2003: ahead to the past

11

Ahead to the Past!Ahead to the Past!Next-Gen, Old-Fashioned VentureNext-Gen, Old-Fashioned Venture

Vinod KhoslaVinod KhoslaKleiner Perkins Caufield & ByersKleiner Perkins Caufield & Byers

[email protected]@kpcb.com

© 2002 The New Yorker Collection from cartoonbank.com. All Rights Reserved.

22

The VC “LANDSCAPE” in 2000The VC “LANDSCAPE” in 2000

# of VC Firms in Existence

# of Professionals

# of First Time VC Funds Raised

# of VC Funds Raised This Year

VC Capital Raised This Year ($B)

Avg VC Fund Size Raised This Year ($M)

Source: NVCA Yearbook 2001; Venture Economics

1980

87

1035

24

57

2.08

36.5

1990

375

3794

14

82

3.20

39.0

2000

693

8368

164

497

105.05

211.4

33

The Illiquid BulgeThe Illiquid Bulge

From 1995-2000:From 1995-2000:

14,463

978

1,529

1,180

10,776

Companies funded

Went public

Were acquired

Went out of business

Remaining

Source: Venture Economics; Venture Source

--

-

44

The Beauty of Small NumbersThe Beauty of Small Numbers(And Board Seats)(And Board Seats)

Initi

al in

vest

men

t: $1

00k

to $

1M, b

oard Date

Jul-95Jun-96Apr-97Sep-98Oct-98Apr-99Jun-00Jan-01

Apr-95May-96Jul-96May-97Apr-98May-98Jul-98Mar-99Sep-99Nov-99Dec-99Dec-99Dec-99Jan-00Mar-00Aug-00Nov-00Jul-01

Company NameExcite (@home)JuniperCerent (Cisco)CorioSiara (Redback)AseraCentrataInfinera

Concentric (XO)ViantCybermediaExtreme NetworksSilicon Spice (Broadcom)CorvisLightera (Ciena)CoSineBBOZapletValiantZambeelOnFiberCoreonKymata (Alcatel)CenixSS8Kovio

Red = Board Seat

< $1M (n=8)> $1M (n=18)

On board (n=16)Not on board (n=10)

AggregateReturn*

192.7 x4.5 x

70.8 x8.0 x

SuccessRate*

100%67%

88%60%

* Numbers include current private companies at cost of fmv

Initi

al in

vest

men

t: $2

.5m

to $

16M

Conclusions– Attention matters– Board responsibility = better due diligence– Quick money makers don’t work– Too many passive deals last few years– One seed per year works for me

55

Capital

Funding to MilestonesFunding to Milestonesaka “Old-Fashioned Venture Capital”aka “Old-Fashioned Venture Capital”

Risk (ß)

Valuation

Idea isFeasible

TechnologyWorks

A CustomerBuys

SeedFunding

R&DCapital

Go-to-MarketCaptial

ExpansionCaptial

P(success) = 30%Req’d IRR = 100%




66

The “Fully Funded” FollyThe “Fully Funded” Folly

A CustomerBuys

FullyFund

IPO(……….pray……………….)

Capital

Risk (ß)

Valuation

Idea isFeasible

TechnologyWorks

77

A Generic Early 90’s ModelA Generic Early 90’s Model

Round Type Date

Amount Raised (MM)

Pre-Money Valuation

(MM) IRR Multiple

1 Seed Jan-90 $ 0.50 $ 2 101% 32.53

2 1st Jan-91 $ 3.00 $ 10 70% 8.13

3 2nd Jan-92 $ 8.00 $ 32 50% 3.30

4 3rd Jan-94 $ 13.50 $ 100 32% 1.32

5 IPO Jan-95 $ 150

Total Private Capital $ 25 Million

88

A Generic Late 90’s ModelA Generic Late 90’s Model

Round Type Date

Amount Raised (MM)

Pre-Money Valuation

(MM) IRR Multiple

1 Seed Jan-97 $ 5 $ 35 79% 18.37

2 1st Jan-98 $ 10 $ 100 65% 7.35

3 2nd Jan-99 $ 25 $ 200 59% 4.04

4 3rd Jan-00 $ 60 $ 600 52% 1.52

5 IPO Jan-01 $ 1000

Total Private Capital $ 100 Million

$200$200

??

99

Why “Ahead to the Past”?Why “Ahead to the Past”?

