Now Hear This!

The high in Austin today is 98 degrees, but for Dirk Meyer it must feel like 198 degrees. The heat is on AMD’s new CEO to rescue the company from its falling stock price, continuing quarterly losses, product delays, and the general embarrassment brought upon it in the last year.

Can Meyer do it? Well, that’s yet to be seen, but general industry sentiment toward a Meyer-run AMD (one that sees Hector Ruiz acting as executive chairman) seems favorable … at least so far. Of course, you wouldn’t know that looking at AMD’s stock, which hit $5.30, one of its highest points in the last 30 days, when AMD announced Meyer’s promotion on July 17, but had slipped to $4.23 by today’s close.

Perhaps the question isn’t really one of Meyer’s specific capabilities, but rather a question of how AMD can be saved at this point. This blog, as well as many analysts, have suggested AMD sell or spin off its fabs, take the cash infusion, and exclusively focus on design to reinvigorate it in its challenge against chief rival Intel.

AMD is no stranger to working with its spin offs, like Spansion, and such a manufacturing shift isn’t unheard of. Texas Instruments announced in January 2007 that it would change gears and work collaboratively with its foundry partners on the next generations of digital process technology, and it still stays very competitive. And as Bernie Meyerson, an IBM CTO, pointed out the recent Semicon West show,manufacturing isn’t cheap and certainly isn’t easy to perfect. Therefore, he noted, more and more companies are partnering to get their designs manufactured.

AMD is already buddy-buddy with IBM and, through its ATI buy, TSMC. Perhaps AMD’s hesitance in announcing its asset-lite strategy plans has to do more with waiting on its options than on its own internal reorganization. Intel is at high-volume 45-nm high K/metal gate manufacturing, planning 32 nm. IBM and TSMC, meanwhile, are moving to 32 nm and duking it out, each claiming superiority to the other. If one could provide long-term proof of manufacturability superior to Intel’s in cost, power, etc, odds are AMD would then unzip its lips on its manufacturing plans.

Still, that’s only one option. While many EDN readers have adamantly commented on our Web site that they agree with the “go fabless” suggestion, many readers have steadfastly stood by the belief that AMD needs its fabs to remain competitive in our Intel-dominated chip world. That was echoed this week in a story EDN published (see “Will AMD separate manufacturing and design?”)  in which an iSuppli analyst maintained AMD should keep its fabs, instead splitting itself into two separate divisions, one for design and one for manufacturing. Within an hour of the story’s online publishing, reader comment was posted from “Adam,” a consultant in Hillsborough, Calif. The comment, below, has been edited for length.

“It both amazes and amuses me how smart people keep on behaving 'stupidly.' Organizing and managing huge companies is indeed a challenging task; but change just for (what appears to be) change's sake is an all too common occurrence. All this off-shoring/out-sourcing bit is practical ONLY for low-tech stuff … The complexity of high-tech just does not justify these remote-nesses. … There is too much at stake in the way of competitiveness erosion, both at the company and national levels. The relatively short-term benefits of lower costs come back to haunt us in the not-to-distant future; by which time it becomes difficult, if not impossible, to rebuild the infrastructure which has degraded in the process. This same logic applies to the fabless idea. Run/manage the business well, and keep as vertically integrated as possible on the HIGH TECH portion (the 'company jewels’) of the business.”

To be true, there isn’t much money out there now for anyone to buy AMD’s fabs, so the company finds itself between a rock and a hard place. Rock: Keep the fabs, benefit from an in-house manufacturing operation, but continue to hemorrhage on the high investment. Hard place: Sell the fabs at a low price or loss, drop the manufacturing, and focus on design innovation to snatch some of Intel’s MPU market share.

What will this new Meyer-headed AMD do? Time will tell. Meanwhile, a second comment to our AMD story has been posted by “AMDfan,” who succinctly sums up the situation: “Whatever you decide to do AMD, just do it right!”

Share your thoughts on AMD’s options and Meyer’s new role below.

--Suzanne Deffree, Managing Editor, News