Loosened, Not Loose

I asked folks a while back what was the absolute most they’d pay for a video card.

Here’s an instance where the raw numbers are useful:

How Much

How Many

$150 Or Less


$175 Or Less


$200 Or Less


$250 Or Less


$300 Or Less


$350 Or Less


$400 Or Less


$500 Or Less




At first glance, this seems to show a signficantly different result than the results of our earlier poll, which indicated that most wouldn’t spend more $150 for a card, and practically no one would spend more than $250.

However, we didn’t ask the same question both times. The first time, we just asked “How much are you willing to pay?” This time, we asked “What is the absolute most you are willing to pay?”

We asked the question differently because we wanted to see what was the absolute maximum market for luxury video cards. Is there really a market for them, like Matrox and 3DLabs apparently think, or are they kidding themselves?

From the comments we got, it looks like most of the difference between the two results were people willing to pay $50-$100 more if they got the sun, the moon and the stars for it.

Indeed, the vast majority of those willing to pay $300 or more went out of their way to say they’d only do it if they got the sun, the moon and the stars for that kind of money.

A few defined what the sun, the moon and the stars meant to them, and if they were representative, the willingness of a very sizable chunk of those who said they’d lay out really big bucks is strictly hypothetical.

However, not all of the difference can be accounted for just by a difference in questioning. On the whole, most people did seem more willing to spend somewhat more than they did about eighteen months ago.

I think most of that is due to stretching a bit to reach the price point of the next level. For instance, somebody otherwise willing to spend $200 today finds himself inbetween the price points of a Ti4200 and a Ti4400, and some will be willing to stretch an extra $30-40 to get the 4400.

I think that indicates clever pricing on nVidia’s part. What I didn’t see was any revolution in shelling out big bucks. There were few “$400 for a video card? So?” type of answers.

In contrast, over 40% simply will not lay out more than $200 for a card, and most of those willing to go somewhat over that are audibly aching in the process.

Per the product that started this up in the first place, the Parhelia, outside of a couple vociferous Matrox defenders and a few more who had professional plans for it, most were pretty skeptical about the prospects of this card.

The problem with the notion of a truly luxury video card is the same for any truly luxury item. If you don’t get economies of scale, the item will cost disproportionately more than the improvement. The issue then becomes, can your product be better enough to persuade an audience to pay that disproportionate price?

Video cards are a uni- or at best bidimensional product. For most people, it either runs fast or it doesn’t. For some, it either puts up high-quality images, or it doesn’t. There’s really not much else to convince you.

In comparison, Rolls Royce owners could cite a lot of different reasons, including intangibles, most especially status, for owning one.

It’s not impossible to create what at least in the minds of their owners is a luxury computer brand: look at Apple. However, also look at Apple’s market share. Also keep in mind that a Mac owner has more reasons (or at least excuses) for wanting a Mac than can be expected from what is, after all, an unseen computer component.

The competition must also be kept in mind. nVidia is a company devoted to economy of scale. It spreads its R&D over millions and millions of GPUs. It spreads its fixed manufacturing costs over those millions, too. That gives them a pretty sizable cost advantage over anybody who has to spend much the same money, but spread them over a small fraction of products.

The two core elements of a video card are the GPU and memory. What is the GPU made from? Silicon. Tough to come up with better silicon. The design of the silicon is very important, of course, but it’s hard to see how you can design much better than nVidia (what’s far more feasible is filling niches nVidia’s not terribly interested in).

Memory is a commodity. Sure, you could get manufacturers to build faster memory than nVidia uses, but again, we have economies of scale. That faster memory will probably cost you a good deal more than it would have cost nVidia, simply because you’ll want much less of it.

None of these preclude the creation of a luxury card. The crunch question in all likelihood would be along the lines of “Will enough people would pay double to get a 30% better card?”

There’s reason to doubt that.

Thanks to all who contributed!


Be the first to comment

Leave a Reply