PicoChip Plays 3G Cost Ace

U.K.-based chipset startup PicoChip Designs Ltd. says it has developed a "software system on a chip" product that will cut the cost of developing and manufacturing 3G base stations by tens of millions of dollars (see PicoChip Cuts 3G Costs).

Vice president of marketing Rupert Baines says the company is offering a multi-processor integrated circuit that, while optimized for use in wireless base stations, is effectively "blank," allowing the OEMs that build the final products to add the required software. The result is a programmable base station that can have the installed software upgraded, replaced, or changed once it is deployed in the carriers' networks, claims Baines.

PicoChip says it can cost $100 million or even more to develop a base station for commercial production, and that its hardware and software tools can cut those development costs for each OEM by as much as 50 percent, as well as reduce costs in the production process.

The product is at the sample stage and is being "looked at and worked with" by "some" of the leading WCDMA base station vendors. So, potentially, the likes of LM Ericsson (Nasdaq: ERICY), Nokia Corp. (NYSE: NOK), NEC Corp. (Nasdaq: NIPNY), and Nortel Networks Corp. (NYSE/Toronto: NT) will be evaluating the chipset.

Although the programmable nature of the product means the chipsets can be used to build base stations for any type of wireless network, PicoChip is focusing its attention on the OEMs building WCDMA products for 3G license holders. "WCDMA is going to be five times as big as CDMA2000 in terms of buildout. That's where the money is." It believes its addressable market was worth about $400 million in 2002 and will be worth $1 billion in 2006.

Another reason for the WCDMA focus is that the needs of the CDMA2000 base station vendors is well met by Qualcomm Inc. (Nasdaq: QCOM), which can offer cost-effective preprogrammed products to the OEMs. In addition, the CDMA2000 market is already maturing in terms of build-out and product specs.

WCDMA, meanwhile, is still in its infancy, with mass market network build-out expected to really take off in 2004 and 2005. As base station development time stands at 18 months or more, Baines says it's feasible that base stations using PicoChip products could be deployed by carriers in 2005.

So what is so great about the PicoChip product? "Our approach is specifically optimized for wireless base stations, which the current options aren't, and we also offer a great deal more flexibility to the OEMs," says Baines. "Our proposition is that a base station would need fewer of our chipsets, they would use less power, and the overall cost is less. Ours is a parallel processing approach, while the current base stations are just using more and more standard components to meet the demands of 3G. Also, using our product saves the OEM the enormous cost involved in setting up the production process of making the specific chips for a base station. With ours they only need to add their software."

He adds that the chips are standards-compliant and that PicoChip checks its products for industry compatibility using test products from Agilent Technologies Inc. (NYSE: A) and UbiNetics Ltd.

So, quite a story for the OEMs. But PicoChip is up against tough industry players, especially Texas Instruments Inc. (NYSE: TXN), the major supplier of chips currently used in WCDMA base stations. And TI is upping its game, too, with a new product that it also describes as "programmable" (see TI Offers New UMTS Chipset).

But IDC research analyst Sean Lavey says there is a vital difference between the two solutions. "PicoChip is offering a truly programmable product, which is a very attractive proposition given the regular tweaks to WCDMA standards. The Texas chipset is configurable, but not completely programmable. However, the new TI product does offer the OEMs greater flexibility and reduced cost in the short term."

Lavey adds that PicoChip's story may be attractive, but it has a lot to prove. "It is a very promising proposition, but there are still a lot of questions to be answered. They are not revealing too much, and it's hard to know just how cost-effective its approach is. What PicoChip needs to do is convince a top-tier OEM to do some prototyping. Then it would be in a strong position. But, for sure, the cost of developing and producing these base stations needs to come down a lot to help the OEMs and the network operators."

With just that sort of validation in mind, PicoChip is seeking additional funding to help take it to the next stage of its development. Baines says the company is "ramping up for round B funding." The $7 million first round came from Atlas Venture and Pond Venture Partners Ltd., and PicoChip is aiming to raise a further $15 million some time late in the first quarter or early in the second.

— Ray Le Maistre, European Editor, Unstrung
Be the first to post a comment regarding this story.
Sign In