Nvidia reportedly cuts program designed to keep gaming GPUs near MSRP pricing — end of OPP pricing-support scheme does not bode well for gamers

Asus ROG Astral 5090
(Image credit: Asus)

A new leak from tech YouTuber der8auer claims that Nvidia is ending a program dubbed "OPP", purportedly an incentive scheme used by Nvidia to ensure that at least some of its graphics cards were sold to consumers at MSRP by AIBs like Asus and others.

By way of historical background, the creator said in his YouTube video that Nvidia’s GPUs were previously being sold for way above the suggested retail price for the last few generations, and consumers were starting to notice and complain about it. So, the company instituted the OPP program, which, according to de8auer’s sources, is some sort of cash back for participating AIB partners. Although the techtuber does not know what OPP stands for, this is how it allegedly works: board partners sell some models at MSRP and then report that back to Nvidia, after which the GPU manufacturer would give them a rebate on the cost of the chip and the memory of that particular GPU they sold.

The Nvidia MSRP Lie - YouTube The Nvidia MSRP Lie - YouTube
Watch On

According to der8auer's leak, Nvidia apparently stopped this program a few days ago. The sources did not say the reason behind the move, although it is easy to surmise that it’s caused by the ongoing memory crisis. We've reached out to Nvidia for comment. There were rumors late last year that the company no longer bundled VRAM with the GPU chips that board partners ordered, but Nvidia refuted this at CES, telling HardwareLuxx [machine translated], “No changes on how Nvidia is handling the memory allocation.”

If the OPP program is indeed real, it seems that the continuous increase in memory pricing is making it difficult even for Team Green, with its record revenues fueled by the boom in AI infrastructure build-out, to sustain the GPU and memory subsidy that allows board partners to at least sell some models at MSRP.

RTX 5070 Ti is not end-of-life, but expect limited supply anyway

Aside from the end of the “OPP,” der8auer also touched on the recent RTX 5070 Ti news. Another YouTube channel, Hardware Unboxed, was told by an Asus representative that the RTX 5070 Ti and RTX 5060 Ti 16GB have been discontinued or marked as end-of-life (EOL). Asus refuted this a few hours later, saying that “Certain media may have received incomplete information from an Asus PR representative regarding these products,” and that “Asus has no plans to stop selling these models.”

However, industry sources reportedly told the tech tuber that Nvidia is prioritizing the RTX 5080 over the RTX 5070 Ti for its GB 203 supply. Both GPU models use the same chip, but since the GB 203 has a high yield rate, it’s been rumored that the GPU manufacturer wants to maximize its profits and prioritize the more expensive RTX 5080. Because of this, some say that the supply for the 5070 Ti will be heavily cut and that it’s going to be shifted towards the 5080.

Nevertheless, that does not mean we expect to see RTX 5080 prices to drop. In fact, the opposite has been true in recent days, with prices for different RTX 5080 models from various AIBs showing an increasing trend on Amazon. And with demand for both memory and AI chips remaining strong, we don’t expect gamers and their wallets to get a reprieve anytime soon.

Google Preferred Source

Follow 3DTested on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Jowi Morales
Contributing Writer
  • phxrider
    With decreased supply, AIBs are going to have to make up for the revenue loss with higher prices, or quit the business. Take your pick.

    And of course, in the face of RAM shortages, they're going to prioritize the version that sells for the most, i.e the 5080 over 5070 ti and 5060 ti, just like AMD is going to prioritize the 9070 XT over the 9070 and 9060 XT.

    This would be a really good time for Intel to get into the RAM business LOL. They already have fabs....
    Reply
  • SkyBill40
    SHOCKER.
    Reply
  • vanadiel007
    The way I look at it, why even set an MSRP. It's clearly a symbolic price that nobody cares to sell at anyways.
    Reply
  • bit_user
    The article said:
    If the OPP program is indeed real, it seems that the continuous increase in memory pricing is making it difficult even for Team Green, with its record revenues fueled by the boom in AI infrastructure build-out, to sustain the GPU and memory subsidy
    I think it's not that they can't afford it, but rather that there's just not enough GDDR7 to go around.

    So, letting prices naturally respond to the supply shortage will have the effect of reducing consumption. That reduced consumption will allow market dynamics to direct GDDR7 where it adds the most value. Sad to say, that will be in their higher-end SKUs, like the RTX 5080 and 5090. Also, the much-maligned 8 GB version of the RTX 5060 Ti.