• We can’t make money with Early 90s IPO market We can’t make money with Early 90s IPO market and Late 90s Cash Burnsand Late 90s Cash Burns

• Some correction alreadySome correction already– Private ValuationsPrivate Valuations– Rent (in Silicon Valley)Rent (in Silicon Valley)

• Some corrections yet to occurSome corrections yet to occur– SalariesSalaries– FocusFocus– AttentionAttention

• VC BehaviorVC Behavior– Funding size?Funding size?– Corrections for the past bubbleCorrections for the past bubble– Entrepreneurial incentivesEntrepreneurial incentives

1010

The Committed Capital BubbleThe Committed Capital Bubble

Am

ou

nt

Investe

d

($B

)

1995 1996 1997 1998 1999 2000 2001

$7.6$12.2

$17.0

$29.8

$50.5

$90.1

$48.2

$6.8 $9.8 $12.8$17.6

$47.2

$91.6

$32.1

$0

$20

$40

$60

$80

$100

Investment Fundraising

Source: VentureOne

FFundraising vs. Amount Investedundraising vs. Amount Invested

1111


Source: VentureOne

0

5

10

15

20

25

30

35

40

1995 1996 1997 1998 1999 2000 2001

UninvestedVenture Capital

Total Over-committed Capital for all Private Equity = $100BTotal Over-committed Capital for all Private Equity = $100B

Accu

mu

late

d C

ap

ital

Over-

com

mit

men

ts (

$B

)

1212


0

5

10

15

20

25

30

35

40

1995 1996 1997 1998 1999 2000 2001

0

1

2

3

4

5

6Uninvested VentureCapital

Years of UninvestedCapital

Years of UninvestedCapital at 1995Investment Pace

Source: VentureOne

Years

Accu

mu

late

d C

ap

ital

Over-

com

mit

men

ts (

$B

)

1313

The VC Firm ModelThe VC Firm Model

• 1.5-2 ventures/partner/year1.5-2 ventures/partner/year

• $8-10M invested per venture (3 rounds)$8-10M invested per venture (3 rounds)

• New fund every 2 yearsNew fund every 2 years

• $12-20M per partner per year$12-20M per partner per year

• $24-40M per partner per fund$24-40M per partner per fund

1414

2001 IPO Companies Are More 2001 IPO Companies Are More MatureMature

Years

4.5

3.12.92.8

3.13.1

4.14

0

1

2

3

4

5

1994 1995 1996 1997 1998 1999 2000 2001

Source: VentureOne

Time From Initial Equity Funding to IPO Time From Initial Equity Funding to IPO

1515

What about the Telecom What about the Telecom “Blowout”?“Blowout”?

CategorySystems (Electronic)

Systems (Optical)

Services

Software

Systems (Electronic)

Systems (Optical)ServicesComponents

Software

Return IF HELD 300x14x228x*3.6xToo early12x*Too early0.3x

0.9x14xToo early12x*0x0.7xToo early12x*0xToo early

CompanyJuniperSiaraCerentCorvisInfineraConcentricOnFiberCoreon

CoSineExtremeSS8LighteraValiantKymataCenixSilicon SpiceIPVerseSigma

On

th

e b

oa

rdN

OT

on

th

e b

oa

rd

* Valued at time of acquisition

1616

Case Study: OnFiberCase Study: OnFiber

• Management team saw a need to cut cash burn in Management team saw a need to cut cash burn in January 2001January 2001

• Reduced burn from $5M plan to $1-1.5M per monthReduced burn from $5M plan to $1-1.5M per month• Turned down capital from investors pushing “get big Turned down capital from investors pushing “get big

fast” strategyfast” strategy• Got out of obligations earlyGot out of obligations early• Used 50% IRR hurdle on all new projectsUsed 50% IRR hurdle on all new projects• Focused on getting critical mass in 6 metros rather than Focused on getting critical mass in 6 metros rather than

2626• OnFiber Team’s Results:OnFiber Team’s Results:

– Current burn = <$1m/monthCurrent burn = <$1m/month– $30M in the bank, no debt$30M in the bank, no debt– Acquiring assets for pennies on the dollarAcquiring assets for pennies on the dollar

1717

Bottom-Fishing Rarely WorksBottom-Fishing Rarely Works

Old joke about a restaurant:Old joke about a restaurant:

““The food is horrible… The food is horrible… but it’s all-you-can-eat!”but it’s all-you-can-eat!”

• Too many “recaps” of uninteresting companiesToo many “recaps” of uninteresting companies• ““Mercy killing” can be a good thingMercy killing” can be a good thing• The good news: a handful of high-valuation The good news: a handful of high-valuation

private companies indicates quality is still being private companies indicates quality is still being chased.chased.

1818

Dot-Com Profitability: Less is MoreDot-Com Profitability: Less is More

• An Empirical Fact: Good financial results out of An Empirical Fact: Good financial results out of several consumer-focused internet-based several consumer-focused internet-based companiescompanies

• Positive growth and EBITDAPositive growth and EBITDA• AmazonAmazon• Blue NileBlue Nile• eHealthInsuranceeHealthInsurance• AutotraderAutotrader• GoogleGoogle

1919

Wrong Goals: Time-to-IPO?Wrong Goals: Time-to-IPO?

No optical systems boards from mid ’98 to 2001

…

2020

Avoid the “Conventions”Avoid the “Conventions”

• Profits in telecom: Invest early & significantlyProfits in telecom: Invest early & significantly

• Less can be more - Small is BeautifulLess can be more - Small is Beautiful

• Attention (and board seats) matterAttention (and board seats) matter

• Bottom fishing – does it work?Bottom fishing – does it work?

• Mercy-kill the uninteresting companies!Mercy-kill the uninteresting companies!

