Foundry 2.0: The Next Generation of Foundry-Fabless Relationships

Foundry 2.0: The Next Generation of Foundry-Fabless Relationships

Loading Social Plug-ins...
Language: English
Save to myLibrary Download PDF
Go to Page # Page of 19

Description: After a stunning run in the nineties, the foundry business model reached a plateau where relationships between foundries and their customers shifted to more adversarial positions. There was a greater focus on wafer prices and daunting technology and cost challenges would soon emerge. Customers began to call for next generation foundry business model after huge hickups at the 40nm and 28nm nodes.

GLOBALFOUNDRIES responded to that call with it's Foundry 2.0 model. Could this be real or could it just be marketing spin?.

Author: G. Dan Hutcheson (Fellow) | Visits: 3614 | Page Views: 5986
Domain:  High Tech Category: Semiconductors Subcategory: Foundry 
Upload Date:
Link Back:
Short URL:

px *        px *

* Default width and height in pixels. Change it to your required dimensions.

Foundry 2.0


The Next Generation of Foundry-Fabless Relationships
Why it’s Different and Why You should Care

July 2013

By G. Dan Hutcheson
VLSIresearch inc


Index of Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Overview and General Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Early Foundry Development — the 1980s . . . . . . . . . . . . . . . . . . . . . 4
Foundry 1.0 — the 1990s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Infrastructure Shifts of the Late Eighties and Early Nineties . . . . . . . . . . . 5
Foundry 1.1 — the 2000s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Market Pressures and Emerging Adversarial Partnering Styles . . . . . . . . . . 7
Technology Pressures for Closer Partnering . . . . . . . . . . . . . . . . . . . 8
Foundry 2.0 — the 2010s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Addendum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
About the Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Notices, Terms of Use, Disclaimers, etc. . . . . . . . . . . . . . . . . . . . . . 17


Foundry 2.0, as presented by Ajit Manocha, CEO of GLOBALFOUNDRIES . . . . 3
Wafer Fab Cost Escalation: 1980-1993 . . . . . . . . . . . . . . . . . . . . . . 6
What Drove the Emergence of the Foundry Business Model. . . . . . . . . . . . 7
R&D Spending Rates Increase. . . . . . . . . . . . . . . . . . . . . . . . . . 10
How Foundry Partnering Styles Became Adversarial. . . . . . . . . . . . . . . 11
The Foundry 1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Key Features of each Foundry version . . . . . . . . . . . . . . . . . . . . . . 14
The Foundry 2.0 Revolution . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Overview and General Synopsis
After a stunning run in the nineties, the foundry business model reached a plateau in
the 2000s. At this plateau, relationships between foundries and their customers shifted from positive partnering styles to more adversarial positions with a greater focus
on wafer prices. Daunting technology and cost challenges emerged as the industry
crossed into the 2010s, which led to calls for new ways of doing business. Customers needed a next generation foundry business model. And they began to call for one
after huge hickups at the 40nm and 28nm nodes.
GLOBALFOUNDRIES responded to that call with Ajit Manocha’s Foundry 2.0
model. His vision was for a completely restructured fabless-foundry relationship
that mimicked what he had experienced as Chief Manufacturing Officer at Philips
Semiconductor. Yet this was the same person that had taken Philips/NXP out of the
strict Integrated Device Model and had migrated them on a long journey through
fab-lite to fabless. Indeed, this man who had once called the IDM model dead was
now calling for a return to something similar to it. Could this be real or could it just
be marketing spin?

It is Time for Foundry 2.0
Foundry 2.0, as
presented by Ajit
Manocha, CEO of

Traditional Model
J Contract

Foundry 1.0

Foundry 2.0



JFoundry Partner

JCollaborative Device

This white paper looks into why this new approach to the foundry model is needed
by first examining the history of how the semiconductor foundry business model
developed. Then, it delves into why things went wrong in the 2000s and shows how
that led to the need for a next generation — a need that Foundry 2.0 answers. It
shows how Foundry 2.0 is far more than marketing hype. It is a completely different
way to run the business.