    This article has a nice table that lays out the price vs. GDDR7 ratio, which should serve as a good predictor of where most of their GDDR7 supply will go:
    https://www.3dtested.com/pc-components/gpus/gigabyte-ceo-explains-nvidias-potential-gpu-supply-strategy-amid-crushing-memory-shortages-gross-revenue-per-gigabyte-of-gddr7-memory-could-decide-what-products-thrive
    Fortunately, their workstation and server GPUs that use GDDR7 all use the 24 Gb dies. So, they aren't actually competing with the gaming cards for supply! I suppose maybe at the Fab level, but not among the 16 Gb GDDR7 chips that are actually produced.
    Reply
  • bit_user
    vanadiel007 said:
    The way I look at it, why even set an MSRP. It's clearly a symbolic price that nobody cares to sell at anyways.
    Plenty of them even dropped below MSRP, during Black Friday sales. That's when I got a RTX 5070 for $480, which is $70 (12.7%) below MSRP!

    Now, I'm slightly kicking myself for not getting a Ti, but it was sort of an impulse buy and already more than I wanted to spend.
    Reply
  • bit_user
    phxrider said:
    This would be a really good time for Intel to get into the RAM business LOL. They already have fabs....
    Sorry, but that can't work. Some of the reasons:
    The manufacturing process for logic and DRAM are fundamentally different. It would require retooling and a huge R&D cycle for them to get a competitive DRAM node, and then those lines wouldn't be able to fab logic wafers. This is why TSMC doesn't make DRAM (or NAND) and companies like SK Hynix and Micron don't make logic dies.
    Requires lots of specialized IP they simply don't have. Another long & expensive R&D effort would be required.
    They recently sold off their NAND business & fabs, in order to focus more on core products. So, it would take them back in the opposite direction of trying to chase too many markets.
    Memory tends to be a very cyclical business. Investing now could leave them trying to launch products in the next big downturn.
    Fun fact: they were once in the DRAM business, but exited back in 1985! Https://timeline.intel.com/1985/farewell-to-dram
    Reply
  • SkyBill40
    bit_user said:
    Sorry, but that can't work. Some of the reasons:
    The manufacturing process for logic and DRAM are fundamentally different. It would require retooling and a huge R&D cycle for them to get a competitive DRAM node, and then those lines wouldn't be able to fab logic wafers. This is why TSMC doesn't make DRAM (or NAND) and companies like SK Hynix and Micron don't make logic dies.
    Requires lots of specialized IP they simply don't have. Another long & expensive R&D effort would be required.
    They recently sold off their NAND business & fabs, in order to focus more on core products. So, it would take them back in the opposite direction of trying to chase too many markets.
    Memory tends to be a very cyclical business. Investing now could leave them trying to launch products in the next big downturn.
    Fun fact: they were once in the DRAM business, but exited back in 1985! Https://timeline.intel.com/1985/farewell-to-dram
    It would take them at least a decade to get back into that market space and by then, what's the point? Even if they decided to make the jump, I doubt it would be anywhere as profitable as their chipmaking process.
    Reply
  • logainofhades
    SkyBill40 said:
    It would take them at least a decade to get back into that market space and by then, what's the point? Even if they decided to make the jump, I doubt it would be anywhere as profitable as their chipmaking process.

    What profit? While much improved, they have yet to right the ship fully.

    Https://www.tomshardware.com/pc-components/cpus/intel-shares-down-13-percent-as-company-only-manages-to-shrink-losses-in-latest-earnings-demand-to-outpace-2026-supply-usd300-million-deficit-comes-despite-more-than-usd20-billion-in-outside-investment-from-nvidia
    Reply
  • ezst036
    The customers are partially to blame here.

    Who thinks its still a secret that Nvidia stopped being a gaming company?

    Who has seen for some time now out of control pricing on Nvidia GPUs?

    Now. Knowing those two things?

    Who (gamers) is still buying Nvidia GPUs?????? Why would you ever do that? Why?!?!?
    Any gamer who has bought a recent Nvidia GPU is at least a part of the problem. The number is not zero. Nvidia's sales of gaming GPUs should be cratering with gamers, but it is not cratering.

    This points to a problem of bad consumers. Nvidia is arrogant and thinks it can do what it wants. Nvidia is actually correct, because its consumers, bad consumers, have told Nvidia to go do what Nvidia wants.

    Bad consumers.
    Reply
  • ezst036
    phxrider said:
    This would be a really good time for Intel to get into the RAM business LOL. They already have fabs....
    That is a really good point, though I don't think fabs can necessarily be turned on a dime.

    But yes. Intel would be wise to get into the RAM business right about now with its fabs.
    Reply