Research report: October 2010

Chips with everything

Lessons for effective government support for clusters
from the South West semiconductor industry
Louise Marston, Shantha Shanmugalingam and Stian Westlake
Executive summary

Modern life pulses to the rhythm set by by intelligent, targeted support from local
semiconductors. They are in a staggering array universities.
of products, from mobile phones and laptops,
to traffic lights and trainers. Their ubiquitous Many challenges facing the semiconductor
nature makes them big business. industry reflect problems common to many
industries: difficulties in recruiting good quality
The UK has a vibrant semiconductor graduates (currently only one in five electrical
industry. UK companies compete in global engineering post-graduates are from the UK),
semiconductor markets by combining adequate access to finance for start-ups and
technological innovation with business model creating incentives for growing large companies
innovation. In doing so, these companies in the UK. The industry is not waiting on
make an important contribution towards government to deliver all the answers, and
the UK’s economic success, with the UK has created a number of skills and research
semiconductor market worth $6 billion in 2009 initiatives through bodies such as Silicon
– third in Europe, behind France and Germany. South West and the National Microelectronics
Equally the industry has an important impact Institute, as well as establishing the European
at the regional level, and the South West Micro-Electronics Academy to try to
semiconductor cluster is one of the largest reduce costs for start-ups and inspire new
silicon design clusters in the world. entrepreneurs.

NESTA’s ‘Rebalancing Act’ highlights how the But there are some issues which fall to
continuing success of innovative companies will government to resolve. Government has
be important to the UK economy. A starting signalled its intentions to address start-up and
point for government in considering how to growth barriers through reviews on access to
support innovative, high-tech companies finance and taxation of Intellectual Property.
and regional clusters is to understand the With the university system in flux, the industry
challenges facing these industries. This report is keen to see how government seeks to
examines the semiconductor industry in the improve the quality of graduates.
UK, particularly in the South West, to identify
how this cluster has grown over the past 30 Finally, with government considering how
years and how government, with very limited to help existing regional clusters, it is worth
resources, can effectively support regional drawing out two important lessons from the
clusters. emergence of the South West semiconductor
cluster. Both are relatively low cost and seek
The South West cluster has grown because of to work with private sector networks, rather
a combination of factors. Large semiconductor than replacing them with public sector delivery
companies present in the region in the 1970s agents.
and 1980s, such as Inmos and Plessey, have
endowed the region with a highly skilled • First, investment in networks, whether
workforce and contributed to attracting university or private sector-led, can help
entrepreneurs, companies and investors to grow clusters by introducing customers and
the region. This has been supplemented

3
 suppliers, providing mentoring and sharing
resources.

• Second, incubation models, embodied by

the SETsquared partnership in the South
West, can work to promote growth of new
companies in these regions and highlights
how universities can effectively pool their
resources to support these companies.

The ongoing success of these interventions

is dependent on sustaining the small base
funding that supports such networks and
incubators, separate from research or company
funding. It provides early-stage funding
from which private sector funds can be
leveraged. This approach will not be suitable
to all industries or regions but offers an
example of how a blended approach can help
private sector investment in innovation and
consequently contribute to economic growth in
the region.

4
Contents

Chips with everything

Lessons for effective government support for clusters from the South West semiconductor industry

Part 1: Introduction 6

Part 2: The South West semiconductor cluster 9

Part 3: Challenges for the semiconductor industry 17

Part 4: Broader implications for government clusters policy 22

Part 5: Conclusions 26

5
 Part 1: Introduction

Semiconductors are an invisible part of in semiconductor technology. The cost of a

1. William Shockley moved to the fabric of modern life transistor has fallen, on average, by 35 per
Mountain View, California in
1956 and founded Shockley cent every year over the past 50 years (when
Semiconductor Laboratory, Semiconductors are ubiquitous in modern adjusted for inflation).3 Moore’s Law (taken
whose alumni went on to
found Fairchild Semiconductor life, encountered in a huge array of products: from a statement made by Gordon E. Moore4
and Intel. from the alarm clock, to the fridge, the family in 1965) states that the number of transistors
2. Fairchild Semiconductor is
commonly referred to as the
car (a recent model will have 50 processors), that can fit into an integrated circuit at
first venture-backed start-up, the traffic lights on the way to work – and of minimum cost doubles every two years. This
first funded in 1959.
course, every computer and mobile phone. predicts exponential growth for the number
3. Mentor Graphics, adjusted
for inflation – presentation Their presence in the very fabric of modern of transistors in a circuit, with knock-on
to Semicon-West, July 2009. life has only come about due to remarkable implications for the increasing speed and power
Data from Mentor Graphics,
SIA, Woodrow Federal Reserve advances in science, technology, software and of those circuits, and decreasing costs. Moore’s
Bank. Available at: http:// manufacturing. Law is part prediction, and partly an expected
www.semiconwest.org/cms/
groups/public/documents/ innovation target for the industry to meet.
web_content/ctr_030755.pdf The semiconductor industry has been
4. Gordon E. Moore is the co-
founder of Intel Corporation,
intimately connected to the story of It has been speculated that the limits of this
and was a founding member technological progress over the past half a law will be reached, as the transistors approach
of Fairchild Semiconductor.
century. The early years of the industry in atomic levels, but a great deal of research
5. Future Horizons forecast,
2010. California are credited with the birth of Silicon is underway into technologies that would
Valley,1 and saw the first venture capital take Moore’s Law beyond its current limits. A
investments.2 In the past couple of decades, common way of categorising research in the
virtually every aspect of technological progress sector is to divide it into ‘More Moore’: scaling
– including the spread of personal computers, processes and components to ever smaller
the internet, digital cameras and mobile levels; and ‘More than Moore’, focused on
phones – has been driven by innovation and integrating existing components in new ways
competition in the semiconductor industry. with such areas as human biomedical sensors
and photovoltaics. While the end of Moore’s
As semiconductors become an essential part law has been pronounced many times, it has
of an ever wider range of goods, standard remained true for decades, and in a testament
designs in areas like memory chips have to innovation within the industry, with each
emerged to serve large parts of the market, and new obstacle, a new technical solution has
this part of the industry behaves somewhat been found.
like a commodity industry, with fierce price
competition and a focus on process innovation.
At the other end of the industry, demanding
customers such as smartphone manufacturers The ubiquity of semiconductors makes
drive constant product innovation, and them big business
pressure to incorporate the latest technology
advances. The global market for semiconductors had sales
of $226 billion in 2009 with 28-30 per cent
Future technological and economic progress growth forecast for 2010.5 The semiconductor
also appears to be dependent on improvements industry is almost unique in its rate of unit

6
 What is a semiconductor?

A semiconductor is a material that has an semiconductors, and a very high level of

electrical conductivity between that of a chemical purity is required to avoid defects.
conductor and an insulator. What makes
them so useful is that the behaviour of In most mass production processes, silicon
the material can be easily manipulated. crystals are grown as cylinders between
The pure silicon crystals are ‘doped’ with 100mm and 300mm in diameter, and these
impurities. Then by introducing an electric are sliced into very thin wafers. Wafers are
field, a current can flow. Semiconductors usually manufactured in areas with very low
can also react to exposure to light, pressure energy costs, because of the large energy
and heat. These properties allow current to consumption needed to work with molten
flow only under certain conditions, creating silicon.6 Large amounts of pure water are
the capacity for a switch that acts as a logic also needed. The wafers are then shipped
gate or for a sensor. Electronic components to semiconductor manufacturers for
made from semiconductor materials processing into chips. These do not need
include transistors, solar cells, many types to be located close by, because shipping
of diode, including light-emitting diodes the thin wafers is relatively cheap. They are
(LEDs) and digital and analogue integrated likely to be located in places with skilled
circuits (ICs). Silicon is used to create most but relatively low cost labour.

