Enabling the next growth wave for semiconductors – a new approach to enable innovative startups

Tarun Verma, Partner Silicon Catalyst

November 11, 2015

MEPTEC role role MEPTEC •

Today’s innovation and startup dilemma dilemma startup and innovation Today’s •

Evolution of semiconductors semiconductors of Evolution •

Agenda

revenue revenue (more silicon) silicon) (more

semiconductor semiconductor applications applications

Increasing Increasing Expanding Expanding

$’s $’s

More R&D (innovation) (innovation) R&D More Lower cost/function cost/function Lower

cycle virtuous industry’s Semiconductor’s

... so everything must be going well? well? going be must everything so ...

Source: WSTS 2/15 2/15 WSTS Source: $335B! $335B!

... revenues record reached just Semiconductors

Semiconductors just reached record revenues ...

... but underlying trends are not good

Development costs Growth is slowing EBITDA is down are skyrocketing

41% Top 5 Semi 186 Other Semi

15% 1%

EBITDA (% of Rev) EBITDA (YoY)

-26%

Source: SIA, AlixPartners Source: 2012 Data, AlixPartners 2/14 Source: Semico 2013

structure chain supply the of Evolution

QPS

1960’s 2010’s

10

DISTRIBUTION: GROSS • RAEBELLO • WALLACE • RAEBELLO • GROSS DISTRIBUTION:

84 BUSINESS WEEK / AUGUST 25, 1997 25, AUGUST / WEEK BUSINESS *Acquired *Acquired

WAYNE MILLER/MAGNUM PHOTOS MILLER/MAGNUM WAYNE

Eli Harari Eli

Federico Faggin, Carver Mead Carver Faggin, Federico

* Systems SDA James Soloman James 1983 1983

Scott Sellers Scott Banatao Dado

cl Integration Scale

Synaptics

1986 1986

H

Gordon Campbell, Gordon Yara, Ron H

Wafer Wafer

3Dfx S3 1983 1983 1994 1994 1989

H

HH

four others four

Kamran Elahian Kamran

Bernard Vonderschmitt Bernard

James Diller, Diller, James lntWeb Planet 1996 Xilinx 1984 1984

Semiconductor H H Gunnar Wetlesen Wetlesen Gunnar Banatao, two others two Banatao,

Dan Floyd, Floyd, Dan

Gordon Campbell, Dado Campbell, Gordon

Sierra Sierra 1983 1983

sn-hn Wu Tsung-Ching

Jack Baletto, Baletto, Jack H

* Technologies &

Prakash Agarwal Prakash Ralph Ungermann Ralph

George Perlegos, Perlegos, George

Technology

Kamran Elahian, Kamran Federico Faggin, Federico

Chips Chips Atmel 1985 1985 1985

H * Zilog VLSI NeoMagic

E 1974 1974 1979 1993

Robert Dobkin Robert

H H H

Robert Swanson, Robert

Technology George Perlegos George

R. Barringer, six others six Barringer, R. others five Elahian, ran

Gordon Campbell, Campbell, Gordon

Linear Linear

1981 1981

G Robert Schreiner, Schreiner, Robert Kam- Hackworth, Michael of launching startups. launching of SEEQ H 1981 1981

E

* Synertek Logic Cirrus execs stay on board instead board on stay execs 1973 1973 1985

HH

ping $20 billion. Most Intel Most billion. $20 ping

digital chips. chips. digital

world, with revenues top- revenues with world,

. . ogr, oel Turiff Lowell Rodgers, J. T.

giant in analog and analog in giant

largest chip company in the in company chip largest Cypress

1983 1983

esesen to others two Weissenstern,

years, building it into a into it building years,

H Jack Schmidt, two others two Schmidt, Jack

processors, Intel is now the now is Intel processors,

Lionel Kattner, Mark Mark Kattner, Lionel

Sporck ran National for 24 for National ran Sporck

John Schroeder, Schroeder, John

The king of PC micro- PC of king The

David Allison, David James, David Allison, David

After leaving Fairchild, leaving After

* Microtechnology

Robert Noyce, Gordon Moore Gordon Noyce, Robert Semiconductor) Philips (now swing is up right now. right up is swing

Charles Sporck, two others two Sporck, Charles

Intel Computer * Signetics and-down rival to Intel. The Intel. to rival and-down 1968 1968 1968 1961

