ECECS 472572 Final Exam ProjectRemember to check the errata

ECE/CS 472/572 Final Exam Project
Remember to check the errata section (at the very bottom of the
page) for updates.
Your submission should be comprised of two items: a .pdf file
containing your written report and a .tar file containing a
directory structure with your C or C++ source code. Your grade
will be reduced if you do not follow the submission
instructions.
All written reports (for both 472 and 572 students) must be
composed in MS Word, LaTeX, or some other word processor
and submitted as a PDF file.
Please take the time to read this entire document. If you have
questions there is a high likelihood that another section of the
document provides answers.
Introduction
In this final project you will implement a cache simulator. Your
simulator will be configurable and will be able to handle caches
with varying capacities, block sizes, levels of associativity,
replacement policies, and write policies. The simulator will
operate on trace files that indicate memory access properties.
All input files to your simulator will follow a specific structure
so that you can parse the contents and use the information to set
the properties of your simulator.
After execution is finished, your simulator will generate an
output file containing information on the number of cache
misses, hits, and miss evictions (i.e. the number of block
replacements). In addition, the file will also record the total
number of (simulated) clock cycles used during the situation.
Lastly, the file will indicate how many read and write
operations were requested by the CPU.
It is important to note that your simulator is required to make
several significant assumptions for the sake of simplicity.
1. You do not have to simulate the actual data contents. We

simply pretend that we copied data from main memory and keep
track of the hypothetical time that would have elapsed.
2. Accessing a sub-portion of a cache block takes the exact
same time as it would require to access the entire block.
Imagine that you are working with a cache that uses a 32 byte
block size and has an access time of 15 clock cycles. Reading a
32 byte block from this cache will require 15 clock cycles.
However, the same amount of time is required to read 1 byte
from the cache.
3. In this project assume that main memory RAM is always
accessed in units of 8 bytes (i.e. 64 bits at a time).
When accessing main memory, it's expensive to access the first
unit. However, DDR memory typically includes buffering which
means that the RAM can provide access to the successive
memory (in 8 byte chunks) with minimal overhead. In this
project we assume an overhead of 1 additional clock cycle per
contiguous unit.
For example, suppose that it costs 255 clock cycles to access
the first unit from main memory. Based on our assumption, it
would only cost 257 clock cycles to access 24 bytes of memory.
4. Assume that all caches utilize a "fetch-on-write" scheme if a
miss occurs on a Store operation. This means that you must
always fetch a block (i.e. load it) before you can store to that
location (if that block is not already in the cache).
Additional Resources
Sample trace files
Students are required to simulate the instructor-provided trace
files (although you are welcome to simulate your own files in
addition).
Trace files are available on Flip in the following directory:
/nfs/farm/classes/eecs/spring2021/cs472/public/tracefiles
You should test your code with all three tracefiles in that
directory (gcc, netpath, and openssl).
Starter Code
In order to help you focus on the implementation of the cache
simulator, starter code is provided (written in C++) to parse the

input files and handle some of the file I/O involved in this
assignment. You are not required to use the supplied code (it's
up to you). Note that you will need to adapt this code to work
with your particular design.
The starter code is available
here: https://classes.engr.oregonstate.edu/eecs/spring2021/cs472
/finalprojtemplatev5.zipLinks to an external site.
Basic-Mode Usage (472 & 572 students)
L1 Cache Simulator
All students are expected to implement the L1 cache simulator.
Students who are enrolled in 472 can ignore the sections that
are written in brown text. Graduate students will be simulating a
multiple-level cache (an L1 cache, an L2 cache, and even an L3
cache).
Input Information
Your cache simulator will accept two arguments on the
command line: the file path of a configuration file and the file
path of a trace file containing a sequence of memory operations.
The cache simulator will generate an output file containing the
simulation results. The output filename will have “.out”
appended to the input filename. Additional details are provided
below.
# example invocation of cache simulator
cache_sim ./resources/testconfig ./resources/simpletracefile
Output file written to ./resources/simpletracefile.out
The first command line argument will be the path to the
configuration file. This file contains information about the
cache design. The file will contain only numeric values, each of
which is on a separate line.
Example contents of a configurati on file:
1 <-- this line will always contain a "1" for 472 students
230 <-- number of cycles required to write or read a unit from
main memory
8 <-- number of sets in cache (will be a non-negative power of
2)
16 <-- block size in bytes (will be a non-negative power of 2)