6. See http://www.solarbuzz.
com/Plants.htm
7. Walden Rhines, Mentor
growth, shipping almost 50 per cent more A landmark event occurred in 1987 when Graphics. Available at:
http://www.eetimes.com/
transistors and 11 per cent more chips year on Taiwan Semiconductor Manufacturing electronics-news/4088277/
year for the last ten years. This compares to Company (TSMC) was spun out of a public Viewpoint-Is-semiconductor-
industry-consolidation-
0.1 per cent annual growth for the automotive research institute to become the world’s first inevitable-
sector and 9.3 per cent for computers.7 semiconductor ‘foundry’, specialising only 8. McKinsey & Company
Underlying this growth trend, the industry is in the manufacture of other company’s chip (2007) ‘Creating Value in the
Semiconductor Industry.’ 2007
highly cyclical, with boom and bust phases. The designs. This consolidation and relentless focus Working Paper.
top companies in the industry have changed on improving manufacturing processes proved
places as the industry has evolved, and new successful, and opened the way for design-only
waves of innovation appear. The industry now companies, who could now easily contract the
seems to have entered a more stable phase: manufacture of their products out to foundries.
Intel has dominated the industry for some
years, and creates significantly more value than This business model is known as a ‘fabless’
others in the sector.8 semiconductor company of which Qualcomm
is a good example. With a foundry company
The industry is global in nature – the supply to take care of the manufacturing, it was
chain is vast, spanning many different possible to set up a semiconductor company
countries and companies, and manufacturing that concentrated on design, outsourcing
occurs mainly in South East Asia, with design manufacturing and no longer needed to own
activities spread across the globe, including fabrication plants (fabs) for their chips. The
the UK. Vertical integration from the raw general trend towards disaggregating research
wafer through to packaged and tested chips, and development, as with other industries,
incorporating design, manufacturing and can benefit small start-ups. It is cheaper and
testing was once the norm. A few Integrated easier to start a company without the need to
Device Manufacturers (IDMs) still exist – Intel manufacture, and small companies are able to
still operates as an IDM – but in a highly supply the innovative ideas that are in demand.
competitive industry, the advantages and
flexibility from breaking this value chain into This disaggregation of the value chain into
separate specialist companies has helped to specialist companies has continued. Rapid
drive improvements, and has changed the unit growth and intense competition means
nature of the industry. that every improvement is valuable, and drives
innovation and specialisation in the sector.
Some, such as the UK’s ARM, specialise only

7
 in design, licensing IP to other companies In 2009, the UK semiconductor market was
who then manufacture chips with the design worth $6.0 billion USD, the European market
company collecting royalties. Other new $29.9 billion USD and the global market
companies, especially in South-East Asia, $226.3 billion USD.13 The UK hosts the
specialise in testing and packaging the operations of more than 500 semiconductor
manufactured output from foundries and other firms, 80 per cent of which are foreign-owned.
plants. Many integrated device manufacturers These employ around 8,000 engineering
have become ‘fab-lite’ by continuing to staff.14 The firms are concentrated in clusters
manufacture some products in-house, while in Cambridge, the South West, Thames Valley
working with partners to produce more and ‘Silicon Glen’ in Scotland. The South West
specialist or advanced products. In addition, has a real international reputation for expertise
the trend towards open innovation across in design and communication technologies,
the technology sector has made firms more as well as perhaps the most vibrant start-up
receptive to the ideas and expertise offered by community. It also forms part of the wider
smaller companies. UK electronics industry worth £23 billion and
employing 250,000 people both directly and
The industry trends toward specialisation in the indirectly.15
different stages of the process, as they become
more complex and expensive, and consolidation
across each stage, to generate economies of
scale that allow continued investment in plant How can government policy help high-
and tools. tech semiconductor companies?

NESTA’s ‘Rebalancing Act’ highlighted how

future growth could be spurred by high-
9. Reference: NESTA
interviews.
UK semiconductor companies innovate tech companies, such as semiconductor
10. Future Horizons – to stay ahead of the game enterprises.16 However the role of government
semiconductor consumption in developing high-tech sectors is patchy if
figure.
The UK has had a long presence in the not poor. Striking a balance between active
11. Department for Business,
Innovation and Skills (2009) semiconductor industry, beginning with investment in certain companies (‘picking
‘Advanced Manufacturing defence and electronics companies such winners’) and promoting the right policy
Strategy.’ London:
Department for Business, as Plessey, and including names such as framework for high-tech sectors is a difficult
Innovation and Skills. Inmos, Acorn, Imagination Technologies and one.
12. See ‘Key Industries’ on
the South West England Cambridge Silicon Radio.
Regional Development In this report, we examine the growth of the
Agency website.
Available at: http://www. The UK is a genuine centre of expertise for semiconductor industry in the UK, particularly
southwestengland.co.uk/ analogue and mixed signal capabilities, that in the South West, to identify how this cluster
key_industries/ict.aspx
are needed for radio functionality, including has grown over the past 30 years and the
13. Future Horizons.
14. UKTI Semiconductor Bluetooth and Wi-fi as well as fibre optic role of government in fostering this growth.
Design fact sheet, Aug technologies. As the industry evolved, This study on the South West semiconductor
2010 http://www.ukti.gov.
uk/uktihome/aboutukti/ companies in the UK were quick to seize on cluster examines how the cluster has grown,
localisation/116366.html opportunities offered by foundries, and moved the challenges it faces and how policy impedes
15. NMI Yearbook 2009-10, up the value chain to focus on high-value or facilitates the growth of these companies.
http://www.nmi.org.uk/
assets/files/annual-reports/ design. UK companies also have a reputation As part of this project, NESTA interviewed
NMI%20Yearbook%202009- for being better at constructing products that founders and managers of companies at
10%201_1.pdf
16. NESTA (2010) ‘Rebalancing
work as part of an integrated system, rather different stages of growth to understand what
Act.’ London: NESTA. than isolated chip designs, an increasingly drove them to establish their companies, locate
important skill.9 in the South West and how they have grown
and expanded.
The UK has a small share of the global
semiconductor market (around 2.7 per cent,
representing 20 per cent of the European
market),10 but it is the third biggest player in
Europe (after Germany and France) creating
high value jobs and significant wealth.11 UK
companies have developed good market share
in specialist areas, like design, with the largest
concentration of semiconductor designers in
Europe based in the South West.12

8
Part 2: The South West semiconductor cluster

The semiconductor design industry forms an engineers, as well as having semi-skilled

important part of the electronics industry, and roles to boost female employment. This led
the South West is its hub in the UK. The cluster to the location of Plessey Semiconductors to
contains around 50 companies and directly Swindon, providing training for engineers in
employs in the region of 5,000 people.17 The a wide range of semiconductor skills. At the
region around Bristol and Bath is home to one time: “Pressed Steel offered more pay, but
of the biggest silicon design clusters anywhere apprenticeships at Plessey were reckoned to
in the world outside Silicon Valley.18 It is be better because of the variety of skills that 17. Silicon South West (2009).
18. See http://www.
double the size of its nearest UK competitor, were learned“.19 southwestengland.co.uk/
Cambridge. In the last decade, start-ups in key_industries/ict.aspx
the South West have attracted more than • The universities in Bath and Bristol and easy 19. See http://www.
swindonweb.com/?m=8&s=
$550 million in investment and returned access to Heathrow airport for flights to the 116&ss=396&c=1327&t=Ple
more than $800 million to shareholders. This United States also contributed to location ssey+is+back...

