Dialogue with Sir Robin Saxby

by Nitin Dahad

ARM has been one of the key players of the UK electronics industry for over a decade. Having started out in 1990 with just 12 engineers, Sir Robin Saxby (57), now the chairman of ARM Holdings, certainly has many stories to tell about growing an electronics company from nothing to becoming a global force in the semiconductor industry. In an exclusive dialogue with The Chilli, Sir Robin offers his thoughts and lessons for other entrepreneurs.

Sir Robin Saxby

The questions:

1. What prompted you to get into electronics, where did you study, what drove you to do what you do?

2. What were some of the early jobs you did, and what was your big break in management?

3. How did you get into ES2 and what were the challenges here as the first pan-European company and a pioneer?

4. How did you get involved with ARM?

5. Describe the scene at ARM in the early days. What were the challenges?

6. Who were the first customers, where did you make your first breakthrough, and what challenges did you face?

7. You hear about ARM 7, 9, 11 and others but don’t hear about ARM 2, 3, 4 etc. Is there any reason for this?

8. The success of ARM in licensing the whole industry with the ARM processor is well now known. What we would like to know is how did you get to charge such a high license fee, and what were the parameters that influenced it?

9. Part of the ARM success has been the vast numbers of partnerships. What makes successful partnerships and what is your advice for startups with regard to this aspect.

10. Other IP companies have tried to emulate the ARM model, but have had mixed results. What do you think differentiates good IP companies from also-rans?

11. As semiconductor IP becomes an integral part of most semiconductor SoC designs, what do you see as the opportunities for IP companies?

12. ARM sits in the middle of all the different design flows and has to work with various EDA companies. As some EDA companies have entered the IP business, how do you think the industry will shape out in the future regarding IP vs EDA design flows, especially when systems become more complex and IP is offered as a complete platform design including tools, OS and applications, as well as the hardware IP?

13. The 1984 recession was blamed for the demise of Cambridge based computer companies such as Acorn, Sinclair, and Tangerine. US based PC companies have had similar recessions (read Dell, HP/Compaq, IBM) and yet they still survived in one form or the other. What factors do you think helped them survive the droughts, while UK based companies withered away? In hindsight, what were the missing elements from the UK side?

14. As you know, a major part of the entrepreneur’s challenge is the management of different risks and crossfire. The skills of surviving the crossfire and endeavours can only come from learning, experience of others, and practice. Again in hindsight, do you think not enough managers were willing to take the risks?

15. Part of the globalisation process requires understanding, adapting, and assimilating foreign markets, cultures, business styles. Do you think that this part is widely recognised in Cambridge and the UK? What can be done to alleviate this? What advice do you have for the next generation entrepreneurs?

16. What message would you like give to engineers thinking of starting up their own venture or new startups in their early stages?

1. What prompted you to get into electronics, where did you study, what drove you to do what you do?

I was given an electronics kit as a Christmas present when I was eight years old, and soon got into radio and television. At the age of 13 I had a radio and TV repair business, coming from my passion for electronics and also for business. I went to grammar school and then studied at Liverpool University in the 1960s – an exciting place to be at the time. At university, my final year essay was on colour television. My interest has always been in life and electronics – I have followed my hobbies all my life, namely business and technology.

I believe that you should follow your passion – do what you enjoy. But it is worth remembering that it takes a lot longer to make money out of engineering; if you want to make a fast buck, then electronics and technology is probably not for you.

2. What were some of the early jobs you did, and what was your big break in management?

I joined Rank Bush Murphy in 1968, designing colour TV receivers and integrated circuits, including the SL901 and SL917 colour TV decoder chips, manufactured by Plessey. At the time these linear ICs just had 50 resistors and 50 transistors. An individual engineer could get his head around the whole design, because it wasn’t very complex. I also gave lectures in Essex and at the Royal Television Society that opened up my horizons. After four years and at the age of 25, I then joined Pye TMC to get some ‘professional experience’, where the old Post Office [and what later became British Telecom] was a key customer.

