While Intel nurses a self-inflicted wound after pulling the plug on its long-anticipated Larrabee graphics chip, analysts said rivals AMD and Nvidia need to make hay with the extra time they've been given.
Last week, Intel announced that it was nixing the release of its Larrabee graphics processor . Larrabee, which had been loosely defined as a graphics chip with between 10 and 100 cores, initially was slated for release in 2009. Then Intel pushed the release date back to 2010. Finally, it pulled the plug on an official release date all together.
Larrabee was to be Intel's first discreet graphics processor.
Nick Knupffer, an Intel spokesman, told Computerworld that the silicon and software for Larrabee were behind schedule so release plans had to be cancelled. He would not say what problems caused the delay.
"We remain committed to getting a graphics product out in 2010," said Knupffer. He declined to say whether that product would be Larrabee or a different graphics processor, but stressed that there would be a many-core graphics chip announced in 2010.
"We have been talking about Larrabee for quite some time - since 2007," said Knupffer. "We are disappointed the product is not where we expected it to be. Obviously, this is not our ideal scenario."
Rob Enderle, an analyst with the Enderle Group, called this a significant cancellation for Intel, one that could help its rivals, AMD and Nvidia pull further ahead in the graphics realm .
"This represented Intel's major push into high-end graphics and their response to AMD's Fusion efforts. It had high strategic importance," said Enderle, adding that there was some doubt in the industry as to whether Intel could pull Larrabee out. "This was a substantial miss, given how much focus they had put on this part. Actually, this is both bad news for Intel and good news for AMD and Nvidia."
Dan Olds, an analyst with the Gabriel Consulting Group, noted that for Intel to be top dog in all things semiconductor , which seems to be on the company's to-do list, it needs to throw its hat into the graphics chip part of the business. Does Intel need the graphics business to be successful and stay in the black? Definitely not. But it does want that business.
"Nvidia has been moving to gain the high ground very quickly over the past few years and they've done a good job of staying ahead of both Intel and AMD," added Olds. "Larrabee was supposed to be an Intel response that would give them a significant advantage of Nvidia. With the death of Larrabee, this gives Nvidia a clear path for the next couple of years, at least.... This definitely gives both Nvidia and AMD more time for them to make chips while the sun shines."
Olds also noted that the fall of Larrabee shows that building a graphics processor is no easy task. However, he said he's sure Intel has learned from this experience and is regrouping to launch a graphics chip in the next year or so - whether it's dubbed Larrabee or not.
"This reinforces the idea that building these kinds of high-performance chips is hard to do and that even the strongest players will sometimes have to take one step back in order to take two steps forward," he added. "While Intel gets a bit of a black eye over Larrabee, they're still pushing forward."
And Ezra Gottheil, an analyst with Technology Business Research, said he thinks that canceling Larrabee shows that Intel is pushing hard when it comes to innovation...and for that, he's applauding them.
"I'm on the side of companies that try hard to make things," he said. "Innovation is risky. If you always do what you're trying to do or planning to do, you're not setting high enough goals. Virtually every significant product in our industry was later than promised and less than we had hoped for.... Making all your dates means you're playing it too safe."
So, Intel's slip gives Nvidia and AMD more time to solidify their lead in the graphics processor market. But if Intel really wants to play in this arena, how much more time can it take before it has to ship a product -- or be left behind?
Enderle said Intel gets more time than usual to design because the graphics processor market is so young. "Intel probably still has around five years to have something competitive in the market," he added. "However, the longer they take, the harder it will be for them to catch up. Both AMD and NVIDIA are driving a technology advancement speed that is impressively fast."