section explores the birth and growth of the decisions of semiconductor companies.
semiconductor industry in the South West.
These factors led to the location of both
Plessey Semiconductors and Inmos in the South
West, in Swindon and Bristol respectively.
Historical background These have been the two most influential
companies in the history of the cluster, creating
As with all clusters, the foundations of the a legacy of skilled people that have remained
South West cluster can be traced to a few key in the area.
decisions and factors, including:

• The decision of Silicon Valley company

Fairchild Semiconductor to locate a design Growth of the cluster
office in Bristol in 1972. Ironically, the
location was chosen partly because there Growth of the cluster since these early
was no cluster present: Fairchild reasoned pioneer companies can be traced back to the
that it could retain engineers better by convergence of several factors, including:
opening offices in locations where there were
few similar companies around. One of the • A highly skilled work force with excellent
original office members would go on to work sector expertise: The South West has a
for Inmos, a government-backed company wealth of talented, highly skilled electronics
founded in 1978 which became a critical engineers. In part, this has been a factor of
training ground for many engineers in the the rise and fall of the two major players,
area. Plessey and Inmos. Those who moved to
the area were frequently reluctant to leave,
• The recognition by councillors in Swindon perhaps reflecting the high quality of life and
that, with the railway industry in decline, attractive surroundings of the South West.
other industries would be needed in the
area, especially those that would employ

9
 Plessey, Inmos and the South West

Plessey was a key driving force for company was divested and the remnants
electronics in the UK from the earliest days of the company were re-acquired and re-
of the industry, originally making radio sets named Plessey Semiconductors in 2010,
for Marconi, and later supplying the Post and the new company aims to make use of
Office and defence communications. Its the Plymouth facility again.
electronics division was formed in 1951,
and the first silicon plant in Europe was Inmos was founded in 1978 by three
opened in 1952. Plessey Semiconductors entrepreneurs from the UK and US and
opened its plant in Swindon in 1957 (the funded with an initial investment of £25
same year that Fairchild Semiconductor, million from the UK government as a
the original Silicon Valley company, was national champion for microelectronics
started in California). Plessey became the through the National Enterprise Board
largest employer in Swindon by 1960, (NEB). The UK company aimed to develop
replacing jobs lost with the decline of the a new generation of microprocessors, which
rail industry. As a large, vertically integrated put the memory on the same chip as the
company, Plessey trained hundreds of processors – dubbed the ‘transputer’.
engineers, taking in many new graduates
and apprentices. It had a small plant in The company was established in Bristol,
Swindon, and in 1985 opened a leading- with additional design expertise in a
edge plant near Plymouth, creating a Colorado office, while a manufacturing
cluster of skills in silicon design and CMOS plant (‘fab’) was constructed in Newport
manufacturing there. in South Wales. The company received a
total of £211 million from the government
In 1989, Plessey was acquired by GEC over six years, and was sold in 1984 to
and Siemens, with GEC taking the Thorn EMI for £192 million, before it had
semiconductor business and Siemens the become profitable. While the impact of
research centre. The resulting GEC Plessey direct government funding is debatable,
Semiconductors (GPS) continued the Inmos has undoubtedly proved to be a
company, but lacked investment, and the fertile training ground for hundreds of
plant technology gradually fell behind the engineers in the cluster, with many former
rest of the world, where places like Taiwan Inmos employees remaining in the South
were making significant investments. West and sustaining the cluster. Spin-offs
Although Plessey did not generate any and companies founded by Inmos alumni
notable spin-offs of its own, it played a include Meiko, Division, PixelFusion,
critical role in training a generation of XMOS and Motion Media Technology.
designers and engineers who went on to Inmos was eventually absorbed into ST
work for other companies in the South Microelectronics in 1994, with the brand
West. After a series of acquisitions, the name being discontinued.

• The rise of the semiconductor serial • Broader network support: Private sector-led
entrepreneur: A virtuous circle has been networks have developed organically and
created by entrepreneurs who have a facilitate the exchange of ideas, and act as
successful company exit and then start a focal point of the semiconductor industry.
again, often investing their own cash, in new This is reinforced with farsighted local
semiconductor companies. universities, and an innovative approach to
incubation of the start-ups.
• Financing new start-ups: Successful
entrepreneurs attract venture capital, and These factors are explored in more detail in the
vice versa. The South West now is a magnet following section.
for venture capital funds seeking to find
start-ups.

10
A highly skilled work force with The rise of the semiconductor serial
excellent sector expertise entrepreneur

The South West now has the dual advantage One of the main residual effects of Inmos,
of a supportive ecosystem and a highly Plessey and other companies is the creation
skilled workforce with good semiconductor of a new set of entrepreneurs. For example,
experience. Critically this provides a ready Hossein Yassaie established Imagination
stream of both semiconductor entrepreneurs Technologies, a mobile graphics and
and a locally based, well trained workforce for microprocessor chip company, after initially
both start-ups and corporations seeking to moving to the area to join Inmos in the late
locate or expand in the region. 1970s.

This workforce comes partly from the local Equally other companies have spurred
universities, although many recruit globally. semiconductor entrepreneurs who have
However, just as in Silicon Valley, companies established companies in the cluster. Stan
within the area are also important sources of Boland, Peter Claydon and Gary Steele were all
trained employees. In particular Plessey and attracted to work at companies with no direct
Inmos are cited as the source of the South links to Inmos and Plessey. With the exception
West’s highly skilled workforce and served to of Simon Knowles, who worked for Inmos and
seed the area with sector expertise. then STM before co-founding Element 14, an 20. Why Peter Claydon set
up picoChip: “I don’t
independent fabless chip company, with Stan want money, but I want
Plessey was attracted to the area, partly by Boland, none of these new companies had commercial success because
that’s how an engineer
Swindon Town council who were well aware direct links to Inmos or Plessey. measures success. I tell
that the railways were in decline and they the VCs I want money,
because that’s what they
needed to attract new employers to the area. Boland, Claydon and Steele are all examples want to hear, but it’s not
of the South West’s serial entrepreneurs, money, it’s the success of
the company in commercial
For Inmos, easy access to trans-Atlantic flights using management expertise gained in their terms that I want.” Source:
from Heathrow, in order to visit customers starting companies to found new ones. Each Electronics Weekly 27
November 2009, http://
and the company’s Colorado Springs research entrepreneur keeps coming back for more, but www.electronicsweekly.
facility, was an important factor. Inmos each is motivated in different ways. com/Articles/2010/06/01/
47541/picochip-loses-its-
may also have been attracted in part by the soul.htm
presence of the Fairchild Semiconductor design All these founders had many reasons for
office in Bristol, which was established in 1972. establishing their companies in the South
The Design Manager at Inmos, Peter Cavill, West. As highlighted above, a highly skilled
was recruited from Fairchild, where he was workforce is a critical factor. Quality of life, well
one of the founding members of the Bristol established networks among individuals and
office. Other factors in the location may have companies and the research expertise of nearby
included the existing industries of the South universities are cited as other key factors in
West: a specialisation in defence companies location decisions.
and research.
These entrepreneurs, though located in the
New companies starting up in the 1990s South West, are focused internationally. That’s
and 2000s could draw on engineers who where their customers are, and that’s where
had located into the region to work at both growth for their companies will arise; there is
companies. Equally, foreign companies seeking no local market for these products. Therefore
to locate new R&D labs or design centres easy access to international companies, via
identify expertise in the area as a key driver nearby Bristol and London airports, influences
of investment decisions. This has undoubtedly their location decisions.
been a fundamental factor in the growth of
the cluster, as the family tree of semiconductor Where successful entrepreneurs reinvest, there
companies in the South West highlights (see are a number of factors involved. Partly it is a
p.12). Other international corporations such matter of professional pride: proving they can
as Intel, Dolby and National Semiconductor do it again, and that it wasn’t just a matter
recognised the growing importance of the of luck the first time.20 It is also a matter of
South West by locating their sales offices there. opportunity: a successful exit provides many
venture capital firms with the confidence
that success can be replicated in another
technological area. Investing in a proven
manager helps to lower their risk, so this gives
these serial entrepreneurs a relatively accessible
audience for any new ideas they have.

