Anthony Noguera, Senior Manager of Sales & Planning Systems at NVIDIA and Padman Ramankutty, CEO at Intrigo Systems, will give you a glimpse into NVIDIA's Supply Chain initiative, driven by a focus on mobile computing, visual computing, and super-computing. [Presented at the 2012 SAP Summit on Suppy Chain Management in Newtown Square, PA]
6. BEST VISUAL EFFECTS
Quadro GPUs were behind all
of the “Best Visual Effects”
Oscar nominees for the past
three years.
Image courtesy of GK Films
7. WORLD-CLASS
SUPERCOMPUTING
In October 2011, Oak Ridge
National Labs announced that it
will use 18,000 NVIDIA Tesla GPUs
to develop the world’s fastest
supercomputer, called “Titan.”
8. MARKET LEADERSHIP
Tegra has taken the market by storm.
Today, more than 30 tablets and 65
phone SKUs are powered by the
super chip.
9. NVIDIA FACTS
Founded in 1993
Jen-Hsun Huang is co-founder,
president and CEO
Listed with NASDAQ under the
symbol NVDA in 1999
Invented the GPU in 1999 and
has shipped more than 1 billion
to date
FY12: $4 billion in revenue
7,000 employees worldwide
2,300 patents worldwide
Ranked #10 “greenest” company
in America by Newsweek in 2011
Headquartered in Santa Clara,
Calif.
10. About Intrigo
We are a Premier Enterprise Consultancy, focused on orchestrating Customer
Value Networks at the confluence of Demand, Supply and Innovation.
Respond Supply Demand Sense
Create
OUR SERVICES Source: AMR
Advisory
Re-architect the planning processes
Usability / workflow re-model Innovate
Enterprise Architecture and Software components
Implementation
Complete turnkey implementation
Assemble to deliver
Support
Offshore project/engineering delivery
Planning Analytics
11. Intrigo Systems’ Clients
“The ability of Intrigo’s leadership team to execute in a
complex, multi-party environment directly contributed to the
success of our project."
-- John Hanna,
Director - iPlan Project, Clorox
“Intrigo's deep industry domain knowledge and their
expertise in SAP made for an effective implementation of
our Business Systems. Their commitment and focus to
making us successful make Intrigo a key partner for
Aptina.”
-- Joe Passarello, CFO, Aptina Imaging
12. State of Supply Chain Planning in
the Semiconductor Industry
• Discrete vs. attribute-based part numbers
• Multiple postponement points
• Optimization planning models
• Long lead time materials (wafers)
• Capacity-constrained, high value materials
• Test equipment, allocation to multiple subcons, cost
• Heuristic planning models
• Order-based planning (customer centric Models)
• Lead time assumptions are vague in fabless model
13. Business Problem
NVIDIA's business has grown more complex
Our ability to project, manage and optimize our wafer supply
has entered the critical path of our growth
Existing spreadsheet-driven processes cannot scale to the
growth of our business (market segments, number of SKUs,
product complexity)
Focus today is simply aligning supply to marketing demand
by product family
Tomorrow, we want to employ production plans optimized for
cost, capacity and flexibility
14. Desired Maturity of Supply Chain
Planning
CODE
program
Excel Complete Scenario Optimized Integrated
Planning Planning planning planning planning
Active links with
Sales, BU, & Finance
tools; i.e., mutual
plans re-align
regularly
CODE – COMPLETE, OPTIMIZED DATA ENGINE
16. Vendor Evaluations
We evaluated several vendors with the final two
being SAP/Intrigo vs i2 (JDA)
Most of our business sponsors preferred i2 over
SAP given i2’s perceived richness of
semiconductor-specific functionalities and more
experience in the space
But IT and Intrigo convinced the business about the
much better TCO of SAP and the feasibility of this
path
17. Planning & Execution Landscape
Planning
APO Demand
