terminatorLast year was a terrible year for Nvidia. The company had a ton of inventory going into the fourth quarter, just as the demand for its graphics chips evaporated. Today, the company’s executives talked about how they plan to get their momentum back.

In 2008, Nvidia let Advanced Micro Devices get ahead in the graphics chip race. Nvidia’s partners also ran into reliability problems attaching its graphics chips to laptop system boards. And then the economic crisis hit, just as Nvidia was carrying more than a billion dollars in inventory.

The company’s long-term comeback plan is to trust in the engineers’ ability to deliver outstanding graphics chips. Another generation of high-end graphics chips is due soon, but Jen-Hsun Huang, chief executive of the Santa Clara, Calif., company declined to say when. He also declined to comment on current quarterly sales. That leaves some doubt in the near term.

The Nvidia plan includes a push into new markets. David White, the company’s chief financial officer, said the company hopes to regain its historical gross margins above 40 percent (currently, it’s at about 32 percent to 34 percent) by selling a richer mix of high-end chips. Those include Tegra mobile chips, Tesla and Quadro heavy-duty computing products, and the new Ion chipset for netbooks. But it’s going to take time, perhaps a year or two, before those product lines all kick into high gear. And Ion remains tied up in a lawsuit with Intel, a fact Huang acknowledged has definitely hurt its sales.

In the meantime, it’s going to be a very tough road, especially if there is no near-term recovery, says JoAnne Feeney, analyst at FTN Equity Capital Markets. That has led to a wait-and-see attitude about the prospect of a turnaround at Nvidia, whose stock is trading around $11 a share, down from $19 a share a year ago.

Nvidia and other computer companies are hoping to see a spur in demand in the fall, when new operating systems debut from both Apple and Microsoft.

huang-big-storyRegaining footing will be difficult. White acknowledged that the downturn has changed the way consumers think, leading to a permanent shift toward low-priced products such as netbooks, or small laptops meant to cruise the web. That means lower average selling prices for Nvidia’s mainstay chips. More consumers will likely settle for computers with low-cost Intel chip sets, which combine low-end graphics with other chip functions. Nvidia will also have to make sure it has a smooth transition to a new manufacturing process.

But Huang believes that the downturn has also accelerated a fundamental shift in computing, away from the costly serial processing of the Intel central processing unit, or CPU, and toward Nvidia’s graphics processing unit, or GPU. While Intel CPUs consist of a hefty processor that handles big tasks one at a time, Nvidia’s GPUs have a lot of little processors that handle small tasks in parallel. Huang believes GPU computing, or the use of graphics chips to perform parallel processing tasks even in non-graphics applications, is the future.

That’s why he believes Apple is spending more money on Nvidia graphics chips than it is on processors in some of its newest laptops. Other computer makers will logically follow Apple’s path, or find that their machines fail to satisfy consumers, Huang said.

Nvidia is also trying to get software developers to make more programs that exploit GPU computing, from the smallest of consumer applications to the heaviest of supercomputers. He says the resulting wave means that graphics processors will always be in demand, even if Intel and Advanced Micro Devices launch their hybrid chips that combine both a CPU and a GPU. Today, numerous analysts asked Huang what he would do when that happens, but Huang stuck to his answer that such hybrid chips will never dominate the market, much the way that a single car model never dominates the entire car market.

Huang says GPU computing has hit the tipping point, since more than 120 million graphics chips have hit the market with the capability. (AMD has its own version of GPU computing as well. If it catches on, Nvidia won’t be alone). This is the way the computer industry will get its sex appeal back, Huang said.

“The GPU has volume sales,” he said. “It is the best way to bring parallel computing to the world.”