11
 Figure 1: Family tree of semiconductor companies in the South West

Fairchild 1972, Swindon Intel 1975 Swindon

Inmos 1978 Bristol Meiko, 1985, Bristol

Thorn 1984 Quadrics, Bristol

Division, 1986, Bristol PixelFusion, 1997, Bristol

Parametric Technology ClearSpeed, 2000,Bristol

Blu Wireless, 2009, Bristol

STMicroelectronics, 1987, Bristol
PETAPath, 2009, Bristol

Element14, Cambridge, Bristol

Broadcom Design Centre,

Icera, 2000, Bristol

AuPix
Motion Media, 1992
EsGem

Gnodal, 2008

XMOS Semiconductor, 2005

Art of Silicon, 2005

Silicon Basis, 2008

Siroyan, Reading, Bristol

3D Labs Design Centre

Tvonics, Wales

Zii Labs, 2008, Bristol

Ensigma Designs, Chepstow

Imagination Technologies

12
Plessey Semiconductor, 1961, Swindon Dialog Semiconductor, Swindon

TEMIC
GEC Plessey Semiconductors, 1989, Siemens Roke Manor,
Swindon, Roborough 1989, Christchurch

Dialog Semiconductor MBO

X-Fab, 1992, Roborough Innovision, Cirencester

Infineon Design Centre, Bristol

Lattice Semiconductor, Chippenham
Plus Semi, 2008, Swindon
IPWireless, Chippenham
Moortec, Roborough

NextWave

Plessey Semiconductor, 2010, Swindon, Roborough IPWireless

HP Labs 1985 Bristol Motorola, Swindon

Elixent 1995 Bristol Ubiquisys, Swindon Lucent Microelectronics, Swindon

PowerOasis, Swindon Pioneer

Panasonic Strategic Semiconductor Brooktree

Development Centre , 2006, Bristol

Oak
Acapella, Southampton Semtec Cascoda, Southampton, 2009

Conexant
Microcosm

Phasor Solutions picoChip, Bath

Conexant

Phyworks

Xintronix
Movement of people and ideas

Transfer of IP/Sale of company

Nanotech Semiconductor

13
 The rise of the South West Serial Entrepreneurs

Serial entrepreneurs have played an unit was wound up. In 2000, he founded
important role in the growth of the picoChip with Doug Pulley, who had also
semiconductor cluster. The experiences worked at Oak and then Conexant. The aim
of Stan Boland, Peter Claydon and Gary was to put together a massively parallel
Steele are highlighted below as example programmable multicore device for digital
entrepreneurs. Obviously, there are many signal processing.
others.
Early conversations with venture capital
Stan Boland company Pond Venture Partners helped to
Stan Boland joined Acorn computers in narrow down the target market to wireless
1997 as Financial Director, having worked and femtocell products, and they invested
at Rolls-Royce and ICL. He became CEO to generate funding of $7 million with the
of Acorn, and proceeded to reorganise founders. When the first devices launched
the company to focus on its chip design in 2003, an additional $17 million was
technology. With Acorn colleagues raised from existing and new investors.
including Sophie Wilson and Simon Knowles The firm now has a total of $112 million
from STM and Inmos, he created Element in venture funding, and major clients for
14 initially as a division of Acorn, and its femtocell technology. He is now with
then spun it out. He became CEO and in Chippenham-based wireless company
1999 they raised £8.1 million in first round Deltenna having left picoChip late last year.
funding. Even this amount was a challenge
– Stan met with nearly 40 funds, and made Gary Steele
77 presentations to raise this amount. Gary Steele is a graduate of the University
Element 14 was bought by Broadcom at of Southampton who has been especially
the height of the internet boom in 2000 for successful as a serial entrepreneur. He
$640 million, before the launch of the first founded Acapella in 1990, the UK’s first
product. fabless chip company, which was sold to
U.S. company Semtech in 1995. He then
Many engineers stayed with Broadcom, co-founded Microcosm, raising two rounds
but Boland, along with Simon Knowles, of venture capital, before selling in 2000 to
who had co-founded Element 14, Steve Conexant Systems, the same company who
Allpress and Nigel Toon went on to found acquired Oak Technology’s Bristol design
Icera in 2002 with the aim of developing centre.
a new generation of chipsets for mobile
broadband cellular communications. They In 2003, Gary Steele moved on to
shipped their first products in 2007, and found another new venture, Nanotech
have also raised around $250 million Semiconductor, with several former
in investment. They bought Canadian Microcosm employees. The company
company Sirific Wireless in 2008 and are designs and supplies integrated circuits for
reported to be considering floating on the the fibre optic communications industry.
London Stock Exchange. Nigel Toon left to The company has so far attracted $11
become CEO of Picochip in 2008. million in funding, and was able to acquire
an experienced design team when Intel sold
Peter Claydon off its fibre optic business.
Peter Claydon worked at the Bristol design
centre of Silicon Valley chip designer, He was also involved in the early days of
Brooktree, which was sold to Pioneer in picoChip and the creation of PhyWorks
1992. The Pioneer Digital Design centre in 2001, and successfully raised further
closed in 1997 and Peter become general venture capital investment in 2003.
manager at the broadband business unit PhyWorks was ranked the 5th fastest
at Oak Technology, another Silicon Valley growing technology company in the UK in
company with an office in Bristol. The Oak 2009, and was sold in September 2010 to
business unit was bought out in 2000, NASDAQ-listed Maxim Integrated Products.
and Peter Claydon stayed on until the

