Will nVidia have discrete desktop graphics to itself within 5 years?

For years, nVidia has enjoyed a huge lead in workstation graphics. No secret. Quadro brings in serious profit. The connection from workstation to desktop was tenuous – as both had very different brands, markets, drivers and sales strategies. But there's one thing that Quadro gives nVidia, which could be crucial in 5 years' time, and that's R&D. KitGuru ponders the challenges being faced by AMD in making future graphic cards: Namely, how do you fund development?

Big secret? The market for graphic cards is shrinking.

Shocked?  Nope?  Us neither. But what will be the fallout from the climate change that's shrinking the GPU ice caps?

First, let us say that competition is a good thing. Without AMD pushing Athlon into the market, Intel may never have been spurred into the Core series. Similarly, when it comes to graphics, unless there is a steady stream of high-end Radeons around – where is the pressure on nVidia to deliver innovation on a regular basis?

On December 4th 2009, the world's biggest processor company said that it would be attacking graphics with its own, unique technology. There was a strong indication that Larrabee-based products would be shipping from Intel's FABs by the end of 2010.

By May 2010, after a series of disappointing internal lab test results, Intel announced that it no longer planned to make a discrete graphics card of its own and the research/learning was absorbed into other various projects – chiefly as a ‘supercomputer on your desktop' concept – which will be productised as the Xeon Phi add-in processing card, with pricing rumoured to be around $400. While not on the shelves yet, support from parallel processing specialists like Rogue Wave is being announced – so expect Xeon Phi to make some splashes in 2013.

So Intel has no interest in making a top-end graphics processor, but will continue to erode the low end of the market. AMD also has strong plans for its APU range to include better and better graphics solutions, PLUS its APU technology will form the basis of the top gaming consoles of the future – which could also eat into mainstream GPU sales.

Right now, the financial drive for monster cards like the 6GB 7970 from Sapphire, comes from the fact that AMD can sell lots of low end products. One AMD board partner recently mentioned that the biggest volume of orders, for them, comes from the low end of the old 5000 series. But if that market dries up – through improved APU performance and an increase in mobile solution sales – how can AMD finance the very top performing cards?

We hope they find a way, because the alternative is not attractive to anyone who believes in healthy competition.

If AMD were to pull back from top end graphics solutions in the 4-5 year time frame, then that would leave Quadro-funded nVidia cards with no competition in the peak-performance part of the market. Without competition at the top, the ‘halo effect' of being the ‘world's fastest' would only shine down the GeForce range – tipping the balance even further. Think reduced ice caps means more of the sun's heat absorbed, which makes the oceans warm even quicker.

KitGuru says: Looking at the available data, there is a revolution coming. Rory Read said, on his Q3 earnings call, that the PC market will change completely in 10 years time. Right now, AMD has no stronger product than Radeon, but there must be huge question marks over how future generations of flagship card will be funded if the larger market for sub-£100 discrete graphics cards continuous to be eroded by integrated solutions from Intel and AMD itself.  If nVidia has limited competition for high end workstation cards, then it will continue to drive that market – with the fall out being next-gen desktop parts. Without the ‘volume products', where will AMD's R&D funding come from?

Comment below or in the KitGuru forums.

Become a Patron!