From the moment I left Rank Bush Murphy, Motorola Semiconductor kept trying to hire me and I kept on saying no. After about one year of trying, they then asked me, “What colour would you like your New Ford Cortina GXL?” I replied ‘white with black upholstery’. I ended up spending 11 years with Motorola, and that turned out to be a good decision. It was a totally different culture to the Post Office (which was British and very formal), and American semiconductor companies were investing heavily in engineering, r & d and new products and establishing a strong European capability, including manufacturing.

My first job at Motorola was that of a sales engineer, selling colour TV chips. I got a lot of design wins, even in the recession of 1974. The big break in management was being fast promoted to head of the computer division.

The downside to working in a huge corporation is how you can personally make an impact [on the corporate organisation]. So I decided to leave for personal reasons and stay in the UK. I was then head-hunted to be the Henderson Security Division CEO, a part of Henderson Garage Doors, who needed someone with an electronics background because of an earlier acquisition in electronic security systems. They said, “Our access control isn’t working”. I remember receiving a letter from the Special Branch saying, “Your electronic gates have written off three of our police cars.”

So I was running the “high-tech” end of Henderson, working with 8-bit microcontrollers. I found then that most businesses have similar features – for example the way you design and specify a security gate is similar to designing and specifying a mass programmable ROM.

3. How did you get into ES2 and what were the challenges here as the first pan-European company and a pioneer?

It was at that time that I received a call from Jean-Luc le Grand Clement who was setting up ES2 (European Silicon Structures), and I was hired as sales and marketing director. Henderson gave me two years of business experience, but mechanical engineers didn’t get paid well. In contrast, ES2 was a well-funded startup, back at the forefront of electronics technology.

While the business model and the cost of producing full custom chips using e-bean machines was prohibitive, the role at ES2 taught me a lot about America and Japan. I was frequently commuting across the Atlantic and set up our US operations and in the other direction to make inroads into Japan.

4. How did you get involved with ARM?

I received a call from a headhunter, appointed by Acorn, who were planning to create a joint venture with Apple Computers and VLSI Technology. The interview with Apple and Acorn was essentially the business plan for ARM. From my experience at ES2 and Motorola, I said we had to make ARM into a global standard and the only way we could do that was to license in parallel to many companies. I also said “We’ll make chips over my dead body,” a reference to the struggle and cost issues we had at ES2 in manufacturing.

5. Describe the scene at ARM in the early days. What were the challenges?

When I joined ARM it was just 12 engineers (spun out of Acorn) and me, with £1.5m intellectual property from Acorn Computers, £1.5m cash from Apple, and £250k cash from VLSI Technology. With the early team (which included Mike Muller, Tudor Brown and Jamie Urquhart) we carried out a SWOT analysis which identified our opportunities in portables, embedded control and automotive.

The first challenge was a friend telling me that joint ventures don’t work. Each of ARM’s joint venture partners had a different agenda and objectives – Acorn was interested in getting the chips at a good price for the Archimedes Computers, Apple was interested in using the technology for the Newton, and VLSI wanted to make the chip. The personal and business dynamics were challenging for a chief executive.

The first year difficulty was that we nearly ran out of money. I remember squeezing salaries and the engineers foregoing salary rises. In the early days we were very lean and mean and clocking up a lot of air miles in economy class. The startup phase was very tough, and getting a credit line was tough. Eventually an old bank manager contact at Barclays gave me a credit line.

The engineers were focused on one major job – to get the first chip out and working for Apple (the ARM610). They also had to do quite a lot for Acorn, and then we had to get a business going. I felt many times that I was like Samson trying to hold up the world in every area except engineering as the engineers were very hard working and overstretched. I did get a lot of help, from, Jamie, Mike and Tudor who were very fast learners.

The focus and setting priorities were important. Sales were also important, so we hired Tim O’Donnell working out of his basement in California. I believe that to be truly successful in technology, you need to think about how you can be global. At that time, the USA was important, and now India and China are equally important. We set out with a clear plan for setting up global operations.