14
Financing new start-ups have been led by the universities who have
worked hard to cultivate links with companies.
Venture capital firms are often eager to back The focus has been on establishing common
managers and technology experts who have ground, rather than securing contracts or
already proved themselves with a successful income. In this way, government funding is
venture, as highlighted above. This can lead to used to work with the grain of private sector
patterns of investment, where the same funds activity rather than creating an alternative
will invest in a series of ventures that have the provider of services. This is illustrated by
same founders. examining the links between companies and
local universities (see Figure 2).
Stan Boland and Simon Knowles had offers
from many VCs after they sold Element 14, for The presence of research expertise at local
consulting and due diligence work as well as for universities has helped to anchor some of
funding. This gave them a receptive audience these companies in the area. Professor Joe
when they needed funding for Icera. McGeehan joined the University of Bath in
1985, bringing his communications research
Amadeus Partners’ Hermann Hauser put money expertise and links with industry; he is also
into Element 14, and then followed up by Managing Director for Toshiba Research
funding Icera, so he recognised the importance Europe. Professor David May was architect of
of backing successful entrepreneurs. Similarly, the transputer at Inmos, and was appointed
Pond Ventures were one of the early backers by the University of Bristol in 1995. He
of Microcosm, and subsequently led the first connects the university with industry very
round of funding for Gary Steele’s Nanotech effectively, influencing the development of
Semiconductor. Bristol’s Computer Science course towards more
practical work, as well as providing feedback
There are a number of mechanisms through to investors on a range of companies over the 21. Funding sources include the
Higher Education Innovation
which government provides support to early- years,22 and starting XMOS to continue the Fund, distributed by HEFCE,
stage technology companies. R&D tax credits transputer legacy. and the local RDA, the
South West of England
seem to be widely used in the sector, and Regional Development
professional bodies offer training sessions These universities are very outward-facing. Agency.
22. Reference: NESTA
on them, along with ways to access other Silicon South West was established by Bath interviews.
R&D funding and grants, both national and Ventures, the technology transfer unit of
European. These all play a part in the financing Bath University. Working with private sector
of semiconductor companies but because the partners, Silicon South West organises
sums of money needed in this industry are so networking events, conferences and industry
large (primarily to fund early research), the and technology updates for semiconductor
majority of company funding has generally companies in the region. They have also
come from venture capital sources. created a test lab in Bath to allow early-
stage companies to access expensive testing
Another way to support start-up financing is equipment at low cost, that would otherwise
to provide access to resources for free or at require a significant capital outlay. Government
low cost, reducing the amount of money that funding for Silicon South West is relatively
they need to raise. Mentor Graphics has been small, in the form of occasional sponsorship
a supporter of this model, offering its design from SWRDA and UKTI, and support for
tools at low cost, as an investment in the future overheads via the University of Bath incubator
custom of these companies. Universities also programme. Staff have leveraged private sector
have a role to play in giving start-ups access to funding through sponsorship and partnerships
their equipment and resources. from semiconductor and associated companies.
This effectively seeds private sector-led
activity and facilitates an open network of
entrepreneurs, companies, investors and
Broader network support university researchers.

A group of far-sighted universities, specifically

Bristol, Bath, Southampton and Surrey,21
have worked hard to create networks which The SETsquared partnership
bring together companies and academics.
From the start, the universities have sought Another key publicly backed investment
to collaborate with each other and more has been in the SETsquared partnership.
importantly with local companies. Key activities SETsquared is unusual as a collaboration

15
 Figure 2: Connections between companies and the universities of Bristol and Bath

University of Bath University of Bristol

XMOS Semiconductor,
Prof David May
2005, Bristol
Centre for Communications Research

Prof Andy Nix

NanoGAN, 2008, Bath ProVision Communications,
Prof Joe McGeehan
2001, Bristol
Prof Dave Bull
IQE, 2009, Cardiff

Toshiba Telecommunications
Research Lab , Bristol
GLOphotonics, 2009, Bath

Wireless Systems International, 1997, Bristol Icera, Bristol, Steve Allpress

Software Radio Technology, Toracomm, 1997, Bristol Andrew Corp, 2002, US

1990, Frome

SRT TETRA PowerWave UK, 2001, Bristol

SRT Marine Technology Systems4Silicon, Bristol Nujira Design Centre, 2008, Bath

AXIS NT, Bristol Silicon Gorge Systems, Bristol

Deltenna, Chippenham
Movement of people and ideas

Transfer of IP/Sale of company

picoChip, Bath

between the universities of Bath, Bristol, The partnership has incubated XMOS, a
Southampton and Surrey, where incubators venture capital-backed company established
from all four universities have been developed by Professor David May, from the University
under the same programme, integrating their of Bristol and a core member of the Inmos
enterprise activities. It develops partnerships in transputer team, and James Foster, former
enterprise activities and collectively supports CEO of Oxford Semiconductor. It is currently
the growth and success of new business housing and helping some promising
opportunities through spin-outs, licensing and companies, and more recently nurtured Gnodal,
incubation. The incubation element enables a company developing innovative networking
companies to keep initial costs down and technology to improve the performance of next
build a strong business case. It also works with generation data centres. Another graduate of
industry through research collaboration and SETsquared is Ubiquisys, a globally-recognised
consultancy. They provide services, and a focus femtocell company based in Swindon with
for service-provision companies such as IP backing from leading venture capital firms
lawyers and accountants. This is not specific to Accel Partners, Atlas Ventures and Advent
semiconductors. However, as microelectronics Venture Partners.
is a local industry, they support a number of
such companies in this area, and their mentor
network is biased towards the industry.

16
Part 3: Challenges for the semiconductor industry

Despite its exposure to economic cycles The success of UK companies and the South
and fluctuating consumer demand, the West cluster should not be taken for granted.
semiconductor industry continues to show The ability of firms to innovate is influenced by
strong underlying growth and this was seen the business and policy environment in the UK.
through the recent recession with demand The current business environment is obviously
for semiconductors holding throughout the very challenging, but there are some specific
industry (Figure 3). UK companies, particularly policy areas that are throwing up challenges
those firms at the design end of the process, for the industry. Some of the challenges
appear to have been less affected by the faced are industry specific, but most go wider
recession than other parts of industry. As and are similar to those that often arise in
innovation continues to drive the industry conversations with UK entrepreneurs and
forward, the services they provide remain in corporate executives. These wider challenges
demand. are outlined below.

Figure 3: Global semiconductor market by region

350

300

250

Revenue 200
in USD
billions
150

100

50

0
1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Asia Pacific Europe Japan Americas

Source: OECD data, based on World Semiconductor Trade Statistics (WSTS)

17
 Skills and those are the people who now lead UK
Recruiting in the UK is becoming challenging companies. Training is costly, and too few
as the pool of graduate electronic engineers companies are able to fund it now, leaving
continues to shrink. These companies also a skills gap. There can also be ‘too much’
need to be able to access and attract global emphasis on engineering skills: leadership,
talent, many of whom are being trained in UK vision, and business skills are also essential.
universities. Apprenticeships and funded work placements
may be part of the solution.
While UK universities are conducting
leading research, and providing courses in This shrinking pool of trained people endangers
microelectronics, large numbers of those the skills base of the South West cluster, and
taking them are not UK residents, and may as the generation of engineers trained at Inmos
not remain in the UK when they graduate. start to retire, there may be a critical shortage
In 2007/08, 68 per cent of electronic and of up-and-coming talent to replace them.
electrical engineering undergraduates and 21
per cent of postgraduates were UK residents. Access to finance
Overall numbers of electronic and electrical A critical issue for a high-tech cluster is
engineering graduates have fallen by 25 per attracting investors to provide financial support
cent since 2002/03.23 Although semiconductor for the high-risk, high-cost endeavours that
companies often recruit globally, it can be hard these companies undertake. While start-up
to attract people to the UK if they are also able activities in a sector such as semiconductors
to work in California, South Korea, Taiwan or are inherently risky, and expensive to test,
any other centre of the industry. many things can be done to mitigate the
risks associated with a new venture. The
The shortage of skills reflects not just the semiconductor industry, unlike some other
23. Higher Education Statistics pool of current graduates, but also those sectors such as digital media, has a more
Agency data.
with industrial experience and business skills. acute problem in that start-up costs are high,
Not everyone in the industry has a degree. with £3 million+ required simply to get a
A generation of engineers were trained at design implemented in silicon. The Innovation
large companies like Plessey and Inmos,

Innovation in semiconductors

Innovation is critical to success in this The high costs are a result of:
industry, with a very competitive market
and a demanding customer base. • Masking costs for patterning the silicon
Innovations in process and production wafers that have risen to around $500,000.
lead to lower prices, which make new
applications viable, which leads to further • Complex design tools needed to handle
product innovation. However, innovation is design tasks that make the best use
expensive, because even the design tools of chip real-estate, whilst operating at
needed to create a new chip design are nanometre scales.
expensive, and a modern manufacturing
plant can cost $3 billion or more. This leads • A demanding customer base which
to a position where incumbents are best seeks smaller, more powerful, lower cost
placed to innovate, because the knowledge and lower power use chips from each
and resources are cumulative, making it generation of technology.
very hard for start-ups to penetrate.
• A complicated, and expensive production
process, requiring clean rooms and
very high precision manufacture. A
semiconductor fabrication plant can cost
$3 billion or more.

