Intel Finds a Fit with Photonics

Stephanie A. Weiss

Intel Corp. has extended its reach from electronics and semiconductors into photonics with three recently announced acquisitions.

In April, Intel announced that it had agreed to acquire LightLogic Inc. of Newark, Cognet Inc. of Los Angeles and nSerial Corp. of Santa Clara. LightLogic manufactures transponders that combine lasers, micro-optics and high-speed electronics. Cognet and nSerial manufacture high-speed optoelectronic networking components.

Electronics meets photonics

Intel's interest in photonics came from a realization that this industry was going through many of the same growing pains that the semiconductor industry has faced: lack of integration and automation, and the size of wafers.

"We saw a lot of automation potential there," said Tony Stelliga, manager of strategic marketing for Intel's Optical Products Group in San Jose. In addition, "there are a lot of electronics used to generate the light."

Intel's most recent acquisition, LightLogic, extends the company's reach further into the space between electronics and photonics.

"We went through the discrete thing. Modules arrived a year ago," Stelliga said. "The 'you are here' is the transponder phase. It's a higher level of integration but in disguise."

He said the next wave will be planar integration. "It offers a higher level of integration at a lower cost. We haven't announced that we're looking at it, but it's inevitable that this single-chip thing happens."

LightLogic's precision computer-automated fiber-attach equipment, borrowed from the disk drive industry, is critical to increasing coupling efficiency.

Stelliga said the increased coupling efficiency and some electronic error-correction will dramatically increase power for the metropolitan network.

"You'll be able to reach 80 km for the price of 20," he said.

Stelliga said Intel has no immediate plans to start making its own lasers, but he also didn't rule out the possibility in the future.

"We don't actually build our own lasers at this stage," he said. "That's not to say that actives aren't going to be continuing to be important. I wouldn't preclude any activity there."

Stelliga said that improving the reach of a transponder by increasing laser power is very expensive. Improving it through electronics means that you can use a less-expensive, less-powerful laser.

"There's a lot more electronics that you can do to improve the performance of the actives. Forward error-correction is relatively straightforward for us," he said. "We think we can get that $80,000 transponder down to $10,000 by improving efficiency, without having to go out and buy a laser supply."

A hint about where Intel is looking?

"As we continue to get transponders down in size and cost, tunability is another trend that is important in the market. We're watching that, too."