But its not all bad news either. Intel ships only mini-PCI slot solutions. The other companies (including Marvell and TI) ship chips used in wireless access points and for consumer electronics, like cell phones. So, Intel will have to branch out to these other markets at some point, or simply plateau when Wi-Fi in computers starts to saturate.
Its easy in chips to commodify: to make a previously special product one that is just cranked out like a jelly bean. Wi-Fi chips are starting to seem like commodity goods. Many vendors produce them for different products, and the prices have been dropping.
One recent weekly flyer for a national BigComputerStore chain touted a Wi-Fi access point and computer card combination for under twenty dollars after rebating. Its hard to see how anyone can make money at that kind of price point.
Intel may try to differentiate its products with Centrino stickers and a big ad budget; but its the wireless functionality that matters to consumers, not the branding.
Consumers will buy the wireless functionality with their new machines or add it on to their old ones. When they buy it new, Intel does well. If they add it on to older machines, then the other makers have a chance to sell something.
Future wireless chips will be software-definable radios that can change physical configurations on the fly (use 802.11a or b or g depending on the situation).
Indeed, Intel has announced that it will produce just such a chip that also has the capability to deliver WiMax functionality.
This is a little surprising, considering that the WiMax standard hasnt been finalized for mobile use yet, but Intel evidently thinks it can tweak the chip no matter which of the competing proposals is eventually adopted.
But Intels competitors wont be standing still, especially a company like TI. TI can do software-definable radios just fine, as it has shown.
The real problem for competitors like Broadcom and the others is that they have been able to sell Wi-Fi chips and use the resultant revenue stream to fund development of other kinds of chips. For instance, the brain of Apples iPod Shuffle is a Broadcom chip. (Fun fact: The chip used in the Shuffle actually has a built-in FM receiver that is not used by the Shuffle. It fit better with the Shuffles target audience to only do stored audio well and ignore mediocre FM reception.)
Losing bread-and-butter Wi-Fi chip sales will have to impact what these manufacturers are able to develop, and so limit their eventual product lineup. Thus, the impact of Intels current sales surge wont show up immediately, but may well end up shaping what its competitors will be able to offer in the future.