Analyst forecasts for femtocell shipments (especially in the early years) have differed widely, and been revised heavily. A recent report from Infonetics grabbed the headlines with a figure of 154% growth*. This related to the growth in femtocell revenues (not number of units) from 2008 to 2009. Given the relatively low base of sales in the last two years, this isn’t surprising. What I’m hearing is that today’s shipments are dominated by Sprint (in the US) and Vodafone (in the UK), but that several other operators could be ramping up volumes during the remainder of this year.

What the femtocell component suppliers are saying

Component suppliers, at the bottom of the food chain, are a good indicator of market traction. They are also notoriously difficult to pin down – understandably unwilling to release commercially sensitive information. I pressed Rupert Baines of picoChip to say whether the forecast given by his CEO Nigel Toon of 1 million femtocell chipset sales by the middle of the year was still valid, and he was only prepared to confirm that they were ontrack to meet this goal. picoChip are the dominant supplier of femtocell chipset designs today, but their figures exclude the CDMA market (supplied by Samsung’s femtocell today and by Airvana for 3G soon). It also excludes chipset sales from their competitor Percello, known to supply Ubiquisys. The latter have been trumpeting a recent order for 100,000 units for the SerComm femtocell based on their design.

I have also been able to determine that around 500,000 crystal oscillators (known by the memorable acronym of TCVCXOs) have been sold by one vendor since January so far.

Since none of these figures comes from exclusive industry suppliers, they need to be supplemented by shipments from their competitors.

What the operators are doing

Operators are extremely cautious about revealing the actual numbers of installed femtocells for commercial and competitive reasons. Their suppliers are also bound to secrecy for similar reasons.

The largest deployments to date seem to be Sprint in the US and Vodafone in the UK. Our earlier interview with Lee McDougal of Vodafone SureSignal reported that they had exceeded their original targets. From personal knowledge (not scientifically valid I know), it’s remarkable how many people I’ve met who have bought or acquired one for their own use. Even those who had scorned the idea a couple of years ago now say they’ve needed one because they use Vodafone with a corporate phone, and have found it easy to install and that it works well. With continued heavy promotional marketing, I’d expect the installed base to continue to grow further. Vodafone see this as a real differentiator in today's very competitive market.

AT&T are slowly rolling out their 3G Microcell to new markets – it was launched in San Francisco earlier this month – and despite some strong criticism that it’s not free, is receiving praise from many using them. I’ve even got enquiries from small businesses who are keen to know when the enterprise version is coming to their area. (Answer: AT&T haven’t released any information about enterprise femtocells to date).

Infonetics reported that 50% of their sample of 16 operators plan to rollout commercial femtocell services before the end of 2011*.

Further growth opportunities

Most operators seem to go through a market trial/soft launch phase before ramping up to large/full scale rollout. As each operator goes through this process, the System Integrators learn more about potential issues, the products become more “battle hardened” and mature, and the time/cost of deployment reduces.

Softbank in Japan have been getting some attention by offering to give femtocells away for free to anyone outside their coverage areas. I’ve heard their CEO is really driving the adoption of this technology, knocking heads together in both mobile and fixed arms of the business. They’ve selected Ubiquisys as their femtocell supplier, who have been making a lot of noise about their sub $100 price point – let’s assume Softbank's order was for sufficient volume to reach that figure.

With the large markets in the US, China, Japan and  Europe at or near this market trial stage, it is easy to see that significant growth is likely to occur over the next 12 months.

Bottom Line

We need to be careful to define what any headline number relates to: component sales are likely to lead installed femtocells by several months, skewing the figure in these early stages considerably. Here, I’m talking about femtocells sold and shipped to customers.

ABI Research estimated around 350K femtocell shipments in 2009 (a revised estimate made in November 2009) but remained optimistic with their longer term forecasts for 2014.

My guesstimate is for around 1.8 million femtocells (of all types) to be sold before the end of 2010, somewhat higher than the 1 to 1.5 million being forecast by analysts about 6 months ago. But given the limited raw data to work with and many variable factors, I’d put an accuracy of +/- 500,000 units on it.

Numbers for 2011 should be considerably higher, and we’ll be in a better position to gauge this at the end of the year when more operators are in advanced trial or commercial launch stages.

--

*References: Infonetics Femtocell Report Summary http://www.infonetics.com/pr/2009/TOC/2009-Infonetics-CRS-Femtocell-Strategies-Service-Provider-Survey-12-11-09-Contents-and-Methodology.pdf