Early Foundry Development — the 1980s
Prior to the first foundries, almost all semiconductor companies owned their own
fabs. The design and manufacture of chips were tightly coupled. Then in the early eighties, the emergence of commercial EDA tools became the big bang that
cleaved design from fab ownership. At first, this resulted in fabless companies going
directly to the then emergent ASIC1 semiconductor segment or contracting out to
semiconductor companies with excess capacity. System companies became more
independent of this approach as EDA tools evolved hierarchical abilities that could
take an electrical schematic, convert it into a physical layout of the chip, and tape it
out to a mask writer to make all the individual layers.
At the same time, node development became stable with few process changes.
Things like Copper and HKMG2 were on the far horizon. So, the electrical effects of
scaling were reasonably model-able and easily constrained by design rules, further
separating process from design. So as the cost of fabs rose, OEMs began to shift
out of captive manufacturing into a new world of contract manufacturing.
This independence of design also led to the first fabless chip companies, the most
famous being Chips and Technologies. There were inherent conflicts of interest with
this early contract manufacturer approach, because the semiconductor companies
doing the contract manufacturing also sold their own, sometimes competing, chips.
The potential for IP theft, real or imagined, was high. This restricted opportunity,
thus opening the window for the foundry model.

Foundry 1.0 — the 1990s
Morris Chang is generally credited with inventing the foundry business model with
the founding of TSMC. While the word ‘foundry’ existed before TSMC, there was
little difference other than its use as a marketing term between what was called a
foundry and contract manufacturing. Chang clarified the difference by declaring that
TSMC would never compete with its customers by selling semiconductors in the
open market. It would only handle the manufacturing part of the process flow, leaving design, marketing, and sales to their customers. Chang also established strong
IP security, so that TSMC could not be a conduit for a fabless company’s ideas to
leak out.
The key breakthrough was Chang’s vision that trust was the essential ingredient
of a true foundry partnership that would stand the test of time. This unlocked the
growth potential of the foundry business model. He then built TSMC around the
mission of being a trusted partner. This distinction catapulted them into the leadership position.

1 ASIC: Application Specific Integrated Circuit
2 High-K Metal Gate

Chang’s model came at a critical juncture in history. Differentiation in the semiconductor market had shifted from manufacturing to design with the development of
VLSI3 scales of transistor integration. Semiconductors were making the leap from
being system building blocks to accounting for significant portions of the system
itself, on a Moore’s Law march that would ultimately lead to SOCs4.

Infrastructure Shifts of the Late Eighties and Early Nineties
The shift of manufacturing to foundries was also aided by several key infrastructure
shifts: The most important of which was that fabs were becoming too expensive to
own and keep.

Average Cost to Build & Equip a Wafer Fab:


Wafer Fab Cost



‘80 ‘81 ‘82 ‘83 ‘84 ‘85 ‘86 ‘87 ‘88 ‘89 ‘90 ‘91 ‘92 ‘93
Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

The foundry movement hit high gear in the early nineties, when the cost of a fab
was just passing the one-billion dollar mark. The cost of a fab was roughly rising
at about half the rate of Moore’s Law, or a doubling every two nodes.5 Moreover, it
was not just the cost of building a new fab, because existing ones had to be upgraded every node for a chip maker to stay in the game. The incremental cost of
keeping a fab up-to-date was a growing capital burden, which became another big
barrier to entry. Too big for venture capital, early fabless customers were beginning
to pool funds in a fab sharing model, where they would get guaranteed access to
capacity without threat of IP theft.
3 Very Large Scale Integration, designating chips with transistor counts above 100K.
4 SOC: System-On-a-Chip
5 Sometimes called “Moore’s Second Law” due to an erroneous quote in Business Week, as Moore never showed the
correlation. It’s also been called “Rock’s Law” and “Hutcheson’s Law.” The observation was first published back in the
late ‘80’s by VLSIresearch, the chart shown here was used then to demonstrate the tendency of fab costs to double
every two nodes. The trend was first named “Hutcheson’s Law” in 1989 by Hartwig Ruell of Siemens; then “Rock’s
Law” in 1994 by Ken Thompson of Intel; and finally “Moore’s 2nd Law” by Business Week in 1996.

