How would you sum up the competition you face in UMTS and Femtocell areas?
In UMTS, there is competition on the big base-station level from Nokia, Ericsson, etc. But innovation comes from the smaller side of base stations. We are going from macro base stations to pico and femto base stations. This is where we can offer cost effective chip solution.
On the femtocell we don’t yet see competition. But I am pretty sure there will soon.be some: companies like ST, TI, ADI and others are all exploring this space.
What about competition in WiMAX area?
WiMAX is a new market and we entered WiMAX three years ago as one of the early entrants. It is more data centric market. Data centric market has traditionally been open to new comers like PicoChip. On the telecom side, change in suppliers is made reluctantly because there are many standard conformations, etc. WiMAX is a much open market and the smaller ones have jumped in. On the big base station for WiMAX it’s mainly TI and Xilinx.
It can also be complicated sometimes. Freescale is sometimes a competitor as well as partner for WiMAX. They have a way to implement PHY which is part of the base station we specialize upon and at the same time we use their processors. ADI competes with DSP but we use other components from them. You have some smaller companies in WiMAX which are focusing on terminal end but may provide solution for base station also.
What sort of customer base do you have?
We have over a 100 customers for WiMAX and UMTS. All big names are using picoChip in WiMAX and now we have a few big names in UMTS as well. I would say there is no OEM in Japan or Korea which are not using picoChip. In China, US and Europe we have a number of significant customers, and are expecting announcements from some Tier 1 OEMs.
How many new customers did you sign up this year? Any indication of shipments you are handling per year?
We signed around 30 or 40: Axiom Wireless in US, ipAccess in UK, Ubiquisys in UK and through them I can say NEC, Motorola, and Thomson. This is mainly UMTS Femtocell. In terms of shipments its somewhere in many thousands this year so far.
Do you use a standard DSP architecture or do you have something of your own?
We have designed our own DSP which is 16 bits RISC processor. The world is moving to multi-core, and we have the leading multi-core DSP architecture. picoChip first generation solution is based on 300 16-bit processors. The reason why we decided to go that way is to be able to propose a parallel solution compared to regular processors that ARM or MIPS offer.
Our architecture gives customers better performance than any other standard DSP product that you have in market. The multicore DSP we have gives you the advantage of less power consumption. And we deliver about ten-times better performance-per-Dollar or performance-per-Watt than traditional DSP.
That said, some of our chips do integrate an ARM core. Whilst our multi-core array is very good for DSP we can use ARM for other tasks that benefit from a more conventional architecture and OS. Protocol stacks, router, and management code runs on an ARM.
DSPs are relatively costlier for ODMs. Why did you go for this option rather than ASICs?
Remember, we are in the infrastructure business. We are not addressing the mobile phone itself. The characteristic of the base station is that it is very complex. It demands a lot of processing power and it is very expensive. So the OEMs want to have those base stations quite flexible and adaptable to the changing and multiple standards like in WiMAX. In mobile as well you have 3G and then you have HSDPA, HSUPA etc. So it is important for them to be able to have a programmable and flexible solution and this is why we are using DSP in the base stations.
You have any academic affiliations and partnerships in terms of R&D?
Yes, we have close relationship with educational institutions like BUPT (Biejing University of Posts and Telecom), ICT (Institute of Communications Technology) which is part of the Chinese Academy of Science, CCL (Coverage Consultant Limited) and ETRI (Electronic and Telecom Research Institute of Korea). We have a mix of strong in-house R&D and partnership with other research institutes for various purposes.
You are a fabless vendor. What is your supply chain?
We source our silicon from TSMC, packaging and test from AST in Taiwan, and we also use some test house in Denmark to implement the high volume tests. But mainly TSMC and AST complete our supply chain and we are happy with the level of support.
You have recently made some Femtocell related announcements. What do you expect from Femtocells?
Femtocell is another new market which will contribute to the growth of chips. I think Femtocell is a logical evolution of the base station or the network architecture. Rather than having small number of big base stations, operators say it is more efficient (especially for 3G) to have a bigger number of smaller base stations. We are growing from the macro-cells to picocells and now to femtocells.
How big an opportunity is Femtocell for you?
In terms of volume and revenue I think femtocell is going to dominate. I will say it is probably going to be 70% femtocell-UMTS and 30% WiMAX.
Is there a commercially available Femtocell in the market?
Not yet. Not to the consumer. There is no commercially available Femtocell in the market. They are undergoing trials and operators like Vodafone and Orange are testing them. We will be able to see some commercial deployments early next year.
Certain vendors like Netgear have plans to integrate femtocell capability in other CPE? Do you think a licensed spectrum product will mix easily with things unlicensed?
Yes, absolutely, the first requirement is that some of the big operators have dual mode femtocell-WiFi CPE. Something like ADSL-WiFi-Femtocell CPE. The advantage of licensed band over the unlicensed band is security and quality of service. I don’t think femtocell is going to kill WiFi. I think they will co-live. It will be a multi-service type of approach for the operators. So there is no reason why they will not co-live.
What sort of revenue growth are you experiencing right now?
We grew 3-fold last year and this year we expect to see 3-to-4 fold increase again.
How is that split between software and silicon?
Our revenue is a mixture of software and silicon. Until now our revenues were predominantly software based - licenses and all that. But starting next year we expect to see silicon sales becoming the predominant part. For this year it will be around 80% software, boards etc and 20% volume silicon and next year we expect to see the reverse.
You recently completed fourth round of funding raising $27 million. What will you spend it on?
On existing products and ASSPs for UMTS, WiMAX, CDMA 2000 and other such areas. This is where main chunk of our funding will go in.