Thinking Ahead: Altair Semiconductor's 4G Vision

Altair Semiconductor was founded in May of 2005 by Oded Melamed, Yigal Bitran, and Eran Eshed, who had shared backgrounds at Texas Instruments and at Libit Signal Processing, which was acquired by TI in 1999.

“We had a very, very extensive background in developing chipsets, and especially broadband chipsets,” says Eshed, now the company’s vice president of marketing and business development. “We developed, over the years, over 25 different mixed signal broadband chipsets in wired and wireless technologies, so we knew that we were in the communications business to stay.”

In looking for the right technology for the new company to pursue, Eshed says, the three saw a clear trend developing towards 4G wireless and decided to do their best to capitalize on it.

However, they did find some fundamental common elements among the competing technologies. “They were all based on OFDMA, they were all making use of advanced antenna techniques like MIMO and beamforming, and they were all based on IP,” he says.

And so, Eshed says, they decided to stay open to all possible outcomes. “We developed an architecture that is based on a very unique processor we’ve developed in-house that allows us to essentially address or implement whichever 4G technology becomes available, with almost the same silicon implementation,” he says.

That processor is the company’s proprietary Optimized OFDMA Processor (O²P), which is used in place of a traditional DSP. “We had to invest very, very heavily in this,” Eshed says. “We had to develop the processor from scratch and develop all of the toolsets around it… and we have the O²P patented in pretty much any direction, so we feel good with respect to the value of the IP and how well it’s protected—and it’s proving itself as a very, very effective cornerstone in the architecture.”

Today, Eshed says, the majority of Altair’s revenue comes from WiMAX—but they’re keeping an open mind. “We’re already heavily invested in LTE,” he says. “We’re one of the most advanced ASIC companies in the world of LTE, with a product that we’ll output in the first half of 2009, based on the same silicon platform.”

And two months ago, the company announced that it was selected by the Japanese operator WILLCOM to provide the chipsets for their proprietary next-generation XG-PHS offering. “With this very, very flexible processor that we’ve developed, we were able to bid on their business and win a significant deal… another 4G technology, with the same silicon platform, with different firmware and software implementations,” Eshed says.

Beyond flexibility, Eshed says, Altair is focused strongly on minimizing size, cost, and power consumption—the point being that most WiMAX chipsets in the market today just aren’t designed for the handset market. “Three, four, five years from now, the devices are not going to be laptops… they’re going to be mobile devices,” he says. “The advanced services will be consumed over different types of devices than we see today, but there’s something that is very common to all of them: they’re all going to have to operate off of batteries, and they’re all going to be handheld and small—so power consumption, size, and cost are going to be very fundamental elements for those solutions, and so we put our focus on these three factors.”

Eshed says that’s reflected in Altair’s product offering today, which consists of the ALT2150 WiMAX baseband chip and the ALT6150 RF transceiver. “What we have in the WiMAX product portfolio is a chipset that is about one-fifth of our closest competitor’s power consumption, is less than half of the size, and is about 20 to 30 percent lower in cost,” he says.

And there’s a real need, Eshed says, for that kind of offering in the marketplace. “If you look at what’s happening in the WiMAX space today, Intel can announce as many customers as they want, Sprint may announce however many ODMs and OEMs that supply devices to them as they want, but ultimately, none of them is a handheld device, and none of them is shipping today,” he says. “And this is mostly because of the lack of a viable semiconductor to integrate them, mainly because of power. So this is what we’re trying to change right now.”

Looking ahead, Eshed sees Altair inevitably partnering with larger baseband suppliers to ensure backwards compatibility with legacy networks. “LTE as a standalone technology will probably not exist in a large scale,” he says. “Any device that will support LTE will have to have backwards compatibility with 2G and 3G. And of course, for that, companies who develop LTE will have to partner with companies who already have and own 2G and 3G technologies. So looking to the future, we definitely see ourselves as a very dominant player in LTE, partnering with some of the larger chipset suppliers that are supplying 2G and 3G technologies today, and we see ourselves as a very significant player in WiMAX—with a focus on handheld devices.”

Jeff Goldman is a frequent contributor to Wi-FiPlanet. He is based in Southern California.