Joep van Beurden

Interview: Joep van Beurden CEO of wireless chip maker CSR

What do you do when the invisible hand of the market gives you a really good shellacking? CSR boss Joep van Beurden explains his strategy for the wireless chip company to E&T.

Wednesday 27 October wasn't the best day in Joep van Beurden's time as head of CSR, the listed wireless chip maker. Before most people's working day had even begun, the company's share price had been marked down more than 18 per cent due to concerns about the quarterly results that van Beurden was due to present later that morning. By the end of a day of cross-examination by press, analysts and investors, the stock was still down 10 per cent. Van Beurden's crime? To present a good set of quarterly results and to highlight possible issues with the next quarter.

This, then, is the life of the CEO of a highly successful fabless semiconductor company operating in the fiercely competitive Bluetooth, Wi-Fi, FM and GPS chip markets. As van Beurden tells it, if your latest quarterly results are bad, the market focuses on what went wrong. If the results are good, the market wants to know what you're going to do for it next. He's sanguine, though: 'At the end of the day the system is designed to keep you on your toes.'

To details, then. In CSR's third quarter the company sales were $222.1m, $12.2m more than the same quarter last year, its gross margin improved 2.1 per cent to 47.4 per cent, its operating profit was up $18.6m at $29.2m and earnings per share were up 10¢ at 19¢. The fly in the ointment was guidance that the fourth quarter would be less lucrative than expected, due to a weakish economy, shortages of manufacturing capacity for GPS chips, and a hole in the product portfolio where a chip for the rapidly growing smartphone sector ought to be. Given the circumstances, van Beurden was prepared for 'something more savage' from the analysts and investors who marked his stock down that morning.

'We're a public company so the shareholders are very important to us,' he says. 'We serve at their pleasure. Our reason for being is to focus on shareholder value.'

Tackling the issues

It's not as if CSR is being negligent about the issues it faces. The company has been caught out by a worldwide squeeze on 65nm CMOS manufacturing capacity, which has left it short of GPS chips to sell.

'The key is access to 65nm process,' says van Beurden. 'We talked about it at the end of the second quarter and it has impacted the third quarter and will impact the fourth.'

At least CSR has a trick up its sleeve to get over the problem. The GPS chips are made on a mainstream CMOS process by one of the Common Platform alliance foundries, where IBM, Samsung and Globalfoundries share a basic process definition. The use of common process rules makes it easier to move chip production from fab to fab, and that is what CSR is considering doing: 'We're talking about moving [production] to another partner.'

It's a similar story with the smartphone chip issue.

'We're underexposed to the smartphone business but we have a plan in place to fix that,' says van Beurden, pointing out that the company already has design wins for a new smartphone chip with two Tier One handset makers that will have products in the market by Christmas.

As for the economy, there is little a company so reliant on consumer demand can do about that other than keep making its products more attractive and relevant to end users.

Moving forward

CSR used the results announcement to introduce a deal with wireless modem chip maker Infineon, to put CSR parts onto Infineon's cellular reference designs. This will make it easy for companies trying to get into the handset business using Infineon modems to choose CSR parts for related functions such as Bluetooth and Wi-Fi.

Van Beurden points out that similar deals to use one company's reference designs as a shop window for another company's chips have happened before, for example in last year's deal to offer Atheros's Wi-Fi chips alongside Qualcomm's Bluetooth parts.

He expects the volume business for the parts to begin in feature phones, but is okay about that: 'Infineon ships hundreds of millions of handsets a year. With a reasonable attach rate we would get to potentially tens of millions of parts a year.'

New products

In the 11 years since it was founded, CSR has created, through development and acquisition, a strong portfolio of wireless technologies, which includes Bluetooth connectivity, Wi-Fi, FM and GPS, plus the ability to integrate different combinations of these standards into single chips to suit different markets. It recently bought APT Licensing for its low-latency codec expertise, which is now being put to work in Bluetooth chips designed to minimise the lip-sync problem that can occur when watching video using Bluetooth headphones.

The company also rides the wave of evolving standards diligently. It has just announced its first chip to support the Wi-Fi Alliance's Wi-Fi Direct protocol, an evolution of the Wi-Fi standard that allows devices to connect to each other without needing a router, and also to act as hotspots, for example when sharing a 3G connection. Some handsets already do this is in a proprietary way - the new standard should make the function more widely available. The CSR9100 chip also supports two other recent evolutions in wireless connectivity: the Bluetooth low-energy protocol; and HD Voice, which some cellular operators are introducing to their networks.

Keeping abreast of standards should open up new markets, but there are no guarantees. Take wireless USB, a version of the ubiquitous computer interconnection standard that appears on billions of devices and which uses an ultrawideband (UWB) radio connection to replace the cable. The advantages of the approach were obvious and the likely opportunity vast, so venture capitalists and large semiconductor companies E F invested hundreds of millions of dollars between them to get chips ready to take advantage. CSR made sure it had an offering, too. Then, nothing.

Is wireless USB dead in the water? 'It is,' says van Beurden. 'But if it revives, we're ready.'

