The RTX 5080 is reportedly Nvidia’s next-gen high-end GPU. We’ve rounded up everything we know about the card, including its price, specs, and performance.
While the RTX 5080 isn’t expected to be as revolutionary as the RTX 5090, the RTX 5080 is still expected to be a powerful card. PC builders looking for a cheaper but feature-packed option will be eager to see what the card can bring to the table.
As with the RTX 5090, there are few concrete details about the RTX 5080. Most of the information available comes from leaks from industry insiders. This means that until Nvidia officially unveils its next-generation cards, all of this information should be taken with a grain of salt.
Rumors about the release window
Recent reports from Taiwanese newspaper United Daily News indicate that the first RTX 50 series card could launch as early as Q4 2024.
The RTX 5090 is expected to be the first to be revealed, but a leak from reliable leaker Moore’s Law is Dead suggests that the RTX 5080 is also slated for that same release window.
A later report from Moore’s Law is Dead, however, indicated that this has now changed. In a more recent video, Moore’s Law is Dead stated that the RTX 5090 would be the only RTX 50-series card launching in 2024, with the others following in early 2025.
This information was disputed by Kopite7Kimi, another leaker. Kopite’s information suggested that in fact, the RTX 5080 would be the first to be released in Q4 2024, followed shortly by the RTX 5090. This would be a change in Nvidia’s usual release pattern if the information is accurate.
Price speculation
There is no confirmed data on the expected price of the RTX 5080. The RTX 5090 is expected to cost around $1,699 at launch, so we can use the price distribution of previous Nvidia GPUs to estimate that the RTX 5080 will cost around $1,000 at launch. This is in line with the current price of the RTX 4080 Super, which launches in early 2024.
Specs leaked
Nvidia officially announced the Blackwell architecture at GTC 2024. While this announcement primarily concerned its AI-focused chips, some of the information available remains relevant for commercial gaming GPUs.
Diamond Foundry
The RTX 5080 is widely expected to use the GB203 chip, a step down from the flagship GB202 chip that is expected to be used in the RTX 5090.
According to regular leaker Kopite7Kimi, the GB203 will be “half the size of the GB202.” This would be in line with the performance difference seen in the Ada Lovelace lineup between the RTX 4090 and the RTX 4080.
For memory, it is increasingly believed that the RTX 5080 will use GDDR7 memory, improving speed and power efficiency. The debate over the memory interface, however, remains open. Another recent leak from Kopite7Kimi suggests that the RTX 5090 will use the 512-bit memory bus.
However, previous rumors indicated that Nvidia plans to reuse the existing 384-bit memory bus for some of its new products. This could mean that the 512-bit bus is reserved for the flagship RTX 5090 card, and that the RTX 5080 and earlier versions will use the older 384-bit memory bus technology.
As with most other information currently available, until Nvidia makes an official announcement, this should be taken with the appropriate level of skepticism.
Twitter/X User Everest I spotted an update to Seasonic’s PSU wattage calculator, which now includes data for the RTX 50 series. While the data may not be completely accurate, it is roughly in line with expected models.
According to Seasonic’s data, the RTX 5080 will have a TDP of 350W. This is a reasonable step up from the RTX 4080, which has a TDP of 320W.
Data provided by Benchlife early September 2024 seemed to confirm the 350 W figure, although Kopite7Kimi has offered a slightly higher figure of 400W for the RTX 5080. However, it is unclear whether this figure refers to TGP (Total Graphical Power) or TDP (Thermal Design Power).
Speculations on features
Once again, we have to thank leaker Kopite7Kimi for the information that the RTX 5080 will support Display Port 2.1, as well as HDMI 2.1. This is an upgrade from the Ada Lovelace cards, which did not support DisplayPort 2.1.
PCWorld
The power connector has sparked an unexpected debate. As more and more reports emerge of the 12VHPWR connectors used on RTX 4090 and 40 Super cards melting and causing significant damage, consumers are hoping that Nvidia will consider a different type of connector for its next-gen cards.
According to Moore’s Law, Nvidia is indeed planning to move to an entirely new connector type for the RTX 50 series.
The leak also suggests that it will be an entirely new type of 16-pin connector, but further details are not available and other sources have yet to confirm this rumor.
If Nvidia persists with the 12VHPWR connector, power consumption will be capped at 600W. However, rumors indicate that Team Green will be pushing the efficiency of the new cards, suggesting that real-world consumption figures will be significantly lower than this.
Speculations on performance
From the current leaks, it seems that the biggest performance improvement is due to the larger bus width. While this may not seem like a big possibility at first glance, the size of the data bus can have a direct impact on the resolution that can be used in games.
Most current 40-series GPUs use 192-bit and 256-bit data buses, which are capable of easily rendering 1440p, but may struggle in 4K.
Even though the RTX 5080 is limited to 384-bit, this would allow it to easily output 4K. A 512-bit bus would allow the card to handle up to 8K.
As the market moves towards 4K as standard, enabling a high-end GPU to handle higher resolutions is a logical evolution.
Should we wait for the RTX 5080?
Cards with the 80 suffix are generally a good option for PC gaming fans who want high-end performance without shelling out a lot of money for a high-end flagship card.
Additionally, those looking for a bargain might be able to snag one in late 2024, when retailers will discount older stock of RTX 4090 or RTX 4080 Super cards to make way for the shiny new GPUs.
