RAM (VRAM), and that has already proved controversial among PC gamers.
Again, like the RTX 5060 Ti, there will only be third-party versions of this graphics card, meaning Nvidia isn’t producing its own Founders Edition model.
To recap on the (already revealed) specs, the RTX 5060 has 3,840 CUDA Cores which is a 17% drop compared to the core count of the RTX 5060 Ti. Clock speeds and other specs are in the same ballpark, but as mentioned, there’s no 16GB spin on the vanilla RTX 5060.
Power usage is also lower with the RTX 5060 at 145W compared to 180W, which could be a consideration for folks who have a lesser power supply and not much room to maneuver when it comes to the overall wattage that their gaming PC can cope with.
Analysis: Lane drain and other VRAM complications
Not to bang on about it, but the obvious problem with the RTX 5060, as I discussed at length yesterday, is Nvidia’s choice to run with 8GB of VRAM. I won’t go over the same ground I’ve already covered (see yesterday’s piece for the full lowdown), but there are some further points to consider now the launch date is primed and ready to go.
The arguments in favor of accepting 8GB (as a pricing compromise) might include ‘it’s fine for 1080p gaming’ meaning those running at Full HD resolution should be more or less okay with this helping of VRAM. However, various third-party tests have illustrated that with some games and graphics settings, 8GB is a bottleneck at 1080p now, slowing down frame rates substantially.
Yes, Nvidia does have its bag-of-AI-tricks in the form of RTX Neural Texture Compression as revealed with the Blackwell generation, which helps VRAM go much further. But the trouble is, nothing’s happening quickly to bring support in with games for that tech. It could be a long way down the road before texture compression makes a meaningful impact (though it may indeed be a big piece of the puzzle for Nvidia eventually).
As things stand in the here and now, that 8GB loadout is looking plain shaky. The central tempting factor for the RTX 5060, then, is that it’ll be a much more affordable outlay – we hope. With that MSRP of $299, if third-party graphics cards are available at that price level, this GPU could be seen as a decent shot at an affordable gaming card with some limitations that you’ll just have to live with.
At least for those with a motherboard that supports PCIe 5.0, because if you have a PCIe 4.0 motherboard, you’ll experience additional slowdown in scenarios where the 8GB of memory isn’t enough, as evidenced with the RTX 5060 Ti. Why? Because when the on-board VRAM can’t cope, the graphics card taps the main system RAM, using that PCIe 4.0 interface to reach it – and it’s a lot slower than PCIe 5.0.
This gets a bit complicated because it’s worsened by the decision made by Nvidia that all RTX 5060 models, Ti versions included, only support half the available PCIe lanes, and so half the bandwidth. Now, PCIe 5.0 is fine because it’s so fast, halving its bandwidth still leaves plenty of room to pipe data through. However, this becomes a real stumbling block with the slower PCIe 4.0 standard, and even more of a problem with PCIe 3.0 should you have a motherboard running that even older take on the interface.
I should note that the RTX 5060 Ti 16GB is okay, because it has that much larger helping of VRAM, and so doesn’t need to run off to grab additional system RAM. And the RTX 5070 above it is fine, too – even if it only runs with 12GB of video RAM, which is arguably still light on the memory – because Nvidia didn’t hamstring that GPU by closing off half the PCIe lanes. The RTX 5070 has the full 16 lanes, rather than just eight as with the RTX 5060 models.
The long and short of it is that those without a new PCIe 5.0-toting motherboard will suffer the more outrageous slings and arrows bogging down performance with any RTX 5060 8GB (Ti, or non-Ti) graphics card. Pricing could also end up being higher than MSRP, too, as we’ve seen with the RTX 5060 Ti 8GB, which is currently selling for something like 10% more than its recommended price (that’s the cheapest model at the time of writing in the US).
So, we could end up looking at $330 (around £250 / AU$510) for the RTX 5060 in the US realistically, or $350 upwards for beefier models of this graphics card (and in line with that elsewhere). At which point, given the mentioned shakiness around performance, would-be buyers may have more pause for thought.
Traditionally, the xx60 series has been a big seller for Nvidia, but maybe that won’t be quite as clear cut this time around – especially given the irony that you need a cutting-edge motherboard to most effectively run your affordable-level Blackwell GPU.
Furthermore, we also need to see what competition AMD’s RX 9060 XT will be – although that may also have an 8GB spin (as well as a 16GB model), we don’t know that for sure yet, or how the full specs will pan out. A lot of eyes are watching this lower-mid-range battle, that’s for sure, with the 9060 XT expected to be revealed just after the RTX 5060 goes on sale.
Leave a comment