Putting our chips on Zarlink

Benj Gallander and Ben Stadelmann
Friday, July 6, 2007

A homely amphibian used to sing, “It’s not easy being green.” Well, shove over, Kermit, and hum along with this refrain: “It sucks to be a semiconductor company.”

True, all those neat little chips and circuits help to entertain us and make our work more productive, but the enterprises that produce them have a much tougher time making consistent profits than. say, those that manufacture cigarettes, which have dubious social value.

The reasons for this conundrum are complex, but the main issues are fast-paced innovation and low barriers to entry. The fuel of innovation is R&D, which means you need a staff of well-paid engineers and sophisticated equipment. That equates to a generous dollop of working capital.

You also have to correctly guess what the market will want and whether, from a profit-making standpoint, it is worthwhile to design and develop. If that assumption is off the mark, or the competition builds something better and cheaper, a decent return on investment is unlikely.

As companies work to stave off competition with new products and stay ahead of the pack, they also tend to undermine sales of their own older product lines, turning them into legacy laggards. This pattern of rapid obsolescence is why you so frequently see large figures alongside the expense line item “Inventory Write-off” on income statements.

Finally, though particular designs and processes can be patented, ideas diffuse quickly through the tech community. Once it is known that there is a good market for chip X that performs function Y, competing labs kick into high gear. Before you can say “pernickety pesky patent protection” three times, some factory in Taiwan is churning out a similar chip.

As mentioned in a recent article, we waded into the sector last December with our purchase of Solectron. On the Canadian side of the ledger, Zarlink, located in Kanata’s Silicon Valley North, was added. In both cases, the ante was a tiny fraction of what these stocks used to go for a decade ago.

We were comfortably familiar with Zarlink as we’d owned its predecessor, Mitel, and did fabulously well on it, watching it climb from a little over $1 in 1991 to our final sale at $22.45 in 2000.

One of the things that piqued our interest was when Kirk Mandy, an old hand from the Mitel days, took over as chief executive officer of Zarlink in 2005. Since then, Mr. Mandy has made progress in stanching the steady string of losses, but the riddle of the corporation’s pattern of falling revenue has proven harder to solve.

Like the bosses of many other North American semiconductor firms, Mr. Mandy has set his sights solidly on the road to “fabless” — i.e., concentrating efforts on the design of chips while outsourcing the manufacture of silicon wafers to foundries, mostly in the Far East. The balance sheet was also strengthened by the sale of some facilities to Intel.

One of the big questions surrounding Zarlink was what it was going to do with its nest egg of cash. There was also speculation that this asset made it a more attractive takeover target. That haziness was dispelled on June 25 with the announcement of the acquisition of Legerity Holdings for $134.5 million (US).

Texas-based Legerity, which has concentrated primarily on chips used in voice over IP (VoIP), had revenues of $113 million in 2006. For Zarlink, strengthening the product mix in its traditional area of strength, voice communications, makes sense to us.

The news was greeted with a burst of enthusiasm that blasted the share price upwards by 17 percent on huge volume. But the euphoria was very short-lived as taciturn comments from analysts rolled in, and the stock closed the day up by a slim penny, at $1.86.

“Prior to this announcement, Zarlink was a poor performing company with a strong balance sheet,” BMO’s Brian Piccioni opined. “Now Zarlink appears like a larger company with limited cash and higher working capital needs.” He cut his price target from $2.50 to $1.50. Ouch!

Despite the difficulties that make this industry so tough, there is one thing we know with absolute certainty: the production of silicon chips is not a registered charity. That means somebody, somewhere, somehow is going to make a profit, even as many others will fail.

We’ve made our bet that Zarlink will eventually join the select few in the winner’s circle, and have set a target of $11.49. Hark — are those the same hecklers in the peanut gallery we heard when Mitel was purchased at $1.01?