The End of the One Trick Wireless Pony. Or is it?

And then there were none.  Last month Silicon Labs acquired Ember – the last independent ZigBee chip manufacturer.  It’s good news for the Smart Metering industry as it’s secured a future for Ember, who have become the chip and protocol stack supplier of choice for a large proportion of smart meters, IHDs and home gateways in the market today.  It’s not such good news for the investment community, as the $72 million initial consideration from SiLabs is a little short of the $89 million investment that had gone into Ember.  But given the fire sales of the other ZigBee start-ups, it’s still not a bad result.

And it could be one of those excellent fits that don’t come along that often.  For Silicon Labs, it extends their radio technology into the hotly contested 2.4GHz band, complementing their very capable sub-GHz range of EZRadio PRO chips.  It also gives them what I’d consider to be the best ZigBee stack on the market.  And it gives Ember what must be a very comforting degree of financial security as well as a ready made range of sub-GHz radios, just at the point where the UK and Japanese smart metering communities are looking at 868MHz. 

But it’s not just Ember getting gobbled up.  A few weeks later, Samsung quietly acquired Nanoradio – the Swedish specialist in low power Wi-Fi for mobile phones.  Both Ember and Nanoradios played the standards card and had essentially become one trick wireless ponies – a fate common to many wireless start-ups.  Perversely, CSR did the opposite thing today, by divesting itself of much of its location technology, (which it had acquired from SiRF), to Samsung, who seem to be getting rather good at acquiring bits of wireless technology. In doing so CSR moved itself back closer to its Bluetooth roots.

Although the prospect of an acquisition is the raison d’être of most wireless silicon start-ups, I wonder whether this flurry of activity indicates that we’re nearing an end-game?  In which case, what comes next?

Ember’s acquisition is mostly about timing.  Smart Metering is one of those “jam tomorrow” markets which is always about to take off, yet stubbornly refuses to generate the volumes everyone hopes for.  In the current climate, with a lot of the stimulus money drying up, this is just what Ember needed.  It has dedicated itself to being the leader in ZigBee, but in doing so, became a wireless one trick pony.  Now it has followed the route of most wireless pioneers.

It’s always a gamble to bet on a single wireless standard, but the economics of wireless mean that for most start-ups, there’s little option.  In a previous article I wrote about the cost of bringing a wireless standard to market.  It’s eye-wateringly huge.  However, one of the things that makes a wireless standard successful amongst its peers is the level of investment that goes into generating start-up companies which fuel the development of chips and protocol stacks.  Often it’s the latter which is the major cost.  Many of these companies start with approximately equal number of silicon and firmware engineers and two to three years in find the latter outnumber the former by at least five to one. 

In general, it’s these more nimble start-ups that make the running, but inevitably, as timescales slip and costs rise, they’re gobbled up until there’s only one dominant company remaining, which is making just enough money to stay afloat.  In Wi-Fi, that was Atheros, which was acquired by Qualcomm last year.  In ZigBee, it was Ember, who have now been subsumed by Silicon Labs.  And in Bluetooth it’s still Cambridge Silicon Radio.  They’ve attempted to diversify with the purchase of SiRF and Zoran, but today’s move sees them moving back towards that original comfort zone of putting all their eggs in the Bluetooth basket.  It will be interesting to see whether they have the scale to retain the accolade of the independent short range wireless king?

It’s a lonely job trying to stay out there as the wireless standard bearer, which is why most disappear somewhere along the route.  But equally, the rewards can be great.  At the less risky end of the market there is still a strong demand for proprietary wireless chips. These are shipped in large volume to companies making key-fobs, remote controls, burglar alarms and a host of incompatible consumer products.  It’s a lower investment cost, as the main requirement is to have a clutch of bright wireless chip designers, without the overhead of a large firmware team.  There’s nothing wrong with that – there are still many applications where a proprietary protocol is the ideal solution.  But that business model starts to fall apart as two things happen:

• Users get the concept of interoperability, which they see as the desirability of connecting their phone or iPad to devices as diverse as door locks, running shoes and toy helicopters, and
• Investors notice the much higher potential for return if you’re lucky enough to be in the right place at the right time for a new wireless standard.  And investors sometimes like a gamble.

However, it can be a poison chalice for an established general purpose wireless company to throw its lot in with an emerging wireless standard.  It can equally be a very profitable chalice if you get the timing right, as Chipcon discovered when they jumped into ZigBee.  Their prime position allowed them to acquire Figure8 – the ZigBee stack provider, and then subsequently be acquired by TI.  We’re probably about to see whether Nordic Semiconductor’s gamble with Bluetooth low energy pays off as well.  But the costs and risks for taking that route can be high and others have fallen by the wayside.

Today consolidation is almost complete, at least for the major short range wireless standards.  That brings another risk to the market, which is whether the standards can continue to evolve.  Start-ups are good at innovation and have the energy to make a standard work. Bluetooth,Wi-Fi and ZigBee would not be here without the energy of the companies that drove them forwards.  Now the custodians of these standards – the companies who contribute engineers to the specification working groups, are much larger companies with a diversified portfolio – companies such as Qualcomm, TI, Freescale, SiLabs and NXP.  They’re less likely to gamble on a technology – they employ too many accountants, which means that the pace of evolution slowly starts to wane as they deem the underlying standard “good enough”.  Instead their focus moves to gaining market share for what they perceive as a largely static, mature standard.

Which brings us back to CSR.  Bluetooth Smart is potentially the next and only remaining billion chip market for 2.4 GHz chips.  (The next green field wireless area is white space, which is another story.)  CSR’s retrenchment goes against conventional wisdom, but may make sense.  Few realise just how enormous the Bluetooth chip market is.  And the new Bluetooth Smart (aka Bluetooth Low Energy / Wibree) could easily double that, possibly even making it the single largest radio chip market.    If CSR can pull that off, they’ll have reinvented the cycle and proven that the one trick wireless pony really can win the race.