For a while know we have been wondering what will happen to the GPU maker Nvidia.
Earlier it looked in danger of being killed off as Chipzilla and AMD made the discrete GPU largely an irrelevance in the budget market by releasing chips that really did not need one. There were rumours that Nvidia was about to be bought because its share price was cheap and life should not be that bad for the outfit really.
This week however Nvidia’s stock price rose by nearly 11 percent, as analysts at the Wells Fargo uprated the outfit. Wells Fargo is better known in the UK as the outfit behind the Pony Express.
Wells Fargo seemed to think that there was an improving PC chip market about to happen. Frankly we think this is about as likely as a second season of the new American Torchwood, or the networks suddenly working out that cancelling Firefly was stupid.
Wells Fargo thinks that there will be a seasonal improvement in PC demand and recovery in the consumer PC end market. This will help both Intel and Nvidia.
But even if there is an unlikely increase in consumer PC sales, that will help Chipzilla, but Nvidia still has the problem that most consumers are likely to buy chips with GPU functions installed. They are not going to want to buy cheap and cheerful discrete GPUs.
But it is starting to look like the tablet is either going to be a long term thing or just another Apple fad. Tablet makers are starting to wish that they had not tried to follow Jobs down this route as sales are light. This means that Nvidia is not going to make much cash out of the so-called mobile boom.
Where Nvidia is hoping to do well is getting its GPUs into notebooks. There is something about the maths of getting notebooks into the shops which makes discrete GPUs still viable in this market. Multimedia notebooks still need decent GPUs and discrete cards are still the cheapest way forward. At the moment Intel has not worked out a way of getting more notebooks using Sandy Bridge.
But, AMD is onto that and could be killing off this market for Nvidia. Nvidia’s only hope is that such punters will want better graphics than Sandy Bridge and Bulldozer can provide, but since it has a similar hope for the standard PC market, it will be doomed if it keeps its fingers crossed.
One of Nvidia’s best bets is the developing markets such as China and India. In this scenario, Nvidia makes dosh from those who need PCs as the IT infrastructure modernises.
Nvidia’s problem here is that the world and its dog is hoping for a slice of the developing market’s pie. While there might be a market for budget discrete chips here, AMD will also be in there too with much cheaper products. Developing markets have been on the table now for a couple of years as a source of potential revenue for a lot of different companies. Everyone has a plan. It is unlikely that Nvidia will do better or worse than any of them.
Earlier this year Nvidia’s stock jumped on the somewhat strange news that it was getting into the CPU business. At the time it announced this the stock was about $15 and it jumped to about $20.
What Nvidia seems to be doing with Project Denver is to stick an ARM chip and attach it to its GPU. In short, do what Intel and AMD are doing. With Microsoft making sure Windows 8 runs on ARM chips, Nvidia can play in the same league as Intel and AMD with surprising less effort.
This, more than any any dreams of a PC boom, is more likely to be Nvidia’s long term salvation. If it can get its gear to go, it could be creating low power hybrid chips suitable for the mobile market just as the world wants them.
It is still a risky strategy. Firstly Nvidia will be following AMD’s Fusion strategy on gear which may or may not be so successful. Secondly it will have to start from scratch selling itself as an unknown against two competitors who know what they are doing.
Although Microsoft seems to think that ARM chips will be the new black, at the moment Steve Ballmer’s predictions are not proving reliable either.
Either way, Nvidia will have its work cut out for it, if it is hoping to really be the Comeback Kid.