We often punched above our weight and also insisted on up-front payment on license fees to help cash flow. In year one we were 20 people, but on the back of the order from Samsung that figure went up to 60 people.

6. Who were the first customers, where did you make your first breakthrough, and what challenges did you face?

The first customers were our partners Apple, VLSI and Acorn. Then came the handshake agreement with Plessey in 1991 where we built working chips for the company on risk, and then the deal with Sharp in 1992. The breakthrough agreements that really put ARM on the map were the TI 1993 agreement followed by Samsung in 1994. It was after this that other partners noticed us and started coming to us.

We probably could have done a deal with LSI Logic earlier but waited as it could have diluted our efforts with VLSI Technology [now owned by Philips Semiconductor].

The first end-equipment customer drivers were Nintendo and Nokia – the former via Sharp, and the latter via TI. Nokia was pushing code density, so that’s where ARM’s ‘Thumb’ came from. Our architecture evolved on the roadmap of our customers.

7. You hear about ARM 7, 9, 11 and others but don’t hear about ARM 2, 3, 4 etc. Is there any reason for this?

WHistorically, Acorn had released ARM 1, 2 (Archimedes) and 3. We started with ARM 6 as the first embeddable core which Apple’s Newton used. ARM 8 was the pre-cursor to ARM 9 and then Digital Equipment [semiconductor division, later acquired by Intel] came along with StrongARM. ARM 7, 9 and 11 were part of the wireless roadmap. We are now using names rather than numbers.

8. The success of ARM in licensing the whole industry with the ARM processor is well known now. What we would like to know is how did you get to charge such a high license fee, and what were the parameters that influenced it?

People pay you if you provide them with value – they pay what you think it’s worth. In IP licensing, you have to spend a lot of money in research and development. We have created a demand for our technology, and you need that kind of license fee to cover the amount of r & d that goes into it. In fact customers are actually getting a bargain because of the all tools, and the bigger software challenge [addressed by a whole community of ARM partners].

9. Part of the ARM success has been the vast numbers of partnerships. What makes successful partnerships and what is your advice for startups with regard to this aspect.

You have to create a win-win situation with partners. We believe we understand the ARM architecture and software tools. What we don’t know as much about is specific nuances of wireless, networking, consumer or automotive chips for example, which is what over 135 ARM semiconductor partners bring to our community.

Honesty and pragmatism is important in making a partnership work – and you also need an awful lot of communications. Pushing the boundaries [of the relationship] is also essential – if you’re not causing pain in the partnership, then it might not work. Too many people use the term partnership too glibly. A real partnership is pain and gain on both sides.

Startups also need to have the best people, so it’s also good to use subcontractors until you can hire your own.

10. Other IP companies have tried to emulate the ARM model, but have had mixed results. What do you think differentiates good IP companies from also-rans?

Artisan and Rambus are examples of other IP companies that have done reasonably well. The key to success is to develop strategies and plans to sustain the business model. To be a successful company, you need a long-term strategy and vision. Other companies [that have not done so well] may not have looked at all at such aspects and have run hand to mouth from one challenge to the next.

11. As semiconductor IP becomes an integral part of most semiconductor SoC designs, what do you see as the opportunities for IP companies?

I would say try to address needs that established players are not addressing. IP is in its early stages, and there is a discontinuity in the 10nm to 65nm range so there are opportunities in that area. There are enormous challenges in every domain of software, from design tools to applications.

12. ARM sits in the middle of all the different design flows and has to work with various EDA companies. As some EDA companies have entered the IP business, how do you think the industry will shape out in the future regarding IP vs EDA design flows, especially when systems become more complex and IP is offered as a complete platform design including tools, OS and applications software, as well as some of the hardware IP?