Semiconductor

HH H Fairchild to found this up- this found to Fairchild

National Flamboyant Sanders left Sanders Flamboyant 1967 1967

F

Micro Devices, and National. and Devices, Micro

..adr I,svnothers seven III, W.J.Sanders

including Intel, Advanced Advanced Intel, including AMD 1969 1969

E

Silicon Valley companies, Valley Silicon

semiconductor industry. industry. semiconductor

It spawned more than 30 30 than more spawned It

cadre for the West Coast West the for cadre

to work exclusively in silicon. in exclusively work to

Wilfred Corrigan Wilfred and became the founding the became and

child was the first company first the was child

LSI Logic LSI

1980 1980

erratic management style management erratic

E Eight” from Shockley, Fair- Shockley, from Eight”

because of Shockley’s Shockley’s of because

Founded by “The Traitorous “The by Founded

the technology. They left They technology. the

Eugene Kleiner Eugene Hoerni, Jay Last Jay Hoerni, East Coast labs to develop to labs Coast East

Caufield & Byers & Caufield Grinich, Julius Blank, Jean Blank, Julius Grinich, eight young men from men young eight

Kleiner, Robert Noyce, Victor Noyce, Robert Kleiner, tor, Shockley recruited Shockley tor,

Kleiner Perkins Kleiner 1972 1972

E Sheldon Roberts, Eugene Roberts, Sheldon Co-inventor of the transis- the of Co-inventor

(From left) Gordon Moore, Gordon left) (From

eight others eight

FAIRCHILD’S OFFSPRING FAIRCHILD’S

* Semiconductor

William Shockley, Shockley, William Lee Boysel, Jack Faith Jack Boysel, Lee

Fairchild Fairchild * Labs Shockley * Phase Four 1957 1957 1955 1969 1969

E E

od” as ely Tak t te school’s the to Thanks Kelley. says ford,” er Yn wr brd o er with tears to bored were Yang Jerry with Linda Himelstein in Menlo Park Menlo in Himelstein Linda with

ie os s fe idbe t Stan- to indebted feel I is jobs time ern ga suet Dvd hl and Fhilo David students grad neering y on ’ . aitn n ao Alto, Palo in Hamilton C. O’ Joan By

nc msl ot n wr to full- two work and out myself knock rie i 19, hr eetia engi- electrical where 1994, in trailer rather, Yahoo! rather,

i am mtr Safr. Te esn I reason “The Stanford. mater, alma his hc satd u i a rny campus grungy a in out started which Amen, say today’s entrepreneurs. Or entrepreneurs. today’s say Amen,

reality the is consolidation now …. startups

n t ser poutdsg porm at program product-design a steer to ing ue Itre drig ao! Inc., Yahoo! darling Internet tured longevity of the situation.” the of longevity

h’ help- he’s , When he’s not running running not he’s When rnusi. e ol nvr ae pic- have never could He preneurship. ucs, te tee s o ii t the to limit no is there “then success, IDEO

h Vle’ aaei/nuty relations. academic/industry Valley’s the el ae ae a 19ssye entre- 1990s-style at gaped have well eciig iio Vle’ eeg and energy Valley’s Silicon describing

in r, xmlfis ht os ih in right goes what exemplifies firm, sign o al i vso, en emn might Terman Dean vision, his all For e r tdy” emn rt i 1973, in wrote Terman today,” are we

and entrepreneurism for haven a once was Semiconductors nutil de- industrial head of Palo Alto’s Alto’s Palo of head rowed to almost nothing.” almost to rowed h patcs ht ae ae s what us made have that practices the IDEO

ai Kle, h successful the Kelley, David n rdcin” as ensy “a nar- “has Hennessy, says production,” in f 2 ilo. A ln a w maintain we as long “As million. $2 of N LIMIT.” “NO

Te a bten eerh n what’s and research between gap “The teach courses at Berkeley. at courses teach Stanford engineering school, with a gift a with school, engineering Stanford

uig H rtre t Safr i 1985. in Stanford to returned He puting. ai M Rst, o eape both example, for Rosati, M. Mario

the in professorship a endow to ever

Com- venture capital, and founded founded and capital, venture odih Rst prnr osn and Sonsini partner Rosati & Goodrich pee bcm te onet individuals youngest the became apiece, MIPS

ensy ok sbaia, asd some raised sabbatical, a took Hennessy hi proa tm. isn Sonsini Wilson time. personal their ag wrh oe hn 10 million $100 than more worth Yang,

hp i te al 18s Then, 1980s. early the in chips et Euly motn, lm donate alums important, Equally ment. n a ie o eun aos Fl and Filo favors. return to time a and

RISC ineig optrsine ok in work computer-science pioneering equip- and money of donor huge a came ucsfl 96 nta pbi offering— public initial 1996 successful

niern da, on ensy did Hennessy, John dean, engineering be- Clark, H. James professor neering atfrad w yas o hugely a to years two Fast-forward

hn aon hr. tnods current Stanford’s here. around thing engi- Stanford by started Inc., Graphics school cut the pair slack. slack. pair the cut school

t em lk nbd de js one just does nobody like seems It Silicon machine. motion perpetual Valley

engineering the goofing, digital the

of calculating brackets.” calculating of Silicon the of part is maters alma their e sts Rte ta cak on on down crack than Rather sites. Web

came a guy who made products instead products made who guy a came show alumni affection and loyalty The idesy aig it o ter favorite their of lists making mindlessly

business-cum-science orientation, “I be- “I orientation, business-cum-science Patterson. says company,” a start point hi PD hss Te bgn almost began They thesis. PhD their

X X

EUR-LA X INT XX XX ASIA 84 XX SVIC 33 SVIC SCT JMP

/ INT’L EURO/LA ASIA DOMESTIC DAD 02 EDIT PASS EDIT 02 K Y M C D: : : OK EDIT PE B

Chip-only ventures headed for extinction?