What Created Foundry 1.0

What Drove
the Emergence
of the Foundry
Business Model

Fab cost
Do not
compete with
customers rule


value lower


shift to design

Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

At the time, it was getting harder to gain a differentiable advantage with process,
while the cost of developing unit processes was rising fast. These were two important infrastructural shifts that favored foundries.
IDMs also played a significant role in making it possible for the emergence of foundries. One was that during periods of sluggish growth and downturns, IDMs would
shed engineering talent. This talent migrated to equipment companies and Asia.
Equipment companies started to offer complete process solutions that came free
with the purchase of their tools. This and an excess of manufacturing talent willing
to move to Asia, Morris Chang being the most notable, made it structurally possible
for the foundries to emerge in Taiwan and Singapore.
Many IDMs responded to meeting the foundry challenge by focusing on cost reduction and automation, which made their fabs inflexible. Worse, they were often
inefficient, leaving an important window of opportunity open for the foundries. The
ultimate response to this was the ‘fab-lite’ movement, in which IDMs ceased to
invest in leading edge manufacturing. In short, they were recognizing that differentiation was now mostly in design.

Foundry 1.1 — the 2000s
By 2000, the foundry business model had moved from an idea for companies who
could not afford fabs to being front and center in the mainstream of manufacturing. IDMs had joined the fabless, with their fablite approach. Moreover, the largest
fabless companies now had the scale to build their own fabs, but they were not
building fabs. This was a powerful demonstration of the foundry business model’s
strength. Many had come to believe the foundry was the future of manufacturing.
But the nature of the business was changing as storm clouds formed on the horizon. These storm clouds grew darker as conflicting market and technology pressures were forcing change.

Market Pressures and Emerging Adversarial Partnering Styles
The market pressures of the 2000s were multifold. There was a rush of capital into
the business, making the number of foundries competing for business rise. Price
competition emerged as a significant force in business relationships. This rise in
competition eroded the foundation of trust between fabless and foundry companies. The thinking on the fabless side was that if process offered no differentiation,
wafers were commodities and so price should be the most important factor in the
buy decision. This downplayed the service foundation of the foundry. Partnering
styles became more adversarial. Fabless companies began to push for a reference
process platform via the FSA6 to easily move production between foundries, thus
giving greater leverage over pricing. They were also aggressively building their own
internal process teams to have a better understanding of manufacturing costs for
the purpose of price negotiation leverage. Also, the very infrastructure shifts and IP
fluidity that had benefited emergent foundries in the 1990s was now biting back in
the 2000s. Now, IP and people were flowing out of the leading foundries.
The leading foundries responded by reducing transparency — limiting access to
early process and electrical results to avoid IP transfer to their competitors. Service,
early in the design cycle, was still where foundries gained their highest price leverage. This was especially true if they could restrict information flows, thus making it
more difficult for customers to enable price competition.
At the same time, consumer and mobility, the fastest growing markets for fabless
companies, demanded shorter design cycle times. Getting to market late resulted
in costly market share losses that could even drive a fabless chip company under.
Design times needed to be shorter and getting the design right at the first tapeout,
called first-time-right, became essential. So the battles over wafer prices were not
easily solvable.

6 Fabless Semiconductor Association, which is now the GSA or Global Semiconductor Association.

Technology Pressures for Closer Partnering
At the same time, the technology imperative to follow Moore’s Law was suddenly
getting more difficult. The era of simple scaling was coming to an end as the industry crossed the 100nm boundary, marking the transition from microchips to nanochips. The rising challenges of new wafer sizes, materials, and processes would
prove daunting. 300mm meant retooling entire factories. Entire processes had to
be reintegrated before they would yield on 300mm. Copper had problems with
thermal-migration and via fill. Lo-K inter-metal-dielectrics were fragile. CMP caused
dishing. Then, variability due to random dopant fluctuations came into play. Many
times, the problems were not known beforehand, causing costly re-spins in the
130nm, 90nm, 40nm, and 28nm nodes. At the same time, the explosion of transistors gave rise to SoC, which added more complexity and density variation, creating
design-specific micro-loading problems across the die. Design and Process were
intertwining, leading to the need to insert a new layer into design flows called DFM7.
The move from tape-out to production kept getting delayed when leading-edge
nodes came out of the foundries. Managing design rules and the risk associated
with them became another problem. In Foundry 1.0, control went in the direction
of fab through the PDK and design rules to the designer. Follow the rules and the
fab would never stand in the way of getting silicon right the first time. As Foundry
1.1 developed, the combination of growing design and process complexity led to
more conflicts between the two. Designers were forced to take on more risk of firsttime failure via exclusion sign-offs. Exclusions are when a design rule is broken, as
invariably happen, and the designers sign-off takes the foundry off the hook for any
failure related to it.
Initially, exclusions were simply ways to manage and assign risk while freeing designers to execute according to their judgment. Initially, exclusions were relatively
small in number, easy to manage and the risk levels easy to comprehend. However
as design complexity exploded, the number of exclusions blew up with it. Delays in
getting to market grew with it. Managing the many design rule exclusions with each
became so bad that EDA companies responded with tools to help. Not to question
the value of these tools, but is this any way to treat a customer? The very fact that it
exists implies lack of trust.
Meanwhile the IDMs, with tight couplings between design and process, were not
experiencing delays in getting to market — Intel the most notable. This highlighted
the growing importance of process and how it no longer could be had for free when
you bought tools from equipment suppliers.
7 DFM: Design for Manufacturing

