So just how widespread are NVIDIA's GPU failure problem. According to NVIDIA, it's nothing to worry about.
OK, let me get this straight. A couple of weeks ago NVIDIA file a report with the SEC (PDF here) stating that the company was taking a $150 - $200 million charge to cover "anticipated customer warranty, repair, return, replacement and other consequential costs and expenses arising from a weak die/packaging material set in certain versions of our previous generation MCP and GPU products used in notebook systems. All newly manufactured products and all products currently shipping in volume have a different and more robust material set." This filing claimed that "products that are impacted were included in a number of notebook products that were shipped and sold in significant quantities" and that certain notebook configurations were "failing in the field at higher than normal rates."
What NVIDIA didn't tell us at the time was what GPUs were affected. However, a new software driver was released to help combat the problem by kicking up the cooling fans a few notches. This lowered the thermal stress on the parts, but at the cost of noise and battery life. NVIDIA also spread the blame about, pointing the finger at the suppliers, OEMs, and customers.
Next came a report by The Inquirer which claimed that the problem wasn't specific to a few parts but that it affected all NVIDIA G84 and G86 GPUs. That turns the problem from being a manageable one into a huge one.
Now Ars Technica has some feedback from NVIDIA that seems to play down the matter:
- The issue is limited to a few notebook chips only; we have not seen and don't expect to see this issue on any NVIDIA-based desktop systems.
- Only a very small percentage of the notebook chips that have shipped are potentially affected, and the problem depends on a combination of environmental conditions, configuration, and usage model.
- We continue to work closely with our partners and have taken the necessary steps to ensure that all NVIDIA chips currently in production do not exhibit the problem.
What I'd like to see from NVIDIA is more details as to which parts are affected. It's no good telling us that the problem is limited to a few notebook chips without being clear as to which they are.
I'm taking a cautious view of NVIDIA's claims here. If the problem really is a small one then the one off charge will be enough to cover it, but if we see NVIDIA taking further charges in future quarters relating to this then we'll know that the problem is bigger than NVIDIA is letting on.
In other GPU related news ...
Was there collusion between ATI and NVIDIA to keep prices high? An email sent by NVIDIA's senior vice president of marketing, Dan Vivoli, to ATI's president and chief operating officer, Dave Orton seems to suggest that there was. Here's the tasty part:
"I really think we should work harder together on the marketing front. As you and I have talked about, even though we are competitors, we have the common goal of making our category a well positioned, respected playing field. $5 and $8 stocks are a result of no respect."