# of Seed/Series A Deals 54 53 42 37 32 28 22 22

4 4 7 6

Source: GSA • Long time-to-revenue • Relatively large investment required • Few IPO successes • Less investment from traditional sources

We are in the “post-chip” era

of ownership ownership of

is needed for sustainability & lower cost cost lower & sustainability for needed is Efficiency Energy ●

WEARABLES & SENSORS SENSORS & WEARABLES MEDICAL MEDICAL TRANSPORTATION TRANSPORTATION

through technology and logistics logistics and technology through Personalization ●

will cause upheaval in IT IT in upheaval cause will Cloud The ●

enterprise and personal personal and enterprise

concern at all levels: government, government, levels: all at concern Security Increase ●

WIRELESS WIRELESS

Next-Gen Wireless Wireless Next-Gen NEXT-GEN NEXT-GEN

Coverage and insatiable bandwidth needs will drive drive will needs bandwidth insatiable and Coverage ●

with ubiquitous RF RF ubiquitous with

will drive mobile processing at low power power low at processing mobile drive will Things of Internet ●

ENERGY ENERGY

drive compute power power compute drive

will continue to converge functions and and functions converge to continue will Computing Mobile ●

MEGA-TRENDS MEGA-TRENDS

silicon in solutions need all these

…. markets technology volume new driving are Megatrends

“We are entering the Era of Devices. This will be led by builders” builders” by led be will This Devices. of Era the entering are “We

revolution in sensors, all tied together by a cloud-like Internet of Things.” Things.” of Internet cloud-like a by together tied all sensors, in revolution

wearables, mobile health, autonomous cars, drones, 3D printing and a a and printing 3D drones, cars, autonomous health, mobile wearables,

cycle of hardware begins to assert itself in the form of watches, watches, of form the in itself assert to begins hardware of cycle

. But that era is ending, as the the as ending, is era that But . … Code of Era the in lived have “We

ever noticed a much longer, 20-year cycle in electronics.” electronics.” in cycle 20-year longer, much a noticed ever

“Most observers appreciate the Valley's four-year cycle, but few have have few but cycle, four-year Valley's the appreciate observers “Most

investments.” investments.”

lack of exciting new products and a decline of early-stage new new early-stage of decline a and products new exciting of lack … “

Devices” of “Era IoT the enter we as silicon in innovate to ever time best the is This

Source: I.B.S. 2011 2011 I.B.S. Source:

<$5M invested Total •

3000 3000 2500 2000 1500 1000 500 0

company semiconductor funded sourced crowd 1 •

st

.18u .18u

processor 50GFLOPS/Watt 1 •

st

.13u/.15u .13u/.15u

2012 2012

90nm 90nm

2013 2013

65nm 65nm 2014 2014

2015 2015

45/40nm 45/40nm

32/28nm 32/28nm

22/20nm 22/20nm

65nm & 45nm have yet to peak peak to yet have 45nm & 65nm

Frugally Done Be Can 28nm Even 130nm has the most design starts starts design most the has 130nm

typical or necessary not is it company, semiconductor a starting $100M spend can you While

incubate obtain

Short me to Hard to find Hard to

connecons influence agency small projects

access or course w/ control / horizon and terms of

Only $’s, not Cons Low $’s Only $’s May go off Partner Short term On metable Diluon

convince least 1) approach

services person to income feedback (at MVP diluve credibility

Just 1 Certainty of Small $’s Pros Network and Market Promotes Oen non- Network and

grants

NRE projects Family Capital Partner funding Programs &

Angel Bootstrap Friends and Venture Strategic Corp. Crowd Gov. Incubator

startups? semiconductor to available are alternatives What

What do semiconductor startups say they struggle with?