Process integration became very difficult as materials and device physics played a
larger role with each successive node. This drove process research back into the
chip manufacturers. As the need for a tighter connection between design and process rose, so did the cost. These factors led to IBM founding the Common Platform
to pool R&D resources across several companies, keeping members on the leading
edge for significantly lower costs. This was a precursor for the transition to Foundry
2.0, as GLOBALFOUNDRIES was involved earlier as a full IDM inside AMD.
But the large traditional foundries adapted slowly to the swiftly changing technical
landscape. Foundries had historically underspent on R&D, at an average of only
4.7% of sales in the mid-90s. They relied on suppliers to carry far more of the load
than traditional IDMs. Even though the spending rate increased by almost half in
the mid-2000s to 6.4%, it was only about a third of what IDMs were spending at
the finished wafer level. It can also be argued that their greater secretiveness and
low participation in global conferences hurt them and their customers. They were
often reacting to problems rather than anticipating them, even though the knowledge needed to anticipate them was public.

R&D Spend Rates
R&D Spending
Rates Increase

IDM Wafer Level R&D



Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

Compounding this, many customers were slowing their adoption of new nodes,
favoring a ‘second mouse to the trap gets the cheese’ strategy. So it was hard for
foundries to get the volume runners that could be used to drive defect densities
down fast. Moreover, their traditional use of SRAMs as yield learning vehicles was
failing to yield predictable first-time-right results with logic and mixed-signal. There
was too much interaction between the design and the process. This was compounded by the rise of SoCs, which made for far more complex chip floorplans.
The relatively lower level of leading-edge volume-production ramps at foundries also
meant that the return on the R&D investments was getting pushed out.

Pricing models broke down, as foundry customers began to insist on factoring yield
variability into wafer prices. Meanwhile, foundries wanted to charge for test wafer
runs not seeing their need as a better yield learning vehicle. It is a key sign of broad
trust erosion in a market when pricing models breakdown.
Meanwhile, the size and cost of Gigafabs drove the large foundries to reduce
flexibility, favoring larger customers who were willing to pay for the volumes
to drive defect density down. They also started tailoring capacity investments
closely to wafer order volumes, versus the nineties when they would build entire
fabs before a market for the next-node had developed. This and the unpredictability of ramps would result in capacity crunches just when a market was taking
off, frustrating customers. Customers also became frustrated as some foundries
started to expand outside of their traditional markets and take on conflicting
roles in advanced packaging.

How Foundry
Styles Became



As the 2000’s came to a close, the leading edge was in danger of becoming the
bleeding edge. IDMs had pulled ahead, their integrated development model giving
them a time-to-market advantage without the disruptive ramp hiccups that Foundry
1.0 had evolved into. Some even speculated the foundry model might be dead.
Fabless companies saw what IDMs could do and wanted the same. The value of integration had awakened after having lain dormant since the nineties. The dual ascent
of the Cloud and Mobile Era were placing a premium on power-performance at a level
never seen before and that put the focus back on transistors. Hence, design-process integration was now a differentiator in the market. Both Qualcomm and Nvidia
called for a closer working relationship with their foundries. Then, Nvidia’s John Chen
caught the world’s imagination with a new fabless-foundry relationship he called the
“Virtual IDM.” What was new, according to Chen, was a critical question to ask in
each fabless-foundry decision: “If we were one company, what would we do?”
One thing was certain: there was a need for something new.

