Intel’s most important technology announcement of the year, was yesterday, when it said it had figured out a way to build chips with a radical new design, a “3-D” transistor, which has a “fin” structure to greatly lower power consumption while increasing performance.
That’s exactly what it needs to better compete against ARM, the British designed microprocessor that dominates the market for smartphones and tablets. But “killing” ARM won’t happen anytime soon. Intel can’t afford it.
Intel is doing extremely well, it recently reported record revenues and profits, and in this current quarter, it will likely report revenues of $1bn a week.
It is the dominant chip supplier in all categories of computers — except two: smartphones and tablets.
It has its Atom microprocessor, which is used in netbooks and competes against ARM, but it has hardly any design wins in the smartphone and Tablet categories. And that rankles Intel, or at least it rankles Intel watchers. Analysts and reporters are continually asking Intel about its plans for the trendy tablet market.
Yesterday in San Francisco, Intel was asked about its plans to use its Ivy Bridge 3-D transistor micro-architecture in its Atom lines. It said that server chips will be the first out of its factories sometime next year, and that Atom will come later, with no date given.
Why the wait when the performance characteristics of Ivy Gate show that it could be used to outperform ARM chips?
If you follow the money you nearly always find the answer. Its Atom microprocessor business was just $370 million in the most recent quarter.
That’s out of a total revenue for the quarter of $12.9 billion, where its data center chip business grew 32%, to $2.01 billion.
Server chips are its most profitable chips, next comes its PC microprocessors, which were a $6.8 billion business in the quarter.
Yes, Ivy Bridge Atoms could be ARM killers but Intel can’t afford to put that at the point of its spear. It has to go after the big money first.
Smartphones and tablets are sexy in that they get a lot of media coverage but those markets don’t add up to much in terms of profits for Intel. Killing ARM will have to wait.
Waiting also means that ARM will be harder to kill because there will be a massive amount of ARM software applications and operating systems. Migrating off of a platform is very tough for manufacturers.
Apple has done it several times and it knows how to do it well but for others, it’s probably not worth changing platforms.
So that means Intel will have to get used to the questions about ARM and get used to living in a world where ARM has the majority of design wins in a highly visible market.
But here’s another scenario: It’s not well known, but Intel possesses an architecture license to build ARM-based chips. It might choose to integrate ARM designs into a future System On a Chip product that gives customers access to both ARM and Ivy Bridge.
An ARM chip built with Intel’s 3-D transistors would be a formidable product: super-low power consumption with great graphics performance.
[Cross-posted at Silicon Valley Watcher]