NVIDIA was hit with a class-action lawsuit filed in the US District Court, Northern District of California on Tuesday. The class action lawsuit (PDF) alleges that NVIDIA violated Federal securities laws by not disclosing to shareholders that certain of its notebook GPUs were failing at higher-than-average rates. The suit seeks class-action status for all NVIDIA stockholders who purchased or acquired common stock between November 8, 2007 and July 2, 2008 and pursues legal remedies under the Securities Exchange Act.
On July 2, 2008 NVIDIA filed a report with the SEC that stated it would take a $150 million to $200 million one-time charge to cover anticipated customer warranty, repair and replacement and other costs relating to a higher than average percentage of failures in some NVIDIA GPUs used in notebook computers. The company clarified the size of that charge??”$196 million??”at its last quarterly meeting.
NVIDIA stated at the time that it was not able to determine the root cause for the failures of the GPUs and that the chance of failure was very slight. The company has maintained its claims regarding both the chance of failure and the number of customers likely to be affected, and has not noticeably changed its position since the initial announcement.
The closest NVIDIA has come to offering an explanation for the higher-than-average failure rates is to say, "testing suggests a weak material set of die/package combination, system thermal management designs, and customer use patters all contribute." Several different OEMs have been affected by the problem, but again, NVIDIA maintains that the total number of impacted users is small.
According to the suit, some market participants have questioned whether the reserve set aside is enough. Sources close to NVIDIA maintain that the $200 million number is adequate for the rate of failure.
The suit documents note that the day after the announcement was made regarding the one-time charge; NVIDIA's stock price plummeted to $12.98 from the $18.78 it was trading out the day before. The drop in stock price cut NVIDIA's market capitalization by more than $3 billion, and the stock has not substantially recovered. In the two months since its announcement, NVIDIA's stock has hovered mostly between $10 and $11 dollars, with a brief peak at just over $14 in the later weeks of August. As of today, NVDA stock was valued at $10.92 per share.
It's only fair to note, however, that there are several other factors that may have played a part in keeping NVIDIA's stock price low. These include a drop in desktop GPU shipments during the second quarter, losing the top GPU spot to ATI, and the Intel marketing machine praising the glories of Larrabee.
As for the lawsuit, it relies on a key statement attributed to NVIDIA's VP for Investor Relations, Michael Hara. On September 4, 2008 at a Citigroup Technology Conference, Hara stated "We've... been working on this problem with the customers for well over a year, going all the way back to August of last year."
In an effort to address the potential issue with the GPU, both HP and Dell released updates that were intended to circumvent the issue by changing the way the cooling system of the notebooks operated. The updates addressed things like fan speeds and runtimes along with how the system throttled the CPU and GPU for thermal control.
The suit alleges that the firmware updates are merely an attempt to postpone the GPU failure until after the warranty period and cites a ZDNet story that reported, "a BIOS update is little more than a band-aid and isn??™t going to fix the underlying problem."
Several statements posted to public message boards by owners of potentially affected notebooks are used as part of the documentation for the suit. Exactly how speculation and hearsay from owners of notebooks made in forum posts, not familiar with the technical aspects of the system or the issues at the core of the suit, have bearing on the outcome of the suit are unclear.
The central allegation made in the suit is that NVIDIA has known about the potential defects in the GPUs since November of 2007 and kept the possible defect and the effect the defect could have the NVIDIA's financial future a secret by issuing a series of false and misleading statements to the investing public.
The question regarding the delay in the announcement of the potential GPU defect in many minds has to be: how long does it take from an HP or Dell coming to NVIDIA with an issue being seen on a GPU to be studied in enough depth to find which side the problem is on. There has to be a process that looks at the abnormal failure rate to determine if the issue causing the failures is from a problem with the cooling design of the notebook in question--failing on the shoulders of HP or Dell--or if the problem is in the actual manufacturing process. The suit claims that NVIDIA knew the answer to that question in November of 2007.
Could it be that NVIDIA knew in November of 2007 that there was a potential issue with the part, and it took until the announcement in early July to determine that the issue was with the GPU manufacturing process, as opposed to a notebook design flaw, a cooling solution issue, or a particular batch of bad parts as opposed to an ongoing problem?
Another question is what actually caused the defect in the manufacturing process? Is the issue here ultimately NVIDIA's sole responsibility, or could it be that one of NVIDIA's manufacturing partners shares in some of the blame? Could one of NVIDIA's partners in GPU construction have provided faulty materials that failed to meet NVIDIA's specifications? Most will want more than unanswered questions, anecdotal evidence, and posts from forums before making a final decision on their own.
Source: Ars Technica