Foundry 2.0 — the 2010s
GLOBALFOUNDRIES, to a great extent, was perfectly positioned to address the
emerging need for a new foundry model: Its roots were as an IDM, having spun
out from AMD in 2009. Having acquired Chartered, it also gained deep roots in the
foundry 1.0 model, allowing it to bridge both worlds. Its CEO, Ajit Manocha, deeply
understood the issues. He had been in the senior management of an IDM, taken
it through a migration to fab-lite, and then fabless, spurring the development of
the original foundry business model in the process. He had been chairman of the
SSMC TSMC-NXP joint venture fab in Singapore. He had started his career in R&D
at Bell Labs. So he understood what needed to be done as design and process
began to interact again. He could relate to the frustrations with Foundry 1.1 the
fabless companies were having -- not just strategically, but also on a personal level.
Unlike other foundry CEO’s, he had no personal stake in a status quo that he had
defined. He was starting all over again at GLOBALFOUNDRIES with a clean sheet
of paper. So it was no surprise that Manocha would be the first foundry CEO to
address the issue, spelling out a new model he called Foundry 2.0.

The Foundry 1.1

The Foundry 1.1 Problem


Foundry 1.1



Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

The core of Foundry 2.0 is about “Collaborative Device Manufacturing,” as Manocha puts it. This new working relationship must meld the seamless collaboration
of an IDM with the flexibility of the fabless-foundry model. A Foundry 2.0 company
must be structured to collaborate seamlessly, meeting Chen’s test of ‘one company
decision making.’ This structure must allow the fabless company to innovate on the
foundry’s platform as an extension of its own strategy starting early in a new process node’s development. This is far beyond Foundry 1.0’s model of primarily being
a source of generic fab capacity.

Foundry Version Features
Key Features of
each Foundry

Seamless parallel collaboration
One company decision making
Cross-company team collaboration
Parallel development
Design-process coupling
Process technology critical
Price focused, solution assumed
Partnership first
Service oriented business model
Foundry does not compete w/customer
Foundry shoulders capital burden




J: Solution Provided



°: Problem Area

Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

With Foundry 2.0, fabless designers must be allowed to innovate early on the
platform in a parallel development process versus Foundry 1.1’s serial development
model. Engineers on both sides must be able to communicate freely and openly
without sales or procurement running interference, like in Foundry 1.1. Designers
can drive the fab, tearing down design rules where possible; as they work with FTS8
improve the PDK and the process bandwidth of the fab. This is the only way to rapidly overcome design-process coupling challenges. This also allows the foundry’s
process technology to be an extension of the customer’s strategy. When the process enhances the design, Foundry 2.0 results become like today’s most advanced
IDMs, rather than the plain vanilla processes of Foundry 1.0.

8 FTS: Field Technical Services

One of the key problems with Foundry 1.1 was that short-term profit got in the way
of partnering for long-term gain. Foundry 2.0 overcomes this limitation by enforcing
Bruck’s rule that “partnering only occurs when common interest rises above proprietary interest” fully down the command chain. This is not as easy as it sounds
because every link in that chain needs to think like a customer, which builds trust.
Trust is what makes customer and supplier willing to go long on profit together by
maximizing the partnership and never sacrificing it for short-term cost gains.
Foundry 2.0 is not business as usual with a new name. Trust breakdowns between
organizations are like a cancer, were things go wrong at small unseen sites and
then metastasize throughout the relationship. So Foundry 2.0 requires a different
culture. Foundry 2.0 is a way of doing business.
To pull this off Ajit Manocha had to build operational platforms and processes to be
Foundry 2.0 compliant. One simply can’t say you’re going to think like one company. You have to build it into the organization so it becomes second nature. Manocha
started by gluing together the disparate IDM culture of the old AMD manufacturing
group and the Foundry 1.0/1.1 culture of Chartered. When he started, these two
parts were largely functioning as two companies. As he integrated them, he also
embedded trust building into GLOBALFOUNDRIES’ infrastructure by assembling a
leadership team with backgrounds from wafer processing, foundries, IDMs, & Fabless companies. These people had similar frustrations with how the foundry model
had evolved and backgrounds to understand how to overcome many of the day-today road blocks that undermine trust.
Having done this, Ajit restructured the company to create what he calls “Mini-IDMs.”
His innovation is to move the sales-buyer relationship up to the corporate level,
thereby removing this wall separating fabless designers from foundry FTSs, and
process engineers in the Foundry 1.1 model. This allows them to execute decisions
like they are a single company.
Can Foundry 2.0 really drive change? The pressures that created Foundry 1.1 still
exist. But the difference today is that it is the fabless companies who see the need
for change the most. Markets change far faster when customers pull than when
vendors push.
Foundry 2.0 has already seen significant successes at GLOBALFOUNDRIES. Those
working under it note that getting a design to market is more like working together
as two internal departments with a common goal, instead of Foundry 1.1’s model
of two independent companies with independent focus on profit share. Working
together is not encumbered by having to go up through procurement over to sales
and down; and then return back through this torturous path.

