In the microcontroller sector it's no longer a matter of providing the silicon: support around the chip can be at least as important as its actual performance. How do you ensure that your evaluation kits, sales support and development tools match the quality of the devices themselves?
Steve Laub: We have always had a high focus on development tools and design flow to inspire and make it easy for our customers to design our chips into their products. This is one of the major reasons why we have success with our 8- and 32bit microcontrollers. We want the customer to experience ease of use and efficiency, whether they buy the least expensive starter kits or the most feature rich in-circuit emulators. For AVR microcontrollers, all development tools are hosted by AVR Studio, which is a high-end debugger provided to our customers free of charge. More than 300,000 users download this software every year, and we sold more then 100,000 development kits in 2008. In addition, we support and highly appreciate the number of third-party suppliers that complete our product range by developing and offering a wide range of kits and software to the market. There are more then 700 different kits and close to 20 different C-compilers listed on the community site AVRFREAKS, a Web site dedicated to the AVR. To ensure the quality of these third-party tools, we are offering design kits for the suppliers to integrate their solutions into the AVR Studio front end. The feedback that we get from customers is very positive regarding the tools' performance and ease of use. They also report that they appreciate that the tools can be used across our entire AVR product line so that they do not have to buy new tools when they do a new design with any of our products.
Atmel recently announced that it was looking at whether it had a serious future within the ASIC market, and was considering the possibility of selling its ASIC business unit. Although the demand for standard-cell ICs is declining, aren't there still useful synergies to your existing smart-card product offerings?
S. L.: Over the last two years, Atmel has been transitioning into a company offering microcontroller-based solutions as its core business. Today we are one of the leaders and the fastest growing providers of 8 and 32bit controllers, and the preferred choice for many customers due to our innovative solutions. We are engaged in a number of rapidly growing microcontroller segments and see a need to strengthen our focus even more to deliver innovative technology that our customers appreciate. While the ASIC business is a market leader in many business segments, it is not sufficiently synergistic to our core microcontroller business. This is why we are pursuing strategic alternatives for this business unit, including a potential sale. The smartcard business utilises our microcontroller core in its product solutions; however, there is little customer overlap with the standard microcontroller business. Therefore, we believe that these customers are best served by the ASIC business unit.
Likewise, is it fair to say that your programmable-logic and EEPROM portfolios won't be added to in the future, but that you will just support existing devices?
S. L.: While our core business is microcontroller-based solutions, there are additional products and solutions that are combined with the microcontroller, such as serial EEPROMs that share both technology and customers. Atmel is the market leader in serial EEPROMs and plans to stay in the serial-memory business as well as introduce new products. The serial-EEPROM market was over $800 million in 2007, and Atmel's market share was nearly 40%.
You acquired UK firm Quantum Research last year in order to benefit from its QTouch capacitive-sensor technology. How much traction are you now seeing in this market, and are there opportunities to tie it in with your MCU offerings?
S. L.: This strategic acquisition was consistent with our strategy to build MCU-based solutions addressing high-growth markets. We began shipping touch-technology products several years ago in partnership with Quantum. This technology is now being fully integrated into our microcontroller line up in order to optimise solutions for buttons, sliders, wheels and multi-touch high-resolution screens. We are very excited about the market opportunities that this technology, combined with our microcontrollers, brings to Atmel. Two out of the top four cell-phone manufacturers have chosen our QTouch technology for their touchscreens. Some industry followers foresee a 6-billion-unit TAM (total available market) for buttons and sliders as more and more designers see the benefits of capacitive touch technology. For Atmel, the combination of our high-performance, low-power microcontrollers and our efficient capacitive-sensing algorithms allows us to provide the marketplace with the industry's most robust solutions. In 2008, Atmel experienced very strong growth in microcontroller-based touch solutions, and we expect that this will be our fastest growing end market over the next several years.
Atmel has a fairly balanced market coverage, with business in the computing, industry, security/identification, consumer and automotive sectors. Are there any particular areas that you are still looking to make further penetration into, or where you see greater opportunities opening up?
S. L.: In addition to touch sensing, we believe that wireless provides very exciting opportunities. AVR is the processor of choice of the ZigBee Alliance, and our recent acquisition of Meshnetics IP means that we can now offer complete AVR and ARM solutions to cover wireless applications from 700MHz to 2.4GHz. Metering and security are just some of the application areas that are growing fast, even in the current economic environment. In addition, the ongoing drive for power reduction is creating new customer opportunities in almost every hand-held product for controllers that include our picoPower technology. This reduces power consumption in both active and sleep modes, thereby lowering overall power requirements and extending the operating time between battery charges. We are also seeing our XMEGA 8/16bit parts displacing competitors' 16 and 32bit devices in some of the higher end processing applications, where designers are looking for power and cost savings.
Atmel was badly hit during the 2000/2001 downturn. How will you go about ensuring that the firm doesn't suffer as badly this time around?
S. L.: The entire semiconductor industry was affected by the 2000/2001 downturn, and Atmel was not immune to the adverse business conditions during that period. However, our microcontroller business grew 50% in this period, and we increased our market share significantly. The biggest change for Atmel over the last few years has been the transition from a commodity memory supplier to a higher margin, proprietary-product supplier providing sophisticated standard products as well as system solutions to our customers. Another key difference is that Atmel began to substantially restructure its business prior to the current downturn. During the last two years, we have closed or divested fourteen different product lines, reduced headcount by over 25%, and saved over $125 million in annual cost. For calendar year 2008, Atmel generated the highest gross margins since 2000. Atmel's performance during this downturn has been relatively better than that of many of our competitors [I believe this is] due to the innovations we have introduced to the marketplace and the proactive actions that we have taken to optimise our cost structure.
A little over a year ago, your competitor Microchip introduced a 8/16/32bit microcontroller platform that allows engineers to easily move their design to a higher performance level if required, or alternatively downgrade to a lower end MCU in order to save money or power. Do you expect to implement a similar platform for your AVR products?
S. L.: Our customers have had this ability for several years. Since 1997, when the first AVR device was launched, they have been able to use a single toolset to develop products using products ranging from our small tinyAVR devices up to our larger, higher performance devices. For example, today they can use the AVRONE emulator to carry out a variety of tasks that range from debuging an 8pin tinyAVR device with a single-wire debug interface to performing real-time emulation of a high-performance AVR32 device.
Who are your major distribution partners here in Europe?
S. L.: Europe is a major market and represented approximately 38% of Atmel's 2008 revenue. Our customers are served by a direct sales force as well as an extensive distribution network. We have two pan-European distributors in Arrow and EBV, supported by a strong network of regional distribution partnerships such as MSC, Ineltek, Silverstar and NuHorizons.
As you are covering such a large number of different product areas, is it hard to make sure that your research, sales/marketing, and other resources are adequately allotted to each of them?
S. L.: Every large company with multiple businesses and product lines confronts the decision as to how best to allocate resources to maximise returns. At Atmel, we allocate resources based on each business unit's and product line's expected return on investment based on objective metrics and adjusting for both market and execution risks. As we are focusing the company more and more towards microcontroller-based solutions, we also allocate resources biased to that business, reflecting the strategic value of those investments. Furthermore, we are spending a significant amount of money in those key technologies to ensure innovative solutions that inspire our customers to partner with us. We believe in the long term: to win requires constant innovation and differentiated performance.
Headquarters: San Jose, California, USA
CEO: Steve Laub
Business: Microcontrollers and microcontroller-based solutions
Revenue: $1.6 billion (FY 2008)
source: EPN 5/09