NVIDIA in the red last quarter due to manufacturing issue

By Joel Hruska |

NVIDIA has published its results for the second quarter of fiscal 2009. The company's performance fell short of previous expectations for multiple reasons, and NVIDIA's forecast for the third quarter isn't much brighter, with projected sales below seasonal levels. NVIDIA reported revenue of $892.7 million for the second quarter, a decrease of some 5 percent. The company dipped into the red this quarter, with an operating loss of $120.9 million (22¢ per share).

Related StoriesAMD takes performance crown from NVIDIA with new 4870 X2
GPU shipments healthy in Q2 as Intel surges forward
GeForce 9 series rumored to launch in late February
NVIDIA's GTX 280: G80, Evolved
The single most important factor that contributed to that loss, of course, is the $196 million warranty charge NVIDIA admitted it was taking a month ago. The company fielded several questions regarding the size and nature of that problem and once again reiterated that the manufacturing issue in question affected a relatively small batch of parts, that the company remains fully committed to repairing those parts, and is working closely with all of its OEM partners. When asked whether or not it believed the problem had affected other GPUs, expected additional costs, or predicted design losses as a result of the issue, NVIDIA said no. Based on the company's profile of the flaw, the one-time $196 million charge will cover it.

A challenging market
NVIDIA could have potentially tried to blame its quarterly loss entirely on that one-time charge, but the company declined to do so. Instead, it presented a very candid picture of its own weak performance during the quarter and the multiple challenges it faced. One problem, according to the company, was that it had "underestimated its competitor," and had mispositioned its new product introductions. This issue has since been corrected, and the GTX 260 / GTX 280 are now far more accurately positioned against the HD 4800 series from ATI than they were when those products launched.

Weak sales in the second quarter drove up inventory levels, and greater-than-expected inventory levels, in turn, have slowed NVIDIA's transition to 55nm. NVIDIA CEO Jen-Hsun Huang claimed that the transition to 55nm technology is ongoing, but that we won't see these cards across the market until existing 65nm supplies have been depleted. The weak US dollar drove US OEM sales in European markets, but it also weakened the sales performance of native European manufacturers. NVIDIA claims this substantially impacted its business, as desktops in Europe have a substantially higher GPU attach rate than their American counterparts.

Huang pulled no punches in his evaluation of the company's second quarter. "Our second quarter financial performance was disappointing," he said. "The desktop PC market around the world weakened during the quarter, and our miscalculation of competitive price position further pressured our desktop GPU business. We have a great product line-up, and having taken the necessary pricing actions, we are strongly positioned again. Our focus now is to drive cost improvements and to further enhance our competitiveness through the many exciting initiatives we have planned for the rest of the year."
He then added, "In contrast, the rest of our businesses did not exhibit the same dynamics as our desktop business. The notebook GPU, MCP, and Professional Solutions groups grew a combined 27 percent year-over-year. Though we approach the near term with caution, we remain very optimistic about the expanding universe of visual computing and the exciting growth opportunities made possible by CUDA, our general purpose parallel computing architecture."

No plans for UMPCs
CUDA, Tegra, and mobile Internet Devices (MIDs) were all hot topics this afternoon. Huang was dismissive of the UMPC market, implying that NVIDIA has no plans to compete in such a space, but talked extensively on how MIDs are the future of computing. The company is also putting a massive amount of weight behind CUDA, which it apparently envisions as the software engine that will power adoption of NVIDIA GPUs within everything from MIDs to supercomputers. None of NVIDIA's forward-looking statements touched on NV70 or any particular aspect of a next-generation architecture, but focused entirely on massively parallel computing and the ways in which the current generation of GeForce products can be adapted to deliver tremendous performance increases.

NVIDIA plans to address its revenue shortfall by significantly cutting operating expenditures, which ballooned by 34 percent year-on-year, and believes strong sale of its 55nm products will improve the situation. The company does not believe it lost market share in its second fiscal quarter, and while it forecast a tough third quarter, it said nothing about a long-term stay within the red zone. Barring further one-time charges, the company may quickly return to profitability; take away that charge and NVIDIA pops back into the black by some $60 million.

As for the rumors and allegations of widespread GPU/chipset failures, NVIDIA has now denied them in its quarterly earnings report. If the company's statements are accurate, there's absolutely nothing to them. If the company is lying about the scope or nature of the problem, it's opening the doors to potential shareholder lawsuits.

Personally, I'm ready for this topic to go away. The scope of the allegations continues to expand, proof remains nonexistent, and NVIDIA has now officially stated that the scope and nature of the problem is confined to a relatively small batch of parts—which, incidentally, is what it has been saying all along.

http://arstechnica.com/news.ars/post/20080813-nvidia-in-the-red-last-quarter-due-to-manufacturing-issue.html