The Foundry 2.0 Revolution
The Foundry 2.0














Chart approved for public release with attribution. Copyright © 2013 VLSI Research Inc. All rights reserved.

One example of how GLOBALFOUNDRIES lowered the competitive barriers is
where they worked concurrently with a major fabless chip company to bring up an
alternate source capability for a new app processor in parallel with another foundry. GF met the challenge of being able to align the technology definition to meet a
single GDS in real time so their customer could ramp at 2 foundries simultaneously.
They pushed aside proprietary concerns in favor of trust-based open collaboration,
giving key access to their technology architecture team during the early technology
definition stage. Giving this level of access showed an incredible level of trust on
GF’s part. Yet, this high level of trust benefited the customer in lowering the risk of a
first-time-right failure.
Trust is a two way street and one critical sign that GLOBALFOUNDRIES earned
was the customer pushed back to the first source, insisting on more commonality
from both foundries at the architectural level. The customer pushed for give and
take on both sides. This told GF that the customer was taking care of their interests. This is an important Foundry 2.0 difference, because alternate sources in the
Foundry 1.1 model had margins compromised in trying to match the first source.
Aligning foundry processes costs the customer little, other than effort, and the trust
benefits are large.
Ensuring your supplier’s profitability is reasonable and fair is critical to long-term
supply base health. If suppliers are not profitable, they won’t stay in business. So
your supply base can shrink down monopolists. But companies only go to this
length when they intend to partner with a supplier over the longer term. So this
customer sent a clear positive signal to GF that it was not abusive. Small reciprocal
trust efforts create positive trust feedback loops in a partnership.

With another major fabless chip company, GLOBALFOUNDRIES went deep with
on-site support around their PDK and 3rd party IP libraries, which got the customer
to market faster, enabling greater R&D efficiency as well. Deep means more than
just training or application support. In this case, the actual PDK evolved as they
worked together. The customer was implementing a high-performance high-voltage
analog mixed signal design that required variable power based on dynamic operating conditions in order to save battery life. In essence, GF allowed this customer to
drive the fab with a living, breathing PDK -- a critical Foundry 2.0 difference.
Another Foundry 2.0 challenge is having an ability to work with smaller customers.
In this case, GLOBALFOUNDRIES worked with an up-start to bring up an ARMbased cloud processor architecture on their 28nm HPP process. Collaboration with
this customer began with an early PDK version, soon after the technology definition
was completed. GF gave full access to GF’s relevant development teams. Together, they used Foundry 2.0 close partnering to take product performance to a higher
level and do it faster without costly delays. They even helped the customer refine
their product’s architecture. This customer had not experienced the level of technical support GLOBALFOUNDRIES provided from any other foundry.
Importantly, this early 28nm design was right on first silicon and quickly moved
to market. This is proof that Foundry 2.0 works, as there were plenty of 28nm
high-performance designs brought up in Foundry 1.1 that met with failed first silicon. Many had delays significant enough to arguably effect quarterly results.
These examples demonstrate that GLOBALFOUNDRIES has been able to take
Foundry 2.0 beyond the buzzword level and deep into how their organization
functions. One of the highest metrics of partnering ability is that you take care
of your partner before you take care of yourself. With the above examples,
GLOBALFOUNDRIES has demonstrated that they have instilled this ability at the
working level and turned Manocha’s Foundry 2.0 into a way of doing business. In
doing so, they have brought a fresh breath of life to the foundry business model.