18
Investment Fund (IIF) launched last year at pure blue sky research, and should seek to
should support some investment in this area. build consortia across Europe which are capable
of competing with Far East ones. Equally,
Incentives for R&D and manufacturing funding criteria for EU projects are often
Several industry players have argued confusing, with criteria difficult to understand
that the current tax framework does not or contradictory (“you only qualify for this
incentivise companies to invest in R&D and money if you don’t need it”). Companies felt
manufacturing, and arguably recent changes to that UK organisations, such as the Technology
capital allowances further tilts the tax system Strategy Board, also strayed into this territory.
away from more capital-intensive companies.
With the UK public purse firmly shut over the
While the case for greater incentives for R&D next few years, ensuring that EU funding is
and manufacturing need to be considered as fit for purpose and can be accessed by UK
part of the UK Government’s review of taxation companies will be important if the UK is to
of Intellectual Property and the support R&D maintain a steady stream of new entrants into
tax credits provide for innovation, this should the semiconductor market.
not ignore other areas where government could
play a role in supporting high-tech industries, No large companies in the UK
including semiconductors. Specifically the An underlying structural problem appears to
role of procurement and EU grant funding is be the lack of large companies in the UK who
considered below. are active acquirers. Large companies play
a number of important roles in the industry,
The best form of financing for a growing from training engineers to acquiring promising
company is its order book: securing a large small businesses to developing the next wave
contract from government has been the making of entrepreneurs. The UK’s lack of large player
of many firms. The Small Business Innovation in this area is a frequent source of frustration, 24. Bound, K. and Puttick, R.
(2010) ‘Buying Power? Is
Research (SBIR) scheme in the United States, leading many talented people to go overseas. the Small Business Research
established under the Reagan administration, A lack of ambition is cited, as is the desire for Initiative for procuring R&D
driving innovation in the
has been very successful at channelling many UK entrepreneurs to cash out early and UK?’ London: NESTA.
government R&D spending to small companies, get out of the business.
and stimulating growth by doing so. The
UK has a similar scheme, the Small Business It is also a reflection of a mainly venture-
Research Initiative (SBRI). After a slow start, funded industry, where access to the markets
the programme was reformed in 2009, and now has become very difficult, so the main exit
seems to be making progress, as highlighted in route is through a trade sale. When the only
a recent NESTA report.24 There are examples of companies large enough to become acquirers
how procurement by international governments are based overseas, this inevitably means that
has helped semiconductor companies. companies are under pressure to create an
exit for investors that involves control of the
Japan provided support for local procurement company leaving the UK.
of semiconductors from the 1960s onwards,
to encourage a local industry. Low-cost
financing was also made available, as well as
co-investment in R&D. Qualcomm, now one of Industry is leading the way
the leading fabless semiconductor companies,
benefited from the US Small Business Industry is not standing still. Many companies
Innovation Research programme in the early have recognised that these challenges need to
days of the company. be addressed as an industry, and a number of
initiatives seek to do this.
As well as funding, such initiatives also provide
a motivator for companies to work with Standards and regulation are long-standing
partners and construct both local and global roles for industry bodies. An example of this
supply chains to respond to a public service is Femto Forum, a not-for-profit organisation
challenge. conceived by Ubiquisys CTO Will Franks,
with global membership. This pushes for
More widely, the UK government could play standardisation and agreement on common
a role in ensuring that European initiatives interfaces, encourages deployment and
have more impact. Companies flagged that EU adoption of femtocell technologies, and
joint technology initiatives should be bigger develops the market.
and more market-targeted, rather than aimed

19
 In research, the Semiconductor Research partnering with their supply chain and
Corporation is a leading technology research providing access to quality people. This would
consortium, involving industry, academics and reduce the seed capital required to get the
government bodies in its research programmes, company to a first major investment round.
in the U.S. and around the world. It sponsors
pre-competitive university research of The cost of getting to a first product has
long-term benefit to the industry. SRC was increased exponentially for microelectronics
created in 1982 to boost the contributions start-ups, and it is now common to need to
of university research for the semiconductor raise £3 million or more simply to test a design
industry. EPSRC is a member of the Global in silicon. This level of risk is off-putting for
Research Collaboration programme within SRC, investors, and too large for more risk-friendly
connecting academics and industry researchers. seed investors and angels.
This programme directs research towards long-
term goals for the industry. The EMA aims to get a start-up to its first
product and initial customer engagement
The UK Electronics Skills Foundation was through brokering risk-sharing relations with
launched this year to improve the quality all of the critical and costly components of the
and supply of employment-ready graduates microelectronics supply chain:
to industry. The National Microelectronics
Institute was instrumental in putting together • ‘Platinum Partners’ – suppliers and
this collaboration, but critical to its creation manufacturers such as foundry, EDA tools
has been the willingness of industry to commit and IP companies.
resources to the Foundation and to sponsoring
individual students. Even a company as small as • ‘Blue Chip Partners’ – who provide access to
five employees is sponsoring a student through the end customer and market intelligence.
this scheme. The activities of the Foundation
extend from student sponsorship through to • ‘Knowledge Partners’ – such as universities
summer schools, scholarships and encouraging and Knowledge Transfer Networks (KTNs).
employers to work with schools.
Alongside these relationships, the Academy has
access to an ‘alumni network’ of entrepreneurs,
leaders and experts to help build teams, to
The European Microelectronics validate opportunities, and potentially to
Academy invest. The Academy aims to tackle some
of the specific challenges facing the UK
Industry is also playing a leading role in the semiconductor industry:
creation of the European Microelectronics
Academy (EMA) to overcome issues raised • Funding: by providing access to tools and
by high start-up costs and foster greater expertise that would be too costly for a start-
entrepreneurship. Successful entrepreneurs also up to acquire, the cost of getting to the first
attract venture capital firms, which are essential product will be lowered.
for the industry.
• Entrepreneurship: Access to a network of key
The founding partners in the EMA will be people and customers will provide support to
Mentor Graphics and NESTA. Mentor Graphics new ventures.
is one of three Electronic Design Automation
(EDA) tool giants whose customers are • Skills: Provide access to experienced people
microelectronics companies. Mentor has very to help support skills.
successfully operated a business development
unit, Cre8Ventures, for the last eight years. It will also help to reinforce the inter-company
This was set up to give them early access to relationships which are a characteristic of the
companies that could become the ‘next Cisco’ – South West cluster.
i.e. a multi-million account for their EDA tools
– via sharing the risk at the start-up phase

The Academy extends the risk-sharing What next for the South West cluster?
approach to other partners, to try to increase
the chances of these companies becoming The South West semiconductor cluster
‘Ciscos’. It is an incubator business that aims highlights how a successful cluster can grow
to de-risk microelectronics start-ups through with targeted, low level support:

20
• Investment in private sector led networks.

• Support for incubators and shared facilities.