• ““HOT” is NOT “good”: Dotcom, Telecom, Nano?HOT” is NOT “good”: Dotcom, Telecom, Nano?

• IPO is NOT a “goal” : building a company is!IPO is NOT a “goal” : building a company is!

• Stock Market is not Validation: CorioStock Market is not Validation: Corio

• Need to bleed off the oversupply of capitalNeed to bleed off the oversupply of capital

2121

How Did We Get Here?How Did We Get Here?The Past Five YearsThe Past Five Years

• Fundamental OpportunityFundamental Opportunity

• GreedGreed

• Funding the “ight-Lay”sFunding the “ight-Lay”s

• Irrational Exuberance : IPO “goal” Irrational Exuberance : IPO “goal”

• Escalating Expectations & Hype: The Khosla BubbleEscalating Expectations & Hype: The Khosla Bubble

• FearFear

• Dislocated Business PlansDislocated Business Plans

• Flight to Quality, Defensibility, Economic Contributions?Flight to Quality, Defensibility, Economic Contributions?

2222

Many Companies Will Not Keep UpMany Companies Will Not Keep Up

LucentLucentNortelNortelCiscoCiscoTellabsTellabs

AlcatelAlcatelAT&TAT&TSprintSprintWorldcomWorldcom

Leading Communications Leading Communications Vendors/CarriersVendors/Carriers

Leading CommunicationsLeading CommunicationsVendors/Carriers (Next Generation)Vendors/Carriers (Next Generation)

??

IBMIBMH-PH-PData Data GeneralGeneralDECDEC

SperrySperryUnivacUnivacWangWangCrayCray

Leading Computer Vendors Leading Computer Vendors (Last Generation)(Last Generation)

IBMIBMSunSunDellDellCompaqCompaqH-PH-P

Leading Computer Vendors Leading Computer Vendors (Current)(Current)

2323

Old vs. New in theOld vs. New in theMainframe to PC TransitionMainframe to PC Transition

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

Microsoft

IntelCompaq

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

01/3

0/87

07/3

1/87

01/2

9/88

07/2

9/88

01/3

1/89

07/3

1/89

01/3

1/90

07/3

1/90

01/3

1/91

07/3

1/91

01/3

1/92

07/3

1/92

01/2

9/93

07/3

0/93

Digital Equipment Corp.

NCRAmdahl

Data GeneralCapital fled legacy systems

New winners emerged

Source: Thomas Weisel Partners

This time the stakes

are larger!

2424

The Investors DilemmaThe Investors Dilemma

• Expect more losers than winnersExpect more losers than winners

• Many will lose everythingMany will lose everything

• More will be won than lostMore will be won than lost

• The value of winners will exceed the The value of winners will exceed the

cumulative market cap of the entire cumulative market cap of the entire

sector todaysector today

• M&A Game is hard to predictM&A Game is hard to predict

2525

5 years out, the group’s market cap has grown…

Winners Take (Almost) AllWinners Take (Almost) All

Industry Structure Today

But leaders far exceed the also-rans

2626

Fundamentals of DemandFundamentals of Demand

• The EnterpriseThe Enterprise– Carrier services purchase decisions based on Carrier services purchase decisions based on

payback periodpayback period– Paybacks< 1 yr => Continued demandPaybacks< 1 yr => Continued demand– 1% increase in IT = 1.5-2% decrease in G&A1% increase in IT = 1.5-2% decrease in G&A– IT spending increased from 0.5% to 3.5% of sales IT spending increased from 0.5% to 3.5% of sales

in 90’sin 90’s

• The Service ProviderThe Service Provider– Enterprise demand => need for equipment Enterprise demand => need for equipment

purchasespurchases– Consumption of bandwidth + shifting computing Consumption of bandwidth + shifting computing

paradigm makes for new revenue opportunitiesparadigm makes for new revenue opportunities– Systems for old revenue don’t support new Systems for old revenue don’t support new

revenue & cost points =>new equipment purchasesrevenue & cost points =>new equipment purchases

2727

Partial View: New Areas for Partial View: New Areas for InvestmentInvestment

• MaterialsMaterials– But only with an economic advantage in an application But only with an economic advantage in an application

that mattersthat matters– Beware the Nano bubble!Beware the Nano bubble!

• Probably more companies named “Nano______” in 2002 Probably more companies named “Nano______” in 2002 than named “_____Light” in 2000than named “_____Light” in 2000

• Information Tech: removing the COWInformation Tech: removing the COW

• Personalized MedicinePersonalized Medicine

• ………….???.???