About the Author
G. Dan Hutcheson is CEO and Chairman of VLSI
Research Inc. His career spans more than thirty years,
in which he became a well-known visionary for helping
companies make businesses out of technology. This
includes hundreds of successful programs involving
product development, positioning, and launch. Dan is a
recognized authority on the economics of innovation
and has a proven track record of being able to predict
trends accurately using the economic models he
develops. He is a senior member of the IEEE and a
recipient of SEMI’s Award for outstanding contributions
in marketing for his extensive development of the
economics behind the semiconductor industry.
He has authored numerous publications on the economic, strategic and tactical
aspects of how to succeed in the business of technology, which includes numerous
articles and weekly analysis of the memory market. He is respected as a Moore’s
Law scholar, having published many papers and invited talks on the subject over his
career. This includes, most notably, work for the IEEE, the Semiconductor Industry
Association, the National Institute of Standards and Technology, SEMATECH, SEMI,
and The Electro Chemical Society. He has twice authored invited articles for Scientific American on Moore’s Law. He has also been the keynote or invited speaker
on many other technology topics at dozens of conferences, including the Robert S.
Strauss Center’s Technology, Innovation and Global Security Speaker Series.
Dan is arguably best known for his forecasting of strategic infrastructure shifts and
his early-eighties development of the first factory cost-of-ownership models, which
are now the basis for most large-scale capital decision-making. He predicted the
shift of the DRAM memory market from the United States to Japan in the 1980’s,
then the shift of it to Korea in 1990s, as well as the driving forces for the rise of
Flash Memory. Most recently, he gave the invited talk, The Future of Memory: Challenges and Opportunities, at the Applied Materials Technical Symposium in Japan.
His pro bono work has included serving as an advisor on innovation to the White
House Council of Economic Advisors, teaching invited courses on Manufacturing
Economics and The Economics of the Internet at Stanford University, and serving
on the Board of Advisors to the Extension School at UC Berkeley. His work at UC
Berkeley helped the Extension School avoid wasting millions on capital acquisitions,
as well as developing the ‘Lifelong Learning’ strategy that was key in turning this
part of the University into a profit center for UC Berkeley.

ABOUT VLSIresearch
VLSIresearch is the leading provider of market research and economic analysis on
the technical, business, and economic aspects within high-technology industries.
The company is known for its unparalleled accuracy, innovation in market research,
and its sharply focused insight into industries covered. VLSIresearch’s primary databases and reports cover the semiconductor, flat panel display, PV cell and module
manufacturing, and associated high-technology industries.
Since the early eighties, VLSIresearch has been the only market research firm to
consistently gauge the direction of semiconductor markets. The company has always driven to deliver data and services with the most advanced methods, such as
the internet and dynamic databases. Its innovations include being the first to use
demand-based forecasting; the first to model semiconductor inventories in forecasts;
the first to use cost-of-ownership models for evaluating equipment; the first to perform customer satisfaction surveys; the first semiconductor industry research firm to
apply capacity models in forecasting.
VLSIresearch was founded in 1976, and has locations in Silicon Valley, the United
Kingdom, Japan, Taiwan and Korea.

Products and Services
Every day, top executives in high technology industries rely on VLSIresearch’s data
and information to guide them through business and technology decisions. VLSIresearch products are designed to provide real-time information to support your decisions. They include subscription services, multi-client studies, as well as consulting
and custom projects.

Subscription Services
VLSIresearch’s subscription services are focused on selected high-technology segments. They all share the objective of providing a real-time knowledge for your business and provide key statistics with frequent updates related to each sector. These
datasets are supported by timely in-depth analysis and VLSIresearch’s proprietary
research. The high-technology segments covered by VLSIresearch’s subscription
services include the semiconductor & electronics supply chain, semiconductor and
display manufacturing, and PV Solar manufacturing.

Multi-Client Studies
Multi-client studies typically provide annual analysis and survey of the market or topics
covered. They vary significantly and VLSIresearch constantly evaluates new projects
to be covered. VLSIresearch’s current recurring multi-client studies are the Customer
Satisfaction Survey, Test Consumable Market Reports, and PV Solar Fab Databases.

Consulting And Custom Projects
VLSIresearch’s team of analysts and consultants constantly engage with major companies and investors in various projects. The projects vary in scope from simple two
day engagments and presentations to multi-month complex industry reviews and
analyses. VLSIresearch can draw from the wealth of experience and expertise within
our team as well as from its international network of affiliates.

Notices, Terms of Use, Disclaimers, etc.
Copyright © 2013 by VLSI Research Inc. All rights reserved.
By accepting this report, opening it, or using it you are agreeing to these terms. This report contains valuable proprietary information developed or acquired by VLSI Research at great expense. You have a limited license to hold
these materials but do not become the owner of any material. The materials provided are protected by copyright,
trade secret, and trademark law. This report has been approved by VLSI Research for public release with attribution. The information in the materials may be used by you on a limited basis in your own documents, provided that
those documents are not-for-sale, VLSI Research is clearly referred to as the source of such information, and you
obtain written approval prior to use.