Rank Issues Rank Expenses Top 3 Challenges

$’s, EDA, Prototypes, Test

Source: Silicon Catalyst, 15 startups surveyed in 2014

Support Support equity common

Space Space

Pizza Pizza

& Legal Legal &

Work Work

Mentoring Mentoring

Lots of of Lots take will

Business Business Secure Secure

Catalyst Silicon

Contributions In-Kind funding, and

services goods, in

Prototyping Prototyping

Tools Tools Prototyping Prototyping

Equipment Equipment

~$2-3M for

Integration & & Integration

MPW MPW EDA

Lab Lab

DFM, System System DFM,

exchange In

fundraising time spending than rather … need they support and tools the with innovators supply will We

Silicon Catalyst minimizes need to raise seed capital and makes any cash investment more impactful

● Founders can get to work building a team and executing

● Founders retain more equity

● Cash goes to innovation, not table stakes

● Startup is more “fund-worthy”

Silicon Catalyst Overview

All All

Transportation Transportation

Analog Analog

Space Space Other Other

<28nm <28nm

8 8 MEMs MEMs

Bio Bio

Eur Eur 6 6 18 18

28-45nm 28-45nm

Industrial Industrial

23 23

31 31 CPU CPU

10 10 USA USA

65-130nm 65-130nm

Energy Energy 6 6

Memory Memory 4 4

CA CA

>130nm >130nm IoT IoT

11 11

6 6

17 17

2 2

Wearables Wearables

SV SV

ASIC ASIC

2 2 3 3

1 1

Mobile Mobile

Data Center Center Data

May May Apr Oct Oct Sept Aug July Jun

Sept Sept Aug July Jun May Apr Oct Oct 0 0

0 0

2 2

10 10

4 4

20 20

6 6 Selected Selected Selected Selected

Applied Applied

Applied Applied 8 8

30 30

Pipeline Pipeline

Pipeline Pipeline

10 10

40 40

12 12

50 50

14 14

Monthly Activity Activity Monthly Cumulative Activity Activity Cumulative

16 16 60 60

profile startup Catalyst Silicon

erentiation ff Di IP IP

pool profit larger into Taps Chip Chip Semiconductor

System System time-to-revenue Faster

Business Model Model Business solution Complete

Software / App App / Software

Silicon in Solutions Software Service Service

Silicon”? in “Solutions by mean we do What

Silicium’s Solution Can Save Space

Misfit Shine Fitbit Flex

The battery in most wearables takes up 50% of the volume! Silicium’s Full Solution For Wearables

Thermoelectric is integrated within the band What’s next: Optics in All Electronics

Get Photonics Into Mainstream Chips

As an As an analog interconnect technology technology

HPC, Datacenters Bulk CMOS IP in CPUs, GPUs, Sensors, chip-based Driverless $29B TAM DRAM Chip-to-Chip, FPGA LIDAR, Cars, IoT, Integrate on-chip with all electronics RF, A/D converters Mobile

harvesting Energy Wearables / IoT •

generation Odor Biomedical •

Vibration sensing Vibration

Industrial / Transportation •

Pressure sensing (dental) sensing Pressure

Wearables / Biomedical •

Location awareness Location

Wearables / IoT •

Human body monitoring: e.g., ECG e.g., monitoring: body Human

Biomedical •

Solutions Silicon create to MEMS using startups Many

enabling the next generation of startups. startups. of generation next the enabling

Great need and opportunity for the MEPTEC community to get involved in in involved get to community MEPTEC the for opportunity and need Great •

semiconductor design and manufacturing manufacturing and design semiconductor

Most entrepreneurs have limited experience in the backend aspects of of aspects backend the in experience limited have entrepreneurs Most •

ensure a viable product especially in SIP, MEMS MEMS SIP, in especially product viable a ensure

Most startups need upfront packaging and supply chain considerations to to considerations chain supply and packaging upfront need startups Most •

most semiconductor startups especially as they strive for solutions solutions for strive they as especially startups semiconductor most

Packaging is an increasingly critical component of the value proposition for for proposition value the of component critical increasingly an is Packaging •

help can MEPTEC How

Strategic partnerships to gain access and awareness to startup needs startup to awareness and access gain to partnerships Strategic • . .

Start up referrals referrals up Start •

In–kind support support In–kind •

Mentor and advisory roles roles advisory and Mentor •

Specifics

Silicon Catalyst’s partner contributions

EDA tool suite PDK access ATE solutions Lab test equipment Application support MPW Shuttle runs Test application support Test suite software

Design support services Test structures and process/ Design, visualization and Design Services, IP and Prototyping design analytics Animation software Supply chain management

Coming soon: more design services, IP, packaging, system prototyping, FPGAs

and revitalizing the semiconductor industry industry semiconductor the revitalizing and

another growth wave like the rise of the foundry/fabless foundry/fabless the of rise the like wave growth another

custom silicon and are considered “cool” …. launching launching …. “cool” considered are and silicon custom Small teams innovating meaningful solutions leveraging leveraging solutions meaningful innovating teams Small

decade next the for Vision

enabling this enabling MEPTEC community has a key role in in role key a has community MEPTEC

vision Catalyst Silicon