I’ve been adding a lot of AIB graphics cards to GD lately. Hundreds, in fact. And through the foggy haze of data entry, one thing stuck out to me - who the heck is buying the premium tier graphics cards in each GPU family?

First things first, there’s the issue with just how many variants of any given graphics card there is. MSI is a particularly egregious offender in this regard, with no fewer than 13 different GeForce RTX 2070 graphics cards, for example, and a dozen each of the RTX 2080 Ti and RTX 2080. In total, MSI has 45 RTX 20 graphics cards split across a range of just four distinct GPUs. That’s a lot for any customer to wrap their head around, but for MSI the motivations are pretty clear.

Shelf space is everything, even in the digital world. The more distinct graphics cards MSI pumps out, the more slots it occupies in Newegg and Amazon searches; the more reviews it can earn, and the more choice it can offer potential customers. What we end up with is 13 different MSI RTX 2070’s, each a little different from the last. Some you’ll have to squint to see the differences, while others will be bolting on water blocks and enough RGB lighting to make Santa’s grotto look like a dingy basement.

Crucially, however, it means MSI can cover each and every price point. Not to pick on MSI specifically here but, as an example, the cheapest MSI RTX 2080 according to Newegg is $699 for the MSI RTX 2080 Ventus, rising to $1799 for the MSI RTX 2080 Sea Hawk EK X. Even discounting this outlier, the second most expensive is the MSI RTX 2080 Sea Hawk X at $945. That's $250 more expensive than another model of the same GPU.

MSI can cover the entire spread of RTX 2080 price points with this tactic, tempting you ever higher, and higher, right on up to the heavens. The smaller the pricing gap between each tier, the more inclined people will be to take the next jump up. It’s a proven trick that absolutely works; we all want the best we can get.

What this boils down to then, is that across pretty much every range of graphics card you can name, whether that’s AMD Radeon or Nvidia GeForce, there are AIB partner variants that spill over the next tier of pricing as they seek to fill all pricing gaps.

You'll find custom Radeon RX 470's, for example, that costs more than a Radeon RX 480. These can’t just be manufactured to sit on store shelves and gather dust, their only purpose to make the other variants appear better value. Or can they?

Coming back to Nvidia RTX for a moment, let’s break down roughly what it all means in terms of performance and pricing when you decide to opt for an enhanced version of a graphics card.

Here we’ve got a couple of charts. The first chart compares the cheapest and most expensive prices for each of the four GeForce RTX 20 GPUs.

As you can see, there’s plenty of overlap between the most expensive graphics card in a series and cheapest graphics card in the next series up.

In this next chart, we see the performance of the RTX 2080 Ti, RTX 2080, and RTX 2070, relative to the RTX 2060 Founders Edition. There are a lot of variables at play here so this obviously isn’t an exact science, but this provides some indication of the actual real-world performance difference.

What this tells us is that users buying high-end third-party graphics cards are spending exponentially more and yet getting comparatively little performance gain in return. In the case of the RTX 2080 Ti, the Zotac Gaming GeForce RTX 2080 Ti Triple Fan 11GB actually ships with slower clock speeds than Nvidia’s Founders Edition, with a $365 surcharge.

Of course, this is all taking into account the two extremes - cheapest and dearest. There are plenty that sit somewhere in the middle and so can better stake a claim to value for money.

What all of this fails to take into account are the numerous other benefits and goodies that AIB partners can bolt onto graphics cards. This includes liquid cooling systems, additional fans, RGB lighting, overclocking facilities, monitoring tools, snazzy shrouds, better grade heat pipes, etc, etc. These are all things that can contribute to a faster graphics card, and can also potentially enable better overclocking. But, at what stage is all this overclocking pointless when you could just buy a graphics card from the next tier up, and run it at stock clocks with zero fuss?

I'm interested to know what your thoughts on this are though, and whether you think it's absolutely worth it pay up the extra cash for premium models, or if you think a standard graphics card will do the job just fine.

Get voting in the polls and let us know why below!