Quite why the standard didn't make it isn't clear. 'You look at the technology, you look at the way it would be used and it was extremely compelling, but for some reason or other it just didn't happen,' he says. Wireless USB has now been overtaken by Bluetooth Hi Speed 3.0, a combination of Bluetooth for protocol handling and Wi-Fi for volume data transfer.

'Compared to what UWB can do, it is substandard,' says van Beurden. But it may turn out to be the market's preferred solution.

CSR is now trying to address markets, including handsets, audio, location, general consumer, auto and health and fitness, with a series of 'platforms', highly integrated combinations of chip function and supporting software that enable it to satisfy as many customers as possible with the minimum number of different chips.

There's good reason for this: although some of CSR's chips are built on what might be called mainstream processes (65nm CMOS), the most demanding and highly integrated devices are being built on 40nm processes. And at 40nm, the mask sets alone cost 'millions of dollars', never mind the costs of creating the design in the first place.

CSR is trying all the tricks it knows to manage its design costs, reusing what intellectual property it can, distributing its design activity, particularly in RF, to centres of excellence in Cambridge, the south of France and Phoenix, Texas, and undertaking some software development in Bangalore. The plan is to spend 8 per cent more on R&D next year than this, but to see R&D costs as a proportion of revenue shrink to 20 per cent through growing sales volumes.


CSR has also increased its technology base through acquisitions, including of echo and noise-suppression software company Clarity Technologies, wireless design house UbiNetics, and GPS specialists NordNav and Cambridge Positioning Systems. In June last year CSR merged with SiRF Technology to further strengthen its GPS portfolio, and followed that with the APT Licensing deal this year.

Van Beurden is excited about the opportunities for streaming audio that the APT deal has opened up, especially at the top end of the market.

'Operators are very interested in the quality of phone audio,' he says. 'They don't want to be building good networks and then have the quality of their audio undermined by the link from the handset to the headset.'

He is also open to other acquisitions, but hasn't seen the right opportunity come along yet.

'It is very clear where we play and what companies are interesting or not,' he said. 'We're looking at digital radio and mobile TV from an M&A viewpoint, but haven't been excited by anything yet.'

He's more enthusiastic about NFC, the near-field radio technology currently being introduced in some low-value contactless payment systems: 'The use cases are quite compelling. I'd love to use my phone as a credit card and get rid of my wallet.'

But acquisition targets have to offer good technology, make sense financially, and be a good cultural fit with CSR: 'If you get the wrong company or the wrong culture then you can't really integrate it.'

Is this a good time to be looking for an acquisition? 'It's almost always a good time - and never,' he says. 'When we did the SiRF deal in early 2009, when things were at their blackest, people thought we were crazy.'

The deal ended up strengthening CSR's position in the GPS chip market.

Overall, though, growth by acquisition isn't an easy option: 'You have to be prepared to say no ten times before you say yes to the one time that makes sense.'

For the moment, then, innovation at CSR will have to come from the shift to 'platforms', rather than new versions of the same chips.

'With things like the high-quality audio chips, we're actively trying to influence the user experience,' as a way of creating new opportunities for the company's products.

In a similar way, CSR is working on 'deep indoor location technology', a subtle integration of data from multiple sensors (accelerometers, altimeters and the like) to extend GPS-like facilities to equipment that is out of range of satellite signals. This could open new markets for equipment that relies on knowing your location indoors.

Van Beurden can see that coupling the technology to back-end servers could offer opportunities to develop service revenues based on the capabilities of CSR chips, opening up another front for innovation.

'I can think of a model where I use that location information to offer services,' he says, 'but it is a very big decision for a company to do that. The roots of the company are in chips, software platforms and differentiation. A move into services would be a big, big decision.'

Maturing business

At 11 years old, CSR is now a mature business focused on serving its shareholders. It is even planning to pay a dividend, 'handing shareholders some of the excess cash it has generated. For a long time, technology companies preferred to believe that they were the best stewards of their shareholders' money and would find a way to make it grow for them. But attitudes have changed recently.

'Historically, companies would move into paying dividends when they were no longer in growth mode,' says van Beurden. 'We are still a growth company, in both the top line and the bottom line, but we have concluded that it would be appropriate to give some of the money back.'

The company is also doing some corporate venturing, recently making an investment in a MEMS company to gain a seat on the board from where it can observe the evolution of the market. Van Beurden says the MEMS technology could be useful for building sensors for the dead reckoning used in deep indoor location, and also for making RF oscillators.

The original premise of CSR was that, using insights that its founders gained from working at Cambridge Consultants, it would do something no one else at the time could: build single-chip Bluetooth chips in standard CMOS processes. The company succeeded spectacularly in carrying out that plan,'and has since grown and diversified into other lucrative, but deeply competitive, adjacent markets. Its growth is reflected in the number of chips it has sold. CSR reached the 'one billion chips sold' landmark in April 2008, nine years after it was founded. According to van Beurden, the company recently held a small celebration to mark selling its two-billionth chip, just a couple of years later.

Given this context, should the market have punished CSR so harshly on that Wednesday, given the signs that the company knows what its issues are and has plans in train to address them?

'The market is always right,' said van Beurden. 'It may just take its time catching up.'

Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.

Recent articles