2828

The HALF-EMPTY viewThe HALF-EMPTY view

• Technology led slowdownTechnology led slowdown• Consumer Reaction pile-onConsumer Reaction pile-on• Sept 11 Sept 11 • The WARThe WAR• 2002 is history!2002 is history!• Financings UnavailableFinancings Unavailable• Stay afloat, tread water, hunker downStay afloat, tread water, hunker down

2929

The HALF-FULL viewThe HALF-FULL view

• Less competitionLess competition• Time to develop technologyTime to develop technology• Focus: sustainable advantage, Focus: sustainable advantage,

value-addvalue-add• Better critique, Better ventures, Better critique, Better ventures,

Higher BarHigher Bar• Longer Term View/HorizonLonger Term View/Horizon• Lower Funding => Lower RiskLower Funding => Lower Risk

3030

RECOIL: The New WorldRECOIL: The New World

• Reaction to the excessesReaction to the excesses

• Unskilled, Unthinking “eBusiness” investmentUnskilled, Unthinking “eBusiness” investment

• Skepticism & Negativism: Prolonged CycleSkepticism & Negativism: Prolonged Cycle

• VC Euphoria NOTVC Euphoria NOT

• Hard ROI – retrospective vs. prospective Hard ROI – retrospective vs. prospective

savingssavings

• Lack of Human CapitalLack of Human Capital

3131

The Weather Forecast …The Weather Forecast …

• Rate of change will accelerate - life will Rate of change will accelerate - life will be more complex, busier . . . be more complex, busier . . .

• Adaptability, agility & momentum will be Adaptability, agility & momentum will be the key to success!the key to success!

• Innovation, opportunities & Innovation, opportunities & entrepreneurship will thriveentrepreneurship will thrive

• Disruption will be the order of the dayDisruption will be the order of the day• Fun, fortunes & failure will be in Fun, fortunes & failure will be in

abundanceabundance

3232

Wake Up CallWake Up Call

• What we take for GrantedWhat we take for Granted• Irrelevance & the “other “ thingsIrrelevance & the “other “ things

– Relationships: Friends & FamilyRelationships: Friends & Family– Satisfaction, Enjoyment & ContentmentSatisfaction, Enjoyment & Contentment– Feeling valuable: making a differenceFeeling valuable: making a difference

• Value SystemValue System• Books: Don’t Sweat The Small StuffBooks: Don’t Sweat The Small Stuff

3434

Tomorrow’s Markets…Tomorrow’s Markets…

3535

(ATM/FR)

(VOIP)

(IP/Bus Class IP)

(Broadcast)

The New IPInfrastructure

Network Issues & Trends Network Issues & Trends

• Everything over IPEverything over IP• Two layer networkTwo layer network• Value added services Value added services • Internet based SW architectureInternet based SW architecture• Skills shortageSkills shortage• Mission critical technologyMission critical technology• Rapid, unpredictable growthRapid, unpredictable growth• Legacy encapsulationLegacy encapsulation

3636

IT Issues & TrendsIT Issues & Trends

• Skills shortageSkills shortage• Legacy architectures & “islands” vs dynamic Legacy architectures & “islands” vs dynamic

architecture needsarchitecture needs• Mission Critical TechnologyMission Critical Technology• Changing applications mixChanging applications mix• Systematic productivity : The Real-Time Systematic productivity : The Real-Time

EnterpriseEnterprise• Random acts of productivity: everyday processesRandom acts of productivity: everyday processes• Technology based competitionTechnology based competition• Operations CostOperations Cost

3737

Results: Market OpportunitiesResults: Market Opportunities

• Outsourcing & growth => a new (datacenter) infrastructureOutsourcing & growth => a new (datacenter) infrastructure

• Value-added Services => new programmable networkValue-added Services => new programmable network

• Operations cost & Complexity => OMAP focusOperations cost & Complexity => OMAP focus

• A new IT architecture => “Real Time” information baseA new IT architecture => “Real Time” information base

• Skills shortage & complexity => ASP, servicesSkills shortage & complexity => ASP, services

• Dynamic demand => Compute utilitiesDynamic demand => Compute utilities

• IT going from 0.5% to 3.5% to ?% of salesIT going from 0.5% to 3.5% to ?% of sales

3838

““Real Time” EnterpriseReal Time” Enterprise“Ciscoize” and “Dellize” Every Business“Ciscoize” and “Dellize” Every Business

• Adaptive architecture, evolvable applicationsAdaptive architecture, evolvable applications• Configuration NOT customizationConfiguration NOT customization• Federation NOT integrationFederation NOT integration• Architecture to connect architecturesArchitecture to connect architectures• Rapid , incremental implementationRapid , incremental implementation• Beyond “database” to an “information base”Beyond “database” to an “information base”• InstantaneousInstantaneous “financials”, metrics, supply “financials”, metrics, supply

chain, customer support .…chain, customer support .…

“Spontaneous transaction flow and information transparency throughout the extended enterprise”

3939

Operations SystemsOperations Systems

• Cost of Operations: Cost of Operations: start/add/change/delete/operate/updatestart/add/change/delete/operate/update

• Fluid Resource allocationFluid Resource allocation• Fluid “service” provisioningFluid “service” provisioning• Hard ROI: Hard ROI: Cost savings & Revenue GenerationCost savings & Revenue Generation

• Total Cost of OwnershipTotal Cost of Ownership

4040

New network operations systems must be designed for adaptability and change (new equipment, multiple vendors, new service offerings/ provisioning).