PDMT Wafer & BE Planning Build Optimization Steelwedge Planning
Collaborative Supply Planner
Inventory Levels are calculated by the system using service targets, demand history,
and supply chain models
All Vendors All Business
Available
Safety Stock can also be defined in days of coverage
Requested
O.R. tools
Committed
Unmet
Unmet %
APO To promise
Requested
Committed
Unmet
Excel
Unmet %
Requested
Committed
S
Unmet
Unmet %
Supply
Requested
Committed
Unmet
Unmet %
Statement A
Planning data management APO ATP P
B
W
Execution Order Management +
Customer Demand Promise B
Orders Dates O
SAP SD
Windchill
Demand HUB
Sourcing Demand
and
SAP supply
(SPNR)
Box color legend
C o l l a b o r a t i v e S u p p l y P l a n n e r
I n v e n t o r y L e v e l s a r e c a l c u l a t e d b y t h e s y s t e m u s i n g s e r v i c e t a r g e t s , d e m a n d h i s t o r y ,
a A l l dV e
n n sd o r s p p l y
u c h a i n m o d e l s A l l B u s i n e s s
R e q u e s t e d
C o m m i t t e d
SAP
U n m e t
New System
U n m e t %
R e q u e s t e d
C o m m i t t e d
U n m e t
U n m e t %
R e q u e s t e d
C o m m i t t e d
Serus SAP MM
U n m e t
U n m e t %
R e q u e s t e d
C o m m i t t e d
U n m e t
U n m e t %
(ABR)
Changes in the
BOM interface
system
to SAP
Production Inventory
No changes
Execution Management
Sub cons
18. Guiding Principles
• Flexible planning models
• Postponement
• Buffer stock
• Usability and maintainability of the planning master data
• Planning rules maintained in Business language
• Ease of maintenance
• Explainability of the planning results
• Scenario and bridge Analysis
19. Planning Solution Overview
Wafer Allocation by
Optimization
Wafer Family:
for Wafer
Wafer Planning by SAPConstraint for
APO Optimizer
Planning
Backend planning
Backend planning
PFG Supply:
to determine Assy for Data Management
Custom UI
Backend planning by SAPConstraint for SLT
APO Optimizer
starts, PFG starts
for Forecast Planning
SLT Planning to Projected
determine PFGSLT planning by
Backend Finished goods
starts for demand CTM
SAP for Order
signal Commitments
20. Reference Model – Material Flow
Fab Bump Sort Assy Test SLT
Unbumped Bumped Wafer Sorted Wafer Die FG
Back Assembly PFG/FG*
Wafer
Grind
Back Grinded Tape
Die and
Reel
Direct Sort Flow (Wafer
Buy)
Direct Sorted Wafer Buy
Assembly Flow Test Flow PFG to FG Flow
Supply Plans generated for:
PFG Postponement Flow PFG to FG Flow
Fab - Sort
Die inv to FG
SLT/T&R adhoc runs Direct Flow
Focus on data, priorities, rules and system generates supply plans
* Test out part numbers
for Direct Flow only.
Required for capacity
planning in SLT
21. Planning Scenarios Overview
POR Version
Wafer Plan
Wafer capacity-constrained version
Wafer capacity-unconstrained version
Backend Plan
Backend material and capacity-constrained version
Backend material-constrained, capacity-unconstrained version
Backend constrained, push production
SLT Plan
SLT material and capacity-constrained version
SLT material-constrained, capacity-unconstrained version
Scenario solves as per the requirement
22. Planning Systems Overview
Input Screens Planning data management tool
APO Planning Engine
Build Rules
Build Strategies
Time Phased Data
Data
Repository
APO Master Data transfer
ECC Master Data transfer
Publish to output
APO Planning Engine
Data
Warehouse
APO Master Data transfer Transaction Data Planning results
Core Interface
Standard SAP component to transfer Planning
Input Data (Master and Transaction Data) and Output Data (Planning results)
Execution instructions to subcon
ECC Confirmations from subcon
23. Implementation Overview
• End-to-end Blueprint Design completed as of 2/15
• Phased project approach but design was end-to-end
• Expected go-live in 2012 and 2013
• Q&A