Will AMD’s new CEO make a smarter choice?

Bad news from AMD this morning: a quarterly loss of $1.2B. That’s rather worse than last quarter’s losses of $360M, and it’s led to the departure of CEO Hector Ruiz.

It’s a shame, but really not that surprising. AMD has some great products, but it doesn’t seem to know what to do with them. In fact, its behaviour in general hasn’t made a lot of sense lately.

I mean, right now, AMD has not one but two really excellent processors on the market: the Phenom CPU, which gives the Core 2 Quad a run for its money, and the RV770 GPU, as found in the Radeon HD 4850 and HD 4870 graphics cards.

But, as I’ve observed before, the company refuses to give the Phenom the promotion it warrants. And while graphics card manufacturers have made a success of its GPUs, AMD insists on pumping these revenues into hopeless conceits.

Platform games

The most recent example of this is Puma – the notebook “platform” launched in early June that really amounts to little more than a handful of new mobile chips and a polite request that manufacturers use them. There’s no marketing budget, and no incentive scheme for system builders. So guess what? Six weeks after the launch, AMD is still waiting with tragic optimism for a single Puma device to go on sale.

And Puma is far from the company’s only recent folly. Last year we met Spider – the quad-GPU desktop platform that, oops, AMD couldn’t afford to promote. Before that, we had AMD Live!, a home entertainment platform taking in both desktops and notebooks. Remember the big splash that made? Seeing a pattern here?

Scaling back

This obsession with the “platform” indicates that AMD wants to define a clear role for itself in the mainstream market. That makes a degree of sense: it’d probably be easier to sell AMD hardware if the man in the street had some notion of what AMD stood for.

But right now it simply doesn’t have the clout to make a success of such grand projects. Intel was able to roll out Centrino 2 with military discipline because it was willing to market the hell out of the thing and (so I’m told) offer generous incentives to manufacturers to get on board. AMD staffers openly admit that that’s way out of their company’s price range. And that’s why, by comparison, Puma has been such a squib.

The sensible thing for AMD to do now is to drop its pretensions. Having your own platform is great if you can pull it off, but the company needs to be realistic. It must remember that it is primarily a chip maker – indeed, as the Phenom and RV770 cores show, a very good one. If it would only focus on this strength, and stop wasting money on doomed fripperies, it could turn its financial position around. In a few years, it could even have enough money put aside for a genuinely effective platform launch.

Precision engineering

And under new CEO Dirk Meyer it could conceivably happen. Some commentators have argued that, as Ruiz’s personally-anointed successor, he stands for “no change.” But given the circumstances of his arrival, we may assume he realises that AMD’s current strategy isn’t working.

And Meyer is an engineer, not a blue-sky visionary. He led the team designing the original Athlon – one of AMD’s greatest successes – so I suspect he’s bright enough, and practical enough, to identify the problem and understand what needs to be done. In his maiden speech as CEO, he promised “sustained profitability through a focus on the core technologies” – a turn of phrase which suggests the axe may indeed be sharpening for distractions like Puma.

I hope it works out. Maybe AMD can’t carve out a new identity for itself right now, but it’s nevertheless played an important role in the industry for many years. Even if you use its rivals’ parts, you can thank AMD for providing competition. You can bet the Core 2 Duo would be more expensive without the Phenom, and the forthcoming GeForce 9800 GTX+ probably wouldn’t even exist without the Radeon HD 4850.

But the bottom line is, well, the bottom line. If AMD’s to keep playing that role, it really can’t afford to keep losing a billion dollars a quarter.