1970-19901970-1990 Minicomputer/Minicomputer/DEC eraDEC era

1:1 people/machine ratio1:1 people/machine ratio

1980-1980- Workstation/Workstation/PC eraPC era


1990-1990- Enterprise networks/Enterprise networks/Cisco eraCisco era

1:100? people/machine ratio1:100? people/machine ratio

2000-2000- Broadband packet Broadband packet networks ?networks ?

1:1000? people/machine ratio1:1000? people/machine ratio

1960-19801960-1980

TimeTime

Mainframe/Mainframe/IBM eraIBM era

Platform(s)Platform(s)


Network Operations ModelNetwork Operations Model

Source: Paul JohnsonSource: Paul Johnson

Old network management systems were single vendor solutions optimized for cost in rigid five-year preplanned networks.

One Analyst’s Explanation:One Analyst’s Explanation:

4141

New Area: “Virtual Computer” New Area: “Virtual Computer”

• Networks of computers as the “Virtual Computer”Networks of computers as the “Virtual Computer”• Scalability of hardware - add & deleteScalability of hardware - add & delete• Self managementSelf management• Geographic distributionGeographic distribution• Load balancing, caching, COS, … servicesLoad balancing, caching, COS, … services• ResilienceResilience• ““Network operating system” Network operating system”

SETI, Napster, RoutersSETI, Napster, Routers

A Computer Distributed Over the InternetA Computer Distributed Over the Internet

4242

Case Study: Router NetworksCase Study: Router Networks

• Behave as “one” machineBehave as “one” machine

• System self-adjusts to “node” failuresSystem self-adjusts to “node” failures

• Capacity can be added/deleted -Capacity can be added/deleted -“self organizing”“self organizing”

• Geographically disbursedGeographically disbursed

• Managed failure modesManaged failure modes

4343

New Areas: Network ServicesNew Areas: Network Services

• Storage servicesStorage services• Database servicesDatabase services• Web servers/HTTP serversWeb servers/HTTP servers• TCP/IP session serversTCP/IP session servers• Application serversApplication servers• Composite servicesComposite services

– ReplicationReplication– Load balancingLoad balancing– DistributionDistribution

The “Decomposed” Computer ArchitectureThe “Decomposed” Computer Architecture

4444

Other…Other…

• Remote Services – multi-trillion global marketRemote Services – multi-trillion global market

• Collaboration –ZapletCollaboration –Zaplet

• NanoNano

4545

Investment ConclusionsInvestment Conclusions

• Internet is changing the model of computingInternet is changing the model of computing

• Software architecture of the internet will become the Software architecture of the internet will become the software architecture for the enterprisesoftware architecture for the enterprise

• Computing infrastructure (datacenter) will be the next Computing infrastructure (datacenter) will be the next surprise; fueled by “ASP Services”, outsourced “compute surprise; fueled by “ASP Services”, outsourced “compute tone”, and internet “geoscale”tone”, and internet “geoscale”

• The “real-time enterprise” will drive carrier demand at The “real-time enterprise” will drive carrier demand at increasing rates of growth & create Oracle/SAP scale increasing rates of growth & create Oracle/SAP scale companiescompanies

• Operations systems to reduce operating cost & total cost of Operations systems to reduce operating cost & total cost of ownership will be a major issue/opportunityownership will be a major issue/opportunity

• Evernet – everyplace, every time, every device thru Evernet – everyplace, every time, every device thru “mobile IP” (but not a lot of money for new ventures)“mobile IP” (but not a lot of money for new ventures)

4646

Fundamental Contributors: Fundamental Contributors: Investment AreasInvestment Areas

• SemiconductorsSemiconductors

• Software in hot boxes Software in hot boxes

• Physics / Optical /NanotechnologyPhysics / Optical /Nanotechnology

• ASP ASP

• Nexgen telecom servicesNexgen telecom services

• Networks: the programmable networkNetworks: the programmable network

• Real Time EnterpriseReal Time Enterprise

• Infrastructure – revamping datacenter Infrastructure – revamping datacenter architectures architectures

• Operations SystemsOperations Systems

4747

But….But….

• Nano is approaching bubble statusNano is approaching bubble status• Storage has it’s “Ightlay’s”Storage has it’s “Ightlay’s”

4949

The Storwidth Tsunami ?The Storwidth Tsunami ?

Vinod KhoslaVinod Khosla

Kleiner Perkins Caufield & ByersKleiner Perkins Caufield & Byers

vkhoslavkhosla@@kpcbkpcb.com.com

March, 2002March, 2002

5050

Hot New Startup: OrstayHot New Startup: Orstay

•Orstay is pioneering a new breed of network storage server that offers unlimited scalability and industry-leading performance in a system that is far easier and more cost-effective to manage than today's solutions

•By offering users the performance and functionality of a Storage Area Network (SAN) with the simplicity and ease of Network Attached Storage (NAS). Our product is an integrated hardware and software solution that operates over standard Ethernet networks and supports both block and file transfers

•Our product team possess a unique understanding of the requirements of the data center. In addition, we understand today's growing need for products that offer high performance in the data center with a focus on high availability, scalability and interoperability.

