AMD Nearly Bought Nvidia, Could Be Building Next-Gen PlayStation GPU

When AMD held its Analyst Day a few weeks ago, the company focused on how it was fundamentally changing its approach to semiconductor research and product development. Years of attempting to compete with Intel drained company coffers and exposed it to tremendous competitive pressure. Consumers unquestionably profited -- Intel's prices for various parts dropped dramatically in the early 2000s as AMD entered specific markets -- but Sunnyvale took a ferocious beating in the process.

Forbes has published a major article on the company's turbulent last few years and current plans. It credits Dirk Meyer with saving the company after Hector Ruiz's tenure as CEO ended in mid-2009, but quotes no less a source than Jerry Sanders, AMD's founder, on why any attempt to branch into cell phones or netbooks would've been a bad move.  “Dirk was digging out of a deep hole,” Sanders told Forbes. “I think Dirk understood that when you’re in a deep hole you stop digging—it would have been a deeper hole if AMD had tried to branch into a new activity, like smartphones.” For the record, we thought exactly the same thing.

Several interesting tidbits have emerged as part of this investigation, including news that ATI wasn't AMD's preferred partner for a graphics merger back in 2005. The firm initially approached Nvidia regarding a potential merger. Talks stalled with Jen-Hsun Huang, Nvidia's CEO, insisted he be named CEO of the combined company. Ruiz went to ATI and got absolutely taken to the cleaners, AMD eventually wrote down half the $5.4B it paid for ATI (thus arguably proving Jen-Hsun's point).

Did Ruiz make the right call? It's hard to say. The ATI acquisition was badly botched; the capital AMD spent (and later wrote down) to acquire ATI was critically needed for fab improvements and product development. It took the combined company five years to deliver the first "Fusion" products when they initially promised to have hardware on the shelf within two.


Nvidia's G80 did a remarkably good job of ruining several years of ATI's life

Back then, AMD and Nvidia were critically linked; the latter's nForce chipsets were the preferred solution for AMD products. In 2005, ATI and Nvidia both reported a shade over $2B in revenue for the year. In 2011, Nvidia did $3.84B in business while AMD's graphics division sold $1.5B worth of GPUs. Hector's long-term vision was accurate, but the execution left much to be desired.

If rumors are to be believed, AMD's graphics revenue could be on the way up. Forbes is also claiming that AMD is hard at work on a GPU to power the next-generation PS3. If true, it would mean AMD would power at least two of the next three consoles; the Wii U is confirmed as relying on a Radeon GPU derived from the HD 4000 series. Given that AMD has also been named as supplying the XBox Next's GPU, could Sunnyvale have pulled a hat trick?

Maybe -- but chances are, it won't mean much to the company's bottom line.



Unlike the PS3, which began life with all the earnings potential of a dying Bantha, both the Xbox 360 and the Wii sold well from Day 1. The Wii, in particular, blew up sales charts -- at the end of June 2007, Nintendo had moved 9.27 million consoles. By June 2010, that number was up to ~74 million, an 8x increase in three years. Nintendo recognized record-smashing profits over this period, while AMD's graphics division struggled to break even. Console sales may create a halo effect around the company's graphics technology, but AMD's extremely aggressive pricing appears to have hamstrung its ability to profit off its own products. Whether Read will change this policy is yet unclear. We've critized the CEO's reliance on meaningless buzzspeak and his tendency to rely on a handful of beleagured adjectives, but that doesn't mean he lacks an overall strategy or the ability to inspire. After years of struggle, AMD deserves a moment in the sun.