The next step for the South West

semiconductor cluster is uncertain. There
is a shortage of venture funding for small
companies, and with some investment firms
pulling out of the sector altogether, the climate
appears to be worsening. Start-up capital is
in short supply, but a greater problem may be
a lack of growth funding to develop the big
companies of the future. Several companies
in the cluster have the potential to develop
to that size or larger, but it is not clear how
they can do this in the current environment.
While some of this may be addressed by the
Innovation Investment Fund, established in
2009, there are other proposals which address
the other barriers faced by companies in the
South West which deserve some attention.

The SPark Science Park

SETsquared and Silicon South West are
effective at bringing on ideas and small teams,
getting some seed funding in place and getting
companies off to a good start. However, the
planned SPark Science Park outside Bristol
is the next stepping stone. It is planned to
offer a good sized incubator that can help
start-ups move up, the kind of facility that
will turn good ideas into global businesses.
To be successful, it will need access to good
transport infrastructure, given the critical role
of global supply chains and funding in the
semiconductor industry.

Shared facilities
Silicon South West set up its test labs from
second-hand equipment donated by a down-
sizing multinational. Shared facilities are an
effective way of supporting small companies,
allowing them to pay only for usage, and save
on capital expenditure. Mentor Graphics/Cre8
ventures has been good at doing this with EDA
design tools and the Silicon South West Labs
are doing the same with test facilities, but
there are other areas such as verification, first
silicon shuttle runs from Taiwan, trade show
attendance and marketing that could benefit
from this approach. Indeed, the National
Microelectronics Institute has already convened
discussions on industry co-operation in the
area of verification.

21
 Part 4: Broader implications for government clusters
policy

At a time where clusters have again come to of governments investing heavily and still not
the forefront of policy discussions, it is worth creating a successful industry in their country.25
considering what lessons can be learned No semiconductor clusters created in the last
from the growth of the South West cluster. In 15 years have demonstrated sustained growth,
particular, it is worth asking the question: what despite significant efforts and investment in a
should governments do when faced with an number of countries.
existing cluster, to support its growth?
25. McKinsey Global Institute McKinsey Global Institute’s estimates of the
(2010) ‘How to compete
and grow: A sector guide levels of government investment are shown in
to policy.’ McKinsey & Table 1.
Company.
26. Ibid. Not a new industrial strategy
• In South Korea, unrestrained subsidised
An industrial strategy approach to cluster government investment led to the financial
development appears to be a high risk strategy. crisis that followed, with the chaebol left
A brief glance at the semiconductor industry with large debts, and oversupply in the
across the globe shows that this strategy industry.
is high risk and more prone to failure than
success. Whilst McKinsey Global Institute • Taiwan also supported its industry, but
points out that no successful semiconductor insisted on each company having a niche and
cluster has been created without government business model, making it much more robust
intervention, there are far too many examples when memory chip prices fell dramatically.

Table 1: Selected countries’ investment and market share in semiconductors

Estimated cumulative government Estimated current global

incentives to 2008 ($ billion)26 market share

United States 12 - 36 16%

Japan 19 - 54 16%

Taiwan 15 - 43 5%

South Korea 9 - 26 12%

Singapore 5 - 16 n/a

Germany 2 - 7 2%

China 6 - 17 21%

Malaysia 1 - 3 n/a

22
Those countries that have intervened creates competitive pressures that improve
successfully have had fortunate timing – when productivity and innovation, while customers
few competitors existed, or by introducing and suppliers that are clustered together
innovation, as in Taiwan’s new ‘foundry’ model. support the existence of specialist providers.30
The industry as a whole is highly cyclical, so
it is important for companies to be able to Porter recognises the problems with industrial
survive through periodic downswings, without policy, but argues that cluster policies can
preventing Schumpeter’s ‘creative destruction’ reinforce the development of all clusters.
from taking place. With funding extremely tight Targeting government support to clusters is a
in the UK, seeking to repeat the investments controversial policy. In defining the boundaries
of Inmos and Plessey is not an option open to of a cluster, industrial definitions are often
anyone. used, which can lead cluster policy to become
industrial policy by another name. This can
mean targeting a specific industry at the
expense of others through subsidies, tariffs or
Academic theories on clusters specific R&D funding.

Cluster research is rich with case studies and

hypotheses for how clusters start and grow. As
early as 1890, Alfred Marshall, the precursor So what works? 27. Marshall, A. (1890)
‘Principles of Economics.’
to modern cluster exponents, identified London: Macmillan and Co.
three factors that make geographic clusters Government involvement in clusters has a 28. Saxenian, A. (1996)
‘Regional Advantage:
beneficial:27 mixed history. In a Harvard meta-study of 833 Culture and Competition
clusters,31 government or chance was identified in Silicon Valley and Route
128.’ Cambridge, MA:
• Sharing of inputs: allows firms to procure as a contributing factor in only in 3.4 per cent Harvard University Press.
inputs at a lower price as part of a cluster of successful clusters. The study identifies 29. Florida, R. (2002) ‘The Rise
than they can in isolation. that government or chance contributions are of the Creative Class. And
How It’s Transforming Work,
observed far more frequently in unsuccessful Leisure and Everyday Life.’
• Labour market pooling: Flexibility and clusters (22 per cent), although the more New York: Basic Books.
30. Porter, M. (1998) ‘Clusters
mobility of labour is a key characteristic, important factor seems to be having a balance and the New Economics of
lowering recruiting costs for companies and of input conditions, competition, demand and Competition.’ Cambridge,
MA: Harvard Business
allowing knowledge spillovers to be shared related industries. Review.
more actively. 31. Van der Linde, C. (2002)
Government therefore has a poor record ‘Findings from the Cluster
Meta-Study.’ Cambridge,
• Knowledge spillovers: Workers learn from at creating clusters. Most successful MA: Institute for Strategy
each other; companies have ready access clusters emerge spontaneously, based on and Competitiveness,
Harvard Business School.
to the latest ideas and innovations; close some geographic or historical advantage. Available at: http://
proximity to customers and suppliers makes Governments can only identify clusters once www.isc.hbs.edu/
MetaStudy2002Prz.pdf
it easier to understand customer needs and they have emerged, and seek to provide the [Accessed 30 July 2010].
trends. This is combined with the ability infrastructure and facilitation that make it 32. Ibid.
to act flexibly and rapidly in response to easier for organisations within the cluster to
this knowledge; being part of a cluster operate, and for the cluster to become better
confers some of the scale benefits of a large known.
vertically-integrated company, without the
disadvantages of being slow to act. In addition to the usual suspects of skills and
transport infrastructure, our research in the
More recently, other researchers who have South West highlights two important, low cost
studied clusters include AnnaLee Saxenian, who approaches to supporting cluster development
examined clusters in Silicon Valley and Boston’s – backing industry-led intermediaries and
Route 128 and concluded that the greater facilitating movement between universities and
success of Silicon Valley was due to a more industry.
entrepreneurial and open culture, in contrast
to the large, closed firms that dominated the Intermediaries
Boston cluster.28 Richard Florida’s thesis is that Clusters consist of many different types of
clusters and innovation are positively correlated organisation, from the core businesses to
with tolerance and the presence of creative support service businesses, universities, trade
people.29 associations, and the customers and suppliers.
Van der Linde’s analysis32 suggests that
Michael Porter is widely associated with cluster competition, demand and supporting industries
research and policy. He argues that proximity are almost equally important in the successful