5151

Hot Startup:Hot Startup:

• ight-Lay is revolutionizing the future of ight-Lay is revolutionizing the future of telecommunications by enabling high-bandwidth, low-cost telecommunications by enabling high-bandwidth, low-cost optical MANoptical MAN infrastructures that are infrastructures that are robust and scalablerobust and scalable enough to handle data and video traffic, while also enough to handle data and video traffic, while also supporting legacy voice. supporting legacy voice.

• ight-Lay combines the ight-Lay combines the intelligence and switchingintelligence and switching capability of a data network with the reliability and capability of a data network with the reliability and protection performance of SONET. The result is protection performance of SONET. The result is multiservice access and transportmultiservice access and transport with unprecedented with unprecedented scalabilityscalability, , provisioningprovisioning flexibility and flexibility and service intelligenceservice intelligence. .

• With a significantly With a significantly smaller capital outlaysmaller capital outlay than competing than competing products and products and fewer management requirementsfewer management requirements than other than other optical networking products, ight-Lay delivers the solution optical networking products, ight-Lay delivers the solution carriers need for the new optical age.carriers need for the new optical age.

5252

The Saga of Ight-LayThe Saga of Ight-Lay

• In 2000 at an “optical investment In 2000 at an “optical investment conference”….conference”….

• Took one line from the website of each Took one line from the website of each metro optical company presentingmetro optical company presenting

• Concatenated them as “elevator pitch” Concatenated them as “elevator pitch” for hot startup “Ight-Lay”for hot startup “Ight-Lay”

• Put presentation on KPCB websitePut presentation on KPCB website• Got plenty of resumes and investment Got plenty of resumes and investment

interestinterest

5353

How to Lose a Lot of MoneyHow to Lose a Lot of Money

How I turned $100,000 in

Stock into $980 in Cash

in Just Six Months!

By I.M. Broke

5454

Being Right is InsufficientBeing Right is Insufficient

Be Wrong Be Right

Buy low,Sell low

Buy low,Sell high

Join Bandwagon

BeContrarian

Buy high,Sell low

Buy high,Sell high

5555

Avoiding the BubbleAvoiding the Bubble

Be Wrong Be Right

Buy low,Sell low

Buy low,Sell high

Join Bandwagon

BeContrarian

Buy high,Sell low

Buy high,Sell high

Is Storage Here??

5656

5 years out, the group’s market cap has grown…

Even in HOT Areas …..Winners Take (Almost) AllEven in HOT Areas …..Winners Take (Almost) All

Industry Structure Today

But leaders far exceed the also-rans

5757

Avoiding the bandwagonAvoiding the bandwagon

• What assumptions are others making?What assumptions are others making?

• Are there scenarios where those assumptions Are there scenarios where those assumptions

are wrong?are wrong?

• Cerent vs. IghtlayCerent vs. Ightlay

• Brocade vs. OrstayBrocade vs. Orstay

• Significant technical contributionSignificant technical contribution

• Enabling a significant new capabilityEnabling a significant new capability

5858

More Recent AssumptionsMore Recent Assumptions

• All Optical Networks, higher speeds are All Optical Networks, higher speeds are cheapercheaper

• No “vision” before it’s time: MEMS, 40GNo “vision” before it’s time: MEMS, 40G• Scalability is more important than costScalability is more important than cost

– In opticsIn optics– In storageIn storage

• Storewidth: performance, SAN,Storewidth: performance, SAN,• Greenfield NetworksGreenfield Networks• Cost of capitalCost of capital

5959

Optical Hype Versus RealityOptical Hype Versus Reality

• Optics will transform the Optics will transform the transport layertransport layer ONLY ONLY

• Since packet switching is necessary for any-any Since packet switching is necessary for any-any

interactive connectivity, an all-optical network interactive connectivity, an all-optical network

remains an optical illusionremains an optical illusion

• Value added services will be done electronicallyValue added services will be done electronically

• DWDM translate from long haul to metro?DWDM translate from long haul to metro?

6060

Storage Hype Versus RealityStorage Hype Versus Reality

• Endless need and limitless opportunitiesEndless need and limitless opportunities

• Declining cost of storage: Declining cost of storage: only in raw bits

• Big silicon, hot boxes, …Big silicon, hot boxes, …

• SAN vs NAS SAN vs NAS

• The big COWThe big COW

• Revolution or Evolution: Revolution or Evolution: Insertion strategies matter

6161

The Good News: Opportunities ExistThe Good News: Opportunities Exist

• The COW•Insertion of Technology•Googlizing

6262

Watch the COWWatch the COW

COW = Cost of Ownership

6363

COW Parts:COW Parts:

Operational CostsAnnual*

$.08-$.10/MB

*Assumes one Sys Admin per 5 TB storage managed 20% Maintenance cost $1000/month std rack hosting cost

4 yr Cost of Storage: $.45-.50/MB!

Filet Mignion: Operational Costs make up 80% of total cost

(4X cost of initial investment)Baby Back Ribs: Not everybody

needs a steak (PFSB)

System Software$.08-$.10/MB

$. 025 Annual amortz.

