GPU MSRP is Not Meaningless

There are a lot of talking heads squawking nowadays about video card MSRP, otherwise known as “Manufacturer Suggested Retail Price.” You might also see this referred to in company slides as, “SEP,” or “Suggested Etail Pricing.” For this conversation, MSRP and SEP mean the exact same thing to us, the consumer.

If you are a computer hardware enthusiast, like me, you are surely aware of the retail/etail cost of video cards currently selling well over AMD’s or NVIDIA’s stated MSRPs. Does this suck? Absolutely, but some of the folks that talk about this subject do not seem to fully understand the way the GPU/video card supply channel works.

While watching one of my favorite TechTubers the other day he suggested that a GPU manufacturer was “cashing in” on inflated MSRPs. I believe that statement is wrong whether it pertains to (in alphabetical order) AMD, hopefully soon-to-be Intel, or NVIDIA. Let’s dig in a little deeper without dragging out the whiteboard but help everyone understand what MSRP is actually all about when stated by the likes of AMD or NVIDIA. We do however need just a paragraph of backstory before we directly address why the MSRP number is not meaningless.

When AMD and NVIDIA sell a GPU to a company such as ASUS, MSI, or Gigabyte (these companies are referred to as AIBs, short for Add In Board partners), it rarely sells the GPU only. Most of the time, what is sold to the card builder is called a “BOM kit.” BOM stands for Bill Of Materials. As “kit” would imply, it is not only a GPU in a BGA package. On a lower end card, one that is usually much easier to build, the BOM kit will usually consist of the GPU and the VRAM. There are several reasons for VRAM to be supplied with the GPU. One, VRAM for that specific GPU needs to meet a particular specification. Two, VRAM availability has been tremendously violative for the last couple of years and may be priced “too high” when needed. Three, due to fluctuating stocks, it may be difficult for the AIB to source VRAM at all. In all these cases, either AMD or NVDIA supply the VRAM to make sure the right VRAM is there to build the video card to specifications that either AMD or NVIDIA have directed to be part of the card’s design. BOM kits might also have more than just a GPU and VRAM as well. On higher end cards we might see AMD or NVIDIA supply power components as well as the less efficient, higher-powered cards get a bit pickier about what components play nicely in delivering a card without issues. VRM, capacitors, and other power components might also be hard to source at the time of need as well, for the same reasons stated above for VRAM. AMD and NVIDIA also have much more buying power and better product foresight than the AIBs and of course AMD and NVIDIA know exactly how many GPUs are going to be sold, so therefore they know how much they need of what, and exactly when. This allows AMD and NVIDIA to purchase VRAM and power components in bulk well ahead of release of the GPUs so it can be sure that the needed components will be there to build the card when it is time to launch into retail.

With that out of the way, let’s talk about MSRP from AMD and NVIDIA. While we are not paying those company’s MSRP at retail, those MSRP numbers from AMD and NVIDIA do mean something. Let’s use the Radeon 6500XT as our example since it has been getting the most attention from the press in the last week. AMD’s MSRP on the 6500XT is $199USD. What that means is, when AMD sells its 6500XT BOM kit to an AIB, the profit margins can support it being sold at retail to consumers for $199USD.

For this discussion, let’s say you see a 6500XT for sale at Newegg for $1000 and you see a 6500XT for sale at Microcenter for $199 on the same day. When both those cards sell, AMD’s profit margin, and even its overall revenue on that BOM kit does not change. AMD and NVIDIA, at least in the USA/Canada and EU, cannot legally dictate prices that products are sold at to consumers. Those dollars over MSRP go into the AIB’s and retailer’s pockets.

Traditionally the profit margin for AIBs and retailers is much lower than those of AMD and NVIDIA on a GPU/video card. However, we also must keep in mind that AMD and NVIDIA are the ones that take the overall risk on the research and development that goes into a GPU. What you may not know is that AMD, NVIDIA, and Intel may offer avenues of protection to its AIBs on retail products when or if market factors contribute to slow moving stock and/or falling prices. If Product X is left stagnant on the shelf due to pricing being too high in retail, those AIBs are often offered different ways to soften the blow since traditionally their margins are much lower. This in itself is a whole other discussion so we will not dive into that here. So again, we see the companies like AMD, Intel, and NVIDIA shouldering most of the risk if not all to a certain extent. There are very real risks to the likes of AMD, Intel, and NVIDIA when you look at the entire channel picture.

Another argument that you might bring up as to retail shelf pricing is companies like Apple, Microsoft, and Sony that enforce MAP (Minimum Advertised Price). Again, this would be another long conversation, but long story short, those companies are delivering a finished product to market where a GPU BOM kit is not when it leaves the hands of AMD or NVIDIA. That is more of a discussion about lower than “MSRP” pricing which of course is not the current focus.

AIBs and retailers have recently made huge profit margins when you look at the channel market historically. It is worth adding that in our “new” GPU economy, AIBs have also had rising costs associated with shipping and manufacturing and are having to deal with those as well. When we see MSRP prices come from AIBs that are hugely inflated over AMD and NVIDIA MSRPs, that is another discussion in its own as well.

The next time someone tells you that AMD, Intel, or NVIDIA are “cashing in” on pricing in the consumer market, that statement is somewhat misleading. There is certainly a lot more depth to the MSRP subject that is not covered here, but I wanted give my thoughts on the core of the GPU MSRP issues and hopefully help you understand it more thoroughly.

Another conversation, which is even more convoluted is what does it mean when a company does not list an MSRP/SEP on a product like the GeForce RTX 2060 12GB? It makes me wonder if we will see an MSRP on the 3080 12GB. I am personally still trying to digest all the ramifications that are subject to that and how it effects us, the gamers and enthusiasts.

I will close with the immortal words of Jimmy McMillan, that I think ring true with all of us. “The rent is too damn high!”

Feel free to cut and paste this post in its entirety with a simple attribution link as with any other post you might read here.