WiMAX: What is Intel Thinking?

Intel will shortly announce its long-awaited "Montevina" Centrino chipset that includes WiMAX functionality. While this is an obvious direction for Intel to go, I'm left scratching my head a little over this one. I have never advocated buying a notebook with a built-in WWAN adapter; that's just too much of a lock-in to a network that might not be the right choice everywhere - and you get to pay for it with hefty monthly charges whether it works for you or not. In the case of WiMAX, which will have only limited commercial availability this year, such a lock-in would be counterproductive for those who travel beyond an area with service availability. Ultimately, it will be critical mass that determines the success of 4G technologies - enough availability in enough places to convince potential customers to make the investment and commitment.

The future of WiMAX should be assured. After all, Intel has poured billions into WiMAX, from chipsets to investments in service providers. But the question, of course, is why. Intel is a chip company, and is always looking for high-volume opportunities to sell vast numbers of chips. Shouldn't, therefore, Intel remain relatively agnostic when it comes to next-generation developments? Perhaps Intel originally thought that WiMAX would achieve critical mass well in advance of LTE, and thus LTE would be no more than a runner up. That's clearly not the case, and there's evidence now that Intel realizes this. Intel CEO Paul Otellini has said that WiMAX is spurring the development of LTE, and that's a good thing. Sean Moloney, who spearheaded WiMAX development at Intel, says he'd like to see the harmonization of WiMAX and LTE. That won't happen anytime soon, but why would he make such a suggestion? Scott Richardson, Chief Strategy Officer for Clearwire, made a similar suggestion today, by the way, during a WiMAX Webcast announcing an intellectual-property consortium.

Intel's a chip company. They live for high volume. They dump products (anyone remember the iAPX 432  or the i860?) that don't achieve high volumes, even if they've sunk copious amounts of cash into their development and marketing. There is little doubt in my mind that the volume of LTE chips sold will far exceed that of those for WiMAX, and why would Intel care whether their foundries are turning out vast quantities of one or the other, as long as all those chips have an Intel logo stamped on them? Conclusion: Intel's working on and will announce LTE chips - and perhaps as early as next year.

I think the right model here going forward is typified by Qualcomm. Qualcomm is best known as an intellectual-property (the other IP) firm that extracts huge royalties from sellers of CDMA-based products. But Qualcomm now has over 700 patents in OFDM, the technology behind 802.11g, .11a, .11n, WiMAX - and LTE. And while we'll undoubtedly return at some point to the IP threat that all of these face from an aggressive Qualcomm, Qualcomm also sells chips that cover essentially all major wireless technologies - but, curiously, not WiMAX. Again, it's all about volume. Qualcomm thus points to where Intel needs to go, bringing Intel's own significant IP position along for the ride. This is not a time to place all of one's eggs in one basket.

Given the uncertainty that is always part of any major new technology's rollout, one cannot afford to be too religious. When all religions appear to offer the same path to salvation, it's best to sell the customer the religion that they want, and avoid trying to convert them. The latter path has always been, absent the use of force, more than difficult.