Raw Storage Costs

$.01-.005/MB

6464

Filet MignionFilet Mignion: Zambeel: Zambeel

Zambeel’s Zambeel’s significant new technology significant new technology contribution attacks the priciest portion of the contribution attacks the priciest portion of the COWCOW

With 4 way independently scalable NASWith 4 way independently scalable NAS

– Dynamic Capacity Allocation; Zero system downtime Dynamic Capacity Allocation; Zero system downtime when changing capacity of user or systemwhen changing capacity of user or system

– Fine grain and fluid assignment of throughput Fine grain and fluid assignment of throughput performance at user or file levelperformance at user or file level

– NFS file system to 100’s terabytesNFS file system to 100’s terabytes– 100’s of simultaneous domains on same FS100’s of simultaneous domains on same FS

6565

Baby Back RibsBaby Back Ribs: : Personal File System BackupPersonal File System Backup

• Personal File System Backup Personal File System Backup (low QOS!)(low QOS!)

– Super Low Cost, low performance , High Latency Back Super Low Cost, low performance , High Latency Back UpUp

– Seamlessly convert data in repository from Windows Seamlessly convert data in repository from Windows FS to a more scalable FSFS to a more scalable FS

– Snap Shot and delta Backup capability to reduce Snap Shot and delta Backup capability to reduce backup processor/bandwidth requirementsbackup processor/bandwidth requirements

– Management/Administration layer to handle Management/Administration layer to handle compilation of 1000’s of desktops and provide search compilation of 1000’s of desktops and provide search capability. capability.

6666

Insertion StrategyInsertion Strategy

• More important than ever!More important than ever!

• Little customer urgency for new Little customer urgency for new

technologytechnology

• Increased focus on COW, not just Increased focus on COW, not just

equipment/software costsequipment/software costs

• Risk MitigationRisk Mitigation

6767

““Googlizing” the compute infrastructureGooglizing” the compute infrastructure

• Cheap PCs in parallel being a better outcome Cheap PCs in parallel being a better outcome than “expensive, specialized hardware” than “expensive, specialized hardware” – Grid computingGrid computing– Low cost unitsLow cost units

• Why does that work for Google but Sun, Why does that work for Google but Sun, NetApp, others are selling real product to smart NetApp, others are selling real product to smart people?people?– Why do enterprises not want “revolution”Why do enterprises not want “revolution”– Linux, GE, … goal or a tactic?Linux, GE, … goal or a tactic?

6868

Reliability - A New ApproachReliability - A New Approach

• The shuttle Challenger: designed not to failThe shuttle Challenger: designed not to fail

• Biological systems: designed to fail Biological systems: designed to fail

gracefullygracefully

• Complex systems: “evolutionary approach”Complex systems: “evolutionary approach”

6969

Current Financial Market DynamicsCurrent Financial Market Dynamics

• Funding the “Or-stays” and “ight-Lay”sFunding the “Or-stays” and “ight-Lay”s

• Hype or reality – real markets? Hype or reality – real markets? (Incubators, exchanges)(Incubators, exchanges)

• Discrimination & typecasting Discrimination & typecasting (USIX, CRIO)(USIX, CRIO)

• Fear vs Greed ; Optimism vs. PessimismFear vs Greed ; Optimism vs. Pessimism

• Capital Supply – perception or reality? (CLECS)Capital Supply – perception or reality? (CLECS)

• Analyst PhenomenonAnalyst Phenomenon– Escalating expectations : Guaranteed disappointmentEscalating expectations : Guaranteed disappointment

– Whisper numbers : need for a new rationalityWhisper numbers : need for a new rationality

• Khosla Bubble : luck, values, belief system…Khosla Bubble : luck, values, belief system…

7070

Valuation RationalizationValuation RationalizationKeep changing the denominator…Keep changing the denominator…

P/E

P/E/G

P/S

P/S/G

P/AM*

P/Cash

Time

Mar

ket

Val

ue

*Price to Addressable Market

7171

Press Release BubblePress Release Bubble#

No

n-F

inan

cial

Pre

ss R

ele

ase

s(3

mon

th r

ollin

g av

erag

e)

Source: Archived press releases from website of a public optical networking company

0

1

2

3

4

5

Feb-9

9

May

-99

Aug-9

9

Nov-9

9

Feb-0

0

May

-00

Aug-0

0

Nov-0

0

1-Feb

1-M

ay

1-Aug

1-Nov

2-Feb

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Press Releases

NASDAQ

7272

Investor’s Weather ForecastInvestor’s Weather Forecast

7373

How Did We Get Here?How Did We Get Here?The Past Five YearsThe Past Five Years

• Fundamental OpportunityFundamental Opportunity

• GreedGreed

• Irrational Exuberance : IPO “goal” Irrational Exuberance : IPO “goal”

• Escalating ExpectationsEscalating Expectations

• FearFear

• Dislocated Business PlansDislocated Business Plans

• Flight to Quality, Defensibility, Economic Flight to Quality, Defensibility, Economic Contributions?Contributions?

