The Continuing Uneven Progress of Knowledge in Semiconductors: A Continuation of Bill Spencers remar

1
The Continuing Uneven Progress of Knowledge in
Semiconductors(A Continuation of Bill
Spencers remarks)

• Hank Chesbrough

2
Areas of Uneven Progress in Semiconductors

• User Requirements
• China
• Services

3
User Requirements The HDD Industry

• 1970s-early 1980s LSI starts to be used in HDDs
• Large quantities of chips, breadboards
• High infant failure rates
• Mixed signal, analog-digital chips
• Most staff are MEs
• Mid to late 1980s VLSI
• Improved reliability and functionality
• Many fewer devices, but more complex
• Getting the ASIC right on the first turn key to
  hit time to market window
• Some staff are now EEs and firmware engineers
• 1990s Advent of DSPs and microprocessors
• More functionality outsourced to chip vendor
• Vendor design tools now used to simply transfer
  of design to manufacturing
• Higher first time successes, faster time to
  market
• Software engineers now half of team

4
HDD Today

• Full digital simulation tools now exist
• Finite state modeling
• Entire disk drive, not just subsystems
• Further advances in density, time to market, and
  ability to utilize new materials
• Software engineers are now the bulk of the team
• With key functional roles for those who develop
  and maintain the simulation tools

5
However..

• Much of the value add of HDDs has leaked outside
  the industry
• To chip suppliers
• To other component suppliers (heads, media,
  motors)
• To computer customers
• One example the MicroDrive in the Apple iPod

6
The State of the Semiconductor Industry in China

• Engineers are cheap in China, while capital is
  scarce
• Disincentive for investment in leading tools
• Over-reliance on brute force design
• Mask development at 180nm, vs. 130 and 90
• Slow time to market
• Little improvement in design process for next
  time
• Like the semi industry in the US 20 years ago

7
The Value of IP in the Chinese semiconductor
industry

• IP is coming to be recognized as vital to the
  development of the domestic industry
• Shanghai Silicon IP Exchange
• ¾ owned by City of Shanghai, ¼ owned by Ministry
  of Industry and Information
• 2nd largest legal repository of semicondcutor IP
  in the world
• Tools, reference designs, cores
• Value to the Chinese industry higher
  productivity, faster time to market, lower device
  costs
• Value to Western IP owners broader market for
  IP, increased adoption of their tools and
  architecture
• In a market otherwise hard to serve

8
Semiconductor Services

• Increasingly, know-how is being bundled with
  products
• Applied equipment recipes
• Foundries Production reference designs and
  tools tuned to their specific processes
• EDA companies advanced design tools designs
  and IP to use them
• Customers use of this value added brings higher
  reliability and faster time to market
• Also creates lock-in to those suppliers

9
Business Model Shifts Toward Services

• IBM Semiconductor
• In 1960s and 1970s exclusively captive pdn
• 1980s 1990s beginning of IDM business
• Intel 486
• PowerPC
• Today an increase in services
• Foundry
• ODM

10
The State of Play in Services Industries Today

• In many leading companies, services are more than
  half of the companys revenue, and usually the
  fastest growing part
• IBM, GE, Xerox, GM
• Yet companies who sell services innovation
  offerings to corporate and government clients
  admit they lack a powerful conceptual model
  underneath their offerings
• Best practices, separated into vertical markets
• The shoemakers children are barefoot
• Paul Horns problem at IBM 6 billion of RD,
  but little of that money is advancing the
  services portion of IBMs business

11
The Impact of Academic Research on Industrial
Performance

• National Academy of Engineering Study (2003) 5
  sectors
• Networking and Communications
• Medical Devices and Equipment
• Aerospace
• Financial Services
• Transportation

12
The Impact of Academic Research on Industrial
Performance

• National Academy of Engineering Study (2003) 5
  sectors
• Networking and Communications
• Medical Devices and Equipment
• Aerospace
• Financial Services
• Transportation
• Found significant impact from academia on the
  first three, and much more limited impact on the
  last two sectors
• the academic research enterprise has not
  focused on or been organized to meet the needs of
  service businesses (p.8)
• DARPA gives more to consultants than academia in
  services
• NSF funding structure badly lags industry
  evolution
• Yet long term university-sited research is vital
  to long term industrial innovation performance

Product
Service
13
An Analogy The Emergence of Computer Science

• Mid-1940s computers used primarily for defense
  purposes, government is main customer
• Calculating ballistic missile trajectories
• Watson, Sr. a world market for maybe 5 computers
• 1946 first computer science class taught at
  Columbia in NYC
• IBM established Watson Scientific Computing
  Laboratory at Columbia in 1945
• Watson Sr. on the Board of Trustees for Columbia
• IBM co-taught first computer science classes with
  Columbia faculty
• 1950s computers actively embraced by
  universities
• But pursued in Engineering, Physics, or Math
  departments
• IBM played key role in convincing universities to
  treat computers as separate field of study
• The single strongest impulse for introducing
  computers on campuses in the mid-1950s did not
  come from the schools themselves or from any
  federal agency, but instead from IBM. (Aspray
  1994) emphasis added
• It was not until 1980s that the NSF gave the
  same institutional status to computer science as
  it confers upon traditional scientific
  disciplines such as physics, mathematics, or
  chemistry. (Aspray 2000)

14
Could Industry Again Stimulate the Emergence of
an Academic Field in Services Science?

• Extensive academic research already underway
• But in unconnected academic silos
• Services operations, Marketing services, IT
  services, etc.
• Analogous to early days of computer research
• No academic unit treats services as worthy of
  study on its own terms (vs. as an application of
  academic theory already developed from another
  field)
• Industry may be persuaded to take a leadership
  role again to catalyze (and fund!) the formation
  of the field of Services Science

15
What is Unique about Services (vs. Products) from
a theoretical viewpoint?

• the close interaction of supplier and customer
  (Fitzsimmons Fitzsimmons 2001
• intangible inputs used to create intangible
  outputs (Vargo Lusch, 2004)
• - the simultaneity of production and consumption
  (Sasser, Olsen, Wyckoff, 1978)
• - the combinations of knowledge into useful
  systems (Herzenberg, Alic, Wial, 1998)
• the decomposition of the exchange into business
  processes and experience points (Pine Gilmore,
  1999)
• the absence of artifacts (which complicate the
  transfer of tacit knowledge) (Nonaka and
  Takeuchi, 1995)

16
Toward a Deeper Structure for Services

• Today, services are vertically conceptualized and
  organized
• Health care, education, government, etc.
• Tomorrow, we may find horizontal structures that
  cut across services
• Data structures, document standards
• Business process modeling
• Business Models
• Customer-Supplier Interactions
• Customization vs. Re-use
• Tacit vs. Codified Information