A dramatic business model change is occurring in the EDA industry. If you’re in the number of design ‘seats’ business [where EDA vendors charge fees according to numbers of design engineers using the EDA tools] it’s becoming a real challenge, but if the EDA company helps their end customer create products quickly and efficiently, then that is more likely to have success. Newer emerging companies are appearing in this space and have opportunities in this area. Established players are evolving their business models. In addressing new opportunities, the board of any company must look at what they do best rather than get into non-core activities.

Different customers use different EDA tools, but we’re not religious about design methodology – we work with everybody, we go with the flow. We invest some of our money in the tools space when necessary.

There is a lot of innovation going on in design methodology as companies require fast time to market but face higher cost of design and manufacturing.

13. The 1984 recession was blamed for the demise of Cambridge based computer companies such as Acorn, Sinclair, and Tangerine. US based PC companies have had similar recessions (read Dell, HP/Compaq, Apple) and yet they still survived in one form or the other. What factors do you think helped them survive the droughts, while UK based companies withered away? In hindsight, what were the missing elements from the UK side?

The industry recessions come in waves, and I don’t see it as UK companies or US companies – all companies now need to operate globally. You start the company somewhere, but the UK needs to create global players.

What we need to do is:
- connect the business world with the technology world
- scale all engineering resources
- be connected to the best universities
- create centres of excellence from the design point of view

The UK doesn’t have a low-cost economy so we need to be more efficient and weigh up the cost of capital and cost of people tradeoffs.

I am optimistic, because we live in a global ecosystem. If government could get more students doing maths and science at schools level that would also help.

14. As you know, a major part of the entrepreneur’s challenge is the management of different risks and crossfire. The skills of surviving the crossfire and endeavours can only come from learning, experience of others, and practice. Again in hindsight, do you think not enough managers were willing to take the risks?

People should do what they are passionate about. You can’t just take great technology and make it into something hoping someone will buy it. Useable great technology comes from satisfying customer explicit needs. In the UK, the biggest gap is in commercial skills – we need to teach sales and marketing skills to the engineers. Otherwise the environment for the technology entrepreneur is good.

In addition, if you want to develop people you need to give them experiences. That’s what young people need – the opportunities for development.

15. Part of the globalisation process requires understanding, adapting, and assimilating foreign markets, cultures, business styles. Do you think that this part is widely recognised in Cambridge and the UK? What can be done to alleviate this? What advice do you have for the next generation entrepreneurs?

Appreciating cultural differences is extremely important, which is why you need to hire local leaders – eg. in China and India. Cambridge has a good entrepreneurial environment and understands some of these issues, but overall the UK needs to spend more effort in this area [to understand these challenges].

With regard to academics, some wouldn’t touch business with a bargepole – so don’t make people do what they don’t want to; it’s a lot of hard work. Succession planning is also important for the high-tech business – I feel good about the team running ARM and the potential and skill of our younger people.

Lastly, people in a startup who are battle-scarred have more chance of success – look at Henderson and ES2 in my case. One of our challenges in the UK is you need global commercial experience, which many UK startups don’t have – but for example startups in California understand. Pushing the technology is one thing, but the pull from customers is actually more important.

16. What message would you like give to engineers thinking of starting up their own venture or new startups in their early stages?

Overall, the UK has the elements of a great entrepreneurial environment, as a result of some of the work going on in the academic and laboratory environments. But when you start a technology company, you shouldn't just focus on the technology. There are many examples of companies around the world that have not been able to exploit their commercial potential. We should also recognise that some failure is part of the learning process. Neither ES2 nor Acorn made it on the “world stage”, but the ARM founders’ experience in those companies lead on to ARM being a success.

I would especially encourage the management of startups and emerging businesses to make great effort in finding and hiring very competent globally experienced sales and marketing people. If they do this and plan with a long-term vision and strategy, then there is no reason why companies originating in the UK or any other country cannot compete with the more established global players.

Comments on this story? Send an e-mail to Editor@TheChilli.com

© Chilli Publishing Ltd 2004

16DEC2004