7474

New Technology has a HistoryNew Technology has a History

0

1,000,000

2,000,000

3,000,000

4,000,000

5,000,000

6,000,000

7,000,000

Co

nn

ecti

on

s in

Lo

nd

on

90 93 95 99

Source: When Old Technologies Were New

“In 1885, Yale students who were getting ‘more light than they relished’ chopped down an electric pole

erected at the corner of the campus…”

7575

Walter Gifford, future President of AT&T, 1921

……We Can’t Predict What Will Happen.We Can’t Predict What Will Happen.

• ““Nobody knew early in 1921 where radio Nobody knew early in 1921 where radio was really headed. Everything about was really headed. Everything about broadcasting was uncertain. For my broadcasting was uncertain. For my own part I expected that since it was a own part I expected that since it was a form of telephony, and since we were in form of telephony, and since we were in the business of furnishing wires for the business of furnishing wires for telephony, we were sure to be involved telephony, we were sure to be involved in broadcasting somehow.”in broadcasting somehow.”

7676


LucentLucentNortelNortelCiscoCiscoTellabsTellabs

AlcatelAlcatelAT&TAT&TSprintSprintWorldcomWorldcom

Leading Communications Leading Communications Vendors/CarriersVendors/Carriers

Leading CommunicationsLeading CommunicationsVendors/Carriers (Next Generation)Vendors/Carriers (Next Generation)

??






7777


NetAppNetAppEMCEMCIBMIBMHitachiHitachiBrocadeBrocadeCompaqCompaq

ZambeelZambeelBlueArcBlueArcSaneraSaneraNishanNishanTrueSANTrueSAN

Leading Storage Leading Storage New class ofNew class ofVendorsVendors VendorsVendors






Some Won’t Some Won’t Make itMake it

ER?BrokeAid?BlowUp?

Others Will Emerge Others Will Emerge as leadersas leaders

?

7878

Old vs. New in theOld vs. New in theMainframe to PC TransitionMainframe to PC Transition

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

Microsoft

IntelCompaq

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

01/3

0/87

07/3

1/87

01/2

9/88

07/2

9/88

01/3

1/89

07/3

1/89

01/3

1/90

07/3

1/90

01/3

1/91

07/3

1/91

01/3

1/92

07/3

1/92

01/2

9/93

07/3

0/93

Digital Equipment Corp.

NCRAmdahl

Data GeneralCapital fled legacy systems

New winners emerged

Source: Thomas Weisel Partners

This time the stakes

are larger!

7979

The Investors DilemmaThe Investors Dilemma

• Expect more losers than winnersExpect more losers than winners

• Many will lose everythingMany will lose everything

• More will be won than lostMore will be won than lost

• The value of winners will exceed the The value of winners will exceed the

cumulative market cap of the entire cumulative market cap of the entire

sector todaysector today

• M&A Game is hard to predictM&A Game is hard to predict

8080

Fundamentals of DemandFundamentals of Demand

• The EnterpriseThe Enterprise– Carrier services purchase decisions based on Carrier services purchase decisions based on

payback periodpayback period– Paybacks< 1 yr => Continued demandPaybacks< 1 yr => Continued demand– 1% increase in IT = 1.5-2% decrease in G&A1% increase in IT = 1.5-2% decrease in G&A– IT spending increased from 0.5% to 3.5% of sales IT spending increased from 0.5% to 3.5% of sales

in 90’sin 90’s

• The Service ProviderThe Service Provider– Enterprise demand => need for equipment Enterprise demand => need for equipment

purchasespurchases– Consumption of bandwidth + shifting computing Consumption of bandwidth + shifting computing

paradigm makes for new revenue opportunitiesparadigm makes for new revenue opportunities– Systems for old revenue don’t support new Systems for old revenue don’t support new

revenue & cost points =>new equipment purchasesrevenue & cost points =>new equipment purchases

8181

Comments?Comments?

[email protected]@kpcb.com

8383

““This wireless music box has This wireless music box has no imaginable commercial no imaginable commercial value. Who would pay for a value. Who would pay for a message sent to nobody in message sent to nobody in particular?”particular?”

-David Sarnoff’s associates in response -David Sarnoff’s associates in response to his urgings for investment in Radio in to his urgings for investment in Radio in the 1920’sthe 1920’s

Past AssumptionsPast Assumptions

8484

““Who the hell wants to hear Who the hell wants to hear actors talk?”actors talk?”

-Harry M. Warner, Warner Bros, 1927-Harry M. Warner, Warner Bros, 1927


8585

““The ‘telephone’ has too The ‘telephone’ has too many shortcomings to be many shortcomings to be seriously considered a means seriously considered a means of communication.” of communication.”

-Western Union Internal Memo, 1876-Western Union Internal Memo, 1876


8686

““There is no reason for any There is no reason for any individuals to have a individuals to have a computer in their home.”computer in their home.”-Ken Olsen, President, Chairman and -Ken Olsen, President, Chairman and Founder of DEC, 1977Founder of DEC, 1977


2003: ahead to the past

Economy & Finance

year vc capital

total private capital

b source

board date

new fund

fund ipo

ipo companies

new equipment