Programmable Notes

AMD's Programmable
Logic Spin-Off

Their competitors are skeptical, but AMD believes
it is the business model for the 1990s.

 by Larry Waller

Whatever the ultimate merits of AMD Corp.'s plan for spinning off its Programmable Logic Division, it has been topic one in the business since the Sunnyvale, CA firm made its sudden March announcement. Notwithstanding whether the move can spur sluggish growth of the one time programmable leader, colorful CEO W.J. (Jerry) Sanders III with customary flair raised the stakes with his ambitious comments.

Sanders contends the independent AMD unit will not only provide faster growth but will quickly set a new organizing standard. "Today's model is the small fabless chip company totally focused on the specific demands of the business," he says. "We can improve on this model by creating a subsidiary that combines the best of both worlds: the flexibility and focus of a small company and the resources of one of the world's premier microchip manufacturers." The new fabless unit will buy its wafers from parent AMD.

Making the Programmable Logic Division a wholly owned subsidiary finds nearly universal approval both inside and outside the company. A chief goal is to build a dedicated sales force that hawks nothing but programmable products. AMD found out the hard way that sales force attention for programmables cannot be shared successfully with microprocessors and other logic products--not when such tightly focused competitors as Altera Corp. (San Jose, CA) and Lattice Semiconductor Corp. (Hillsboro, OR) have hundreds of sales and applications people making programmable customers happy.

AMD officials point to this major deficiency in their business approach as the driver behind their move. "We were frustrated by the PLD business, it requires a focus that our (shared) sales people didn't have the time to give," notes Dave Chavoustie, new vice-president of sales and strategic marketing for the subsidiary. Moreover, the problem doesn't stem from sales focus alone. It is complicated by a market that AMD says has evolved from a standard commodity business to one heavily emphasizing design software and proprietary architecture. "Our problem was that we couldn't even get enough time to train them (sales and applications engineers) with our software."

Nothing controversial about needing a dedicated sales force. The industry holds it as a basic tenet of programmables. Both Intel Corp. (Santa Clara, CA) and Texas Instruments Inc. (Dallas, TX) ran into the sales limitation barrier earlier in the 1990s and opted out of the business.

In fact, it was the "Silicon Valley" grapevine churning out stories that AMD was about to take a similar path that precipitated the sudden announcement. "It was a little premature, even from our point of view," Chavoustie concedes. "Our competitors were spreading rumors that we were leaving the business, and we had to dispel that myth."

According to observers, AMD's programmable operation really had been on the block and several firms had looked closely at it. The hang-up, they say, was sky-high pricing. Set at about three-times sales, the total tab was in the $1 billion neighborhood. "This was too much for anybody," says one source.

AMD declines comment, with officials emphasizing improvements in the works, corporate commitment of resources, and executive talent to make things happen. Reportedly, CEO Sanders helped pick Chavoustie, formerly a top official at the Embedded Processor Division. He also influenced the choice of Al Frugaletti, previously vice-president for distribution, now vice-president for worldwide distribution and programmable sales in North America.

Frugaletti is a veteran programmable executive, joining AMD when it acquired Monolithic Memories Inc. in 1987. Rich Forte, AMD group vice-president, is president of the subsidiary, and Frank Barone, vice-president and general manager, will run the day-to-day business. Chris Henry also continues as director of distribution and reps.

Commentary on the spin-off was generally favorable, especially from AMD's customers, according to Frugaletti. "We were concerned about their reaction, but there were no discouraging words." Most of them made remarks along the lines of 'what took you so long,' and 'it sure makes sense,'" he notes.

Industry watcher Rhondalee Rohleder, president of Pace Technologies (Scottsdale, AZ), sums up the reaction of most observers, "it puts them on a more level playing field" with the competition. Sales people who concentrate on one product are a necessity as devices reach higher complexity, in her opinion. This became apparent for AMD, which had a more entrenched programmable position than other broad-line chip suppliers dropping out of the contest.

Altera sees the AMD spin-off as a rather ho-hum event, affirming what everyone already accepts, contends Bob Beachler, director of communications and strategic planning. "It's clear that broad-line semiconductor suppliers have great difficulty in the programmable market, and are unsuccessful in gaining a significant share," he says. "As for AMD, the company needed re-energizing and re-invigorating, but it's still unclear what they'll end up doing."

After acquiring MMI, AMD was the clear programmable sales leader, but fell behind Xilinx Corp. (San Jose, CA) in total sales two years ago and Altera in the past year. AMD had CMOS PLD device revenue of $256 million in 1995, with $95 million coming from the more complex MACH device lines. The fading bipolar units added $68 million for a total of $324 million PLD sales, which puts AMD as a solid number three in the industry.

One significant fact about the spin-off that emerged after the announcement concerns the fate of the AMD bipolar products. They don't go with the deal. Instead, the new entity acts only as their sales agent. The corporate parent retains their profit/loss management.

While some AMD folks are calling the spin-off "a startup," veterans such as Frugaletti point out important differences that can present problems if not handled correctly. "We're not starting from scratch with a white sheet of paper," he explains. Hundreds of customers need continuity of support, even while the supplier structure is undergoing radical surgery, and new people are being added. "Insuring a smooth transition with little or no interruption, that's the challenge we have in front of us," he says.

Disentangling from the corporate "mother ship" is no simple matter. Frugaletti lists numerous decisions to be made quickly that will broadly impact the operations of a new business structure. One example is the array of services AMD now provides, such as information systems that track orders, sales, inventory, etc. "We have to decide what to take and what to leave," he says. Similar decisions must be made with people who need to be in place as fast as possible.

On the plus side, the attraction of a fast-moving, new operation is proving strong to many people seeking to go along. Distributor personnel, who presently handle the AMD programmable line, and sales representatives interviewed for the new rep organization see career advantages in specializing. Once the word got out that AMD Programmable was signing on reps "we saw people of a quality and caliber significantly better than we had any reason to expect," notes Frugaletti. The subsidiary, scheduled to be named in mid-May, will be up and running before July 1, but tuning will continue through the year.

If the new programmable organization works as well as AMD hopes during a proving period through 1997 or so, more change could be in store. Jon Joseph, semiconductor security analyst with Montgomery Securities of San Francisco, speculates that AMD could well launch its programmable unit as a separate company. The fact that the plodding bipolar line was held out of the entity and that AMD is sparing nothing to make it go hints that "AMD is cleaning it up to spin off to the public." Also, Sanders and AMD were early backers of Xilinx and profited greatly from its Wall Street success.

Larry Waller is a contributing editor for Integrated System Design.

To voice an opinion on this or any Integrated System Design article, please e-mail your message to michael@asic.com.

integrated system design  July 1996