23
 Intermediary research centres – the example of the Interuniversity
Microelectronics Centre in Leuven, Belgium (IMEC)

A very different kind of government IMEC’s leading-edge research. Key elements

support is through research centres, such include:
as IMEC in Belgium, a major European
centre for semiconductor research that has • The centre seconds researchers from
been successful at attracting funding from partner companies to work at the
industry as well as government. institute. Of its 1,700 staff, about 30 per
cent are guest researchers and residents.
In recent months, both Hermann Hauser More than 40 per cent of staff come from
and Sir James Dyson have suggested that outside Belgium.
the government should consider developing
new technology research centres as a means • IMEC has created a classification system
to facilitating the development of products for IP which allows both open sharing
from research ideas and stimulating and retention of private rights for IMEC’s
greater exchanges of researchers between industrial partners.
academia and industry. IMEC is often cited
as a model for these centres. • IMEC has successfully created a centre
for expertise which attracts partners
IMEC is a world-leading interdisciplinary and generates an environment for co-
research centre, and often held up as a operative research. They work with 600
successful example of a public-private companies and 175 universities.
research institute. The centre continues
33. Porter, M. (1998) ‘Clusters to receive government support from the • They aim to lower the costs in certain
and the New Economics of
Competition.’ Cambridge, Flemish government, amounting to around technology areas by focusing their
MA: Harvard Business 16 per cent of revenue in 2008. The development. Research programs are
Review.
government has supported IMEC since its organised into two main streams: ‘More
creation in 1984, but it was ten years before Moore’ – to examine ways to scale the
the centre developed substantial external existing CMOS technology into ever
revenues, which come from companies and smaller sizes; and ‘More than Moore’ –
academic institutions that want to access looking at application-oriented research.

growth of a cluster. Porter suggests a number The key elements are to help organisations
of roles that a trade association or other to organise into supply chains; to seek out
intermediary can play to support a cluster:33 the infrastructure and services that there is a
collective need for; and to help promote and
• Provide a forum for the exchange of ideas market the cluster, to ensure it becomes better
and a focal point for collective action. known as a source of expertise and output.
The South West highlights how intermediaries
• Establish university-based testing facilities can provide links to mentors and management
and/or specialist training programmes. talent, and the industry working together
can make equipment or resources available at
• Collect information and data relating to the favourable rates to allow companies to test
cluster. ideas or prototypes at lower cost.

• Offer help on common managerial problems. The success of university/industry partnership

approaches in the South West, whether through
• Investigate solutions to environmental issues. Silicon South West or the SETsquared partnership,
highlight how Porter and Van der Linde’s theories
• Organise trade fairs and delegations. can be applied in practice. Equally given the
relative low cost models adopted in these
• Manage purchasing consortia. partnerships, these are the sorts of investments
that government can usefully make in a fiscally
constrained environment.

24
Facilitating movement between universities • University courses which combine technology
and industry training with business can help to provide the
There is a high level of job mobility in Silicon graduates that are most sought after, as well
Valley,34 due to a number of factors, including as providing an opportunity to engage the
the modular nature of the computer industry, businesses themselves with the curriculum
where external economies of scale are and even delivery.
particularly important; a ‘California’ effect
due to the unenforceability of non-compete
agreements in California law; and an industry
where gains from new innovations are both
large and uncertain.

The origin of the South West silicon design

cluster was arguably the opening of the
Fairchild Semiconductor design office in Bristol,
which was said to be motivated by a desire by
Fairchild to open offices in locations where it
was easier to retain engineers, because they
would not be lured away by local competitors.
Ironically, one of the original members of that
office went on to become Design Manager of
Inmos.

As the cluster has developed, this movement

of individuals across companies has been
supplemented by movement of individuals 34. Fallick, B., Fleischmann, C.
and Rebitzer, J. (2005) ‘Job
between universities and the private sector. Hopping in Silicon Valley:
For example, Professor David May was initially Some evidence concerning
the micro-foundations of
at Inmos, later joined the University of Bristol a high-technology cluster.’
and has now also established his own company, NBER Working Paper Series,
No. 11710. Cambridge, MA:
XMOS. This movement has been facilitated by NBER.
the work of far-sighted universities who have
actively facilitated this lateral movement (see
Figure 2).

The movement of people between companies

and institutions supports and defines the
cluster, and can lead to greater levels of
collaboration. A number of measures can help
to support greater mobility:

• Events that bring the community together

around a common purpose are where much
of the ‘gossip’ that sustains mobility occurs:
who is starting something new, who is an
authority, who is good to work for. Sharing a
task, such as building a consortium for a bid
can create deeper links than more general
networking.

• Opportunities for work placements and

training build links between companies and
institutions, as well as increase the pool of
skills. While confidentiality is a major concern
in an IP industry, there are many companies
whose skills are complementary rather than
being rivals.

25
 Part 5: Conclusions

Semiconductors are not considered a particular Clusters can be reinforced by measures and
strength in the UK – we don’t have an Intel institutions that anchor the industry to
or a Samsung here. And yet, semiconductor the area, and promote the inter-industry
design is considered a real UK strength links that promote the cluster and support
by the global industry, and contributes to smaller companies. Intermediaries can rally
regional growth in several areas of the UK, but companies around common causes such as
most notably in the South West, attracting promotion and standards; incubators provide
significant investment and people into the support to smaller companies and make
area. critical connections to people who can make
a difference to the business; shared resources
This cluster is anchored on an historic skills between companies or provided by academia
base dating back several decades. But the can lower costs for companies in an industry
cluster would not be sustained on that basis where every part of the process is characterised
alone. It survives because those skills are by incredibly expensive tools and processes.
retained in the area, the expertise is renewed,
and new companies continue to be formed.
With the retirement of a generation of Inmos-
trained engineers, the South West will face a
new challenge in finding skilled engineers who
understand the processes involved as well as
the business issues.

As this report highlights, many of the issues

of the semiconductor industry are shared with
other industries. Connecting the university
courses to the demands of employers is
a common theme in many science and
engineering industries, and sector-spanning
people like Professor David May can be
invaluable in opening up these links.

Mobility is a key attribute of a cluster: freedom

of people to move between companies, and
between industry and academia, creates greater
innovation and enhances the attractiveness of
the cluster as a whole. A critical mass of similar
companies means that people can locate there
with confidence that if any given company
fails, there will be others who are seeking the
same skills.

26
Acknowledgements

NESTA is grateful for the time and insights of the following people. All views and omissions are, of
course, our own.

Dr Nick Appleyard, Technology Strategy Board

Rupert Baines, PicoChip

Simon Bond, Bath Ventures and Silicon South West

Carson Bradbury, Mentor Graphics and Crea8 Ventures

Adrian Buckley, Mentor Graphics

Peter Claydon, PicoChip

Steve Cliffe, formerly Xintronix, now ProVision Communications

Russell David, Clearspeed

Guillaume d’Eyssautier, ADD Semiconductor and GDE Consulting

Helen Finch, Infineon

Nick Flaherty, Silicon South West

Tony Ford, Gnodal

James Foster, formerly XMOS, now Apple

Chris Garcia, SWRDA

Matt Hatch, formerly Gnodal, now ProVision Communications

Fred Homewood, Gnodal

Libby Kinsey, NESTA Investments

Simon Knowles, Icera

David May, XMOS

Professor Joe McGeehan, Toshiba

Malcolm Penn, Future Horizons

Dr Rufus Pollock, Open Knowledge Foundation

Chris Smart, Acacia Capital Partnerships

Gary Steele, Nanotech Semiconductor

Professor Eric Thomas, University of Bristol

27
NESTA
1 Plough Place
London EC4A 1DE
[email protected]
www.nesta.org.uk

Published: October 2010

SWS/60