This report is provided on an “AS IS,” “WHERE IS”, “WHERE AVAILABLE”, “WITH ALL FAULTS” basis. VLSI
Research does not warrant these materials or the information provided therein, either expressly or impliedly, for any
particular purpose and VLSI Research specifically disclaims any express or implied warranties, including but not
limited to, any express or implied warranties of TITLE, ACCURACY, NON-INFRINGEMENT, MERCHANTABILITY or

The sources of the information in this report include numerous individual reports, memos and bulletins from various
segments of the industry, annual reports, financial reports, interviews, questionnaires, surveys, technical symposia,
trade journals, technical journals and individual assessments by knowledgeable company or industry representatives as well as our own analysis and judgment. Some companies are more cooperative about providing information than others and some companies decline to provide or validate the accuracy of any information. Although the
information provided is obtained or compiled from sources VLSI Research believes to be reliable given the oftentimes difficult circumstances under which it is collected, VLSI Research cannot and does not warrant or guarantee
the accuracy, validity, truthfulness, timeliness, or completeness of any information or data made available to you for
any particular purpose. In no event will VLSI Research be liable to you or any third party, whether in contract, tort
or under any other legal theory, for any direct, indirect, special, consequential or incidental damages, or any other
damages of any kind even if VLSI Research has been advised of the possibility thereof.

We receive letters and e-mails on current topics covered in our services and/or reports that are of interest to our
subscribers, as well as comments on our reports. We value that subscriber input and like to use it. By submitting such material to us, unless you tell us specifically not to publish it, or except to the extent that you give us an
embargo date before which you instruct us not to publish it, you authorize us to publish and republish it in any form
or medium, to edit it for style and length, and to comment upon or criticize it and to publish others’ comments or
criticisms concerning it, as the case may be.

This report may contain information concerning stocks that is obtained from the opinions of industry analysts.
Quoted past results are not necessarily indicative of future performance. None of the information should be seen
as a recommendation to buy or sell any securities. We are not stock analysts or investment advisors. You should
contact a registered investment advisor as to the nature, potential, value or suitability of any particular investment
action. No information provided is investment advice and any such information is just an opinion and is not tailored
to the investment needs of any specific person. Certain statements in this report, other than statements of historical fact, and other written or oral statements made by VLSI Research may be forward-looking. In some cases,
you can identify forward-looking statements by terminology such as “may”, “will”, “should”, “expects”, “intends”,
“plans”, “anticipates”, “believes”, “thinks”, “estimates”, “seeks”, “predicts”, “potential”, and similar expressions.

Although VLSI believes that these statements are based on reasonable assumptions, they are subject to numerous
factors, risks and uncertainties that could cause actual results and outcomes to be materially different from those
stated or projected. Those factors, among others, could cause actual results and outcomes to differ materially
from the results and outcomes stated or projected in, or implied by, the forward-looking statements. You should
understand that forward-looking statements are not guarantees of results or outcomes. New risks and uncertainties
arise from time to time, and VLSI Research cannot predict those events or how they may affect you, the reader.
VLSI Research Inc does not have any intention or obligation to update forward-looking statements after the date of
this report.
No part of this report may be used in any legal proceedings nor may any of these materials or the information contained therein be disclosed to any third party, including investors or affiliated firms belonging to investors, outside
directors or to your affiliated companies, or reproduced or transmitted to any third party, in any form or by any
means – mechanical, electronic, photocopying, duplication, microfilming, videotape, verbally or otherwise – without
the prior written permission of VLSI Research.

The Chip Insider®, The Industry Pulse Pro®, and the CSS 10 BEST logo are registered trademarks of VLSI Research Inc. All other trademarks, service marks, and logos are the property of their respective owners.

Disclosure Statement
Dan initiated this white paper after hearing Ajit present his views on a new foundry business model he called
Foundry 2.0. Dan’s intent was to identify if it was truly different, detail what these differences were, and identify how
GLOBALFOUNDRIES was executing this model differently from the traditional foundry model.

GLOBALFOUNDRIES has subscribed to VLSIresearch’s services since its inception and regularly interacts and
seeks advice from them via a retainer. Dan Hutcheson has known Ajit Manocha since he ran operations at Philips.
They have often shared knowledge and advice over the course of this relationship.