The ATI acquisition
In July 2006, AMD announced that it planned to acquire Canadian graphics giant ATI for US$5.4 billion. This was presented as the creation of a ‘processing powerhouse’, combining AMD’s strengths in CPUs with ATI’s in GPUs, chipsets and consumer electronics.
On one level it was hard to argue. What chance would Intel and NVIDIA have when AMD/ATI could provide everything? AMD soon started talking about its ‘Fusion’ vision – one that is now coming to fruition in the form of Phil Hester's APU concept – and the line between CPUs and GPUs was potentially becoming blurred.
The only question was whether the deal represented value for money. At the time, the BBC seemed to think it did but, at $20.47 per share, it represented a 24 percent premium on the going rate. Furthermore, the last time ATI’s share price had been over $20 had been November 2004 and it had fallen to as low as $11.80 in June 2005.
Associated costs
Of course there were costs associated with the acquisition, not least in managing the human resource implications of such an undertaking.
The feeling at HEXUS is that, while the first 12-18 months of a merger are always the most difficult, perhaps AMD could have initially run significant elements of the two companies 'as is' – operating more or less independently of each other – and not rushed straight into full integration.
Perhaps this would've allowed a more considered cherry picking of both company's human resource assets, in conjunction with the benefit of time for reflection on the implications of its strategic decisions.
The biggest potential conflict was that, to some extent, divisions of sales people within the 'new AMD' were effectively competing against each other.
Those responsible for AMD CPU sales needed to flog processors on whichever platforms their customers were demanding. At that time, in the UK at least, this meant promoting AMD CPUs as the processor of choice for systems with NVIDIA chipsets and graphics cards.
Meanwhile, those responsible for ATI graphics sales weren't bothered whether a system had an AMD or Intel processor - or indeed AMD, Intel or NVIDIA chipset - as long as it had an ATI graphics card in it.
As Intel had finally regained the desktop-processor performance-crown, this inevitably meant that ATI graphics were actively being promoted as the GPU of choice for systems with Intel processors.
The integration created a lot of confusion and dissent internallyPresumably this wasn't exactly what AMD had in mind in its oft used mantra of being 'customer centric'.
As it was, our intelligence is that the integration of the two companies created a lot of confusion and dissent internally. Again, perhaps some of this was unavoidable, but we still feel it could have been handled much better and think that you'd struggle to find an informed AMD, or ex-ATI, employee say much different with any conviction.
Reshuffle
In the post-acquisition reshuffle, the decision was made to decentralise the decision-making at AMD from its HQ in Austin, Texas. In Europe this meant increased influence in AMD's regional power-bases of Italy and Germany. It would be revealing to know which country has the better of that equilibrium right now.The other danger of such an acquisition is the diversion of corporate attention from its day-to-day business. In 2007, AMD was consistently beaten to the punch by its two main rivals – Intel on the CPU side and NVIDIA in GPUs.
Even the deal's value for money has been called into question. Tellingly, by the end of 2007, AMD had reported a running total of ATI-related charges amounting to over half the value of the original buy-out – $2.7 billion and counting.
To better illustrate the sequence of events described, we've created a timeline of events, each hyperlinked to either an AMD press release or a piece of HEXUS reporting, for easy reference.