3 <-- level of associativity (number of blocks per set)
1 <-- replacement policy (will be 0 for random replacement, 1
for LRU)
1 <-- write policy (will be 0 for write-through, 1 for write-back)
13 <-- number of cycles required to read or write a block from
the cache (consider this to be the access time per block)
Here is another example configuration file specifying a direct-
mapped cache with 64 entries, a 32 byte block size,
associativity level of 1 (direct-mapped), least recently used
(LRU) replacement policy, write-through operation, 26 cycles to
read or write data to the cache, and 1402 cycles to read or write
data to the main memory. CS/ECE472 projects can safely ignore
the first line.
1
1402
64
32
1
1
0
26
The second command line argument indicates the path to a trace
file. This trace file will follow the format used by Valgrind (a
memory debugging tool). The file consists of comments and
memory access information. Any line beginning with the ‘=’
character should be treated as a comment and ignored.
==This is a comment and can safely be ignored.
==An example snippet of a Valgrind trace file
I 04010173,3
I 04010176,6
S 04222cac,1
I 0401017c,7
L 04222caf,8
I 04010186,6
I 040101fd,7
L 1ffefffd78,8

M 04222ca8,4
I 04010204,4
Memory access entries will use the following format in the trace
file:
[space]operation address,size
· Lines beginning with an ‘I’ character represent an instruction
load. For this assignment, you can ignore instruction read
requests and assume that they are handled by a separate
instruction cache.
· Lines with a space followed by an ‘S’ indicate a data store
operation. This means that data needs to be written from the
CPU into the cache or main memory (possibly both) depending
on the write policy.
· Lines with a space followed by an ‘L’ indicate a data load
operation. Data is loaded from the cache into the CPU.
· Lines with a space followed by an ‘M’ indicate a data modify
operation (which implies a special case of a data load, followed
immediately by a data store).
The address is a 64 bit hexadecimal number representing the
address of the first byte that is being requested. Note that
leading 0's are not necessarily shown in the file. The size of the
memory operation is indicated in bytes (as a decimal number).
If you are curious about the trace file, you may generate your
own trace file by running Valgrind on arbitrary executable files:
valgrind --log-fd=1 --log-file=./tracefile.txt --tool=lackey --
trace-mem=yes name_of_executable_to_trace
Cache Simulator Output
Your simulator will write output to a text file. The output
filename will be derived from the trace filename with “.out”
appended to the original filename. E.g. if your program was
called using the invocation “cache_sim ./dm_config ./memtrace”
then the output file would be written to “./memtrace.out”
(S)tore, (L)oad, and (M)odify operations will each be printed to
the output file (in the exact order that they were read from the
Valgrind trace file). Lines beginning with “I” should not appear
in the output since they do not affect the operation of your

simulator.
Each line will have a copy of the original trace file instruction.
There will then be a space, followed by the number of cycles
used to complete the operation. Lastly, each line will have one
or more statements indicating the impact on the cache. This
could be one or more of the following: miss, hit, or eviction.
Note that an eviction is what happens when a cache block needs
to be removed in order to make space in the cache for another
block. It is simply a way of indicating that a block was
replaced. In our simulation, an eviction means that the next
instruction cannot be executed until after the existing cache
block is written to main memory. An eviction is an expensive
cache operation.
It is possible that a single memory access has multiple impacts
on the cache. For example, if a particular cache index is already
full, a (M)odify operation might miss the cache, evict an
existing block, and then hit the cache when the result is written
to the cache.
The general format of each output line (for 472 students) is as
follows (and will contain one or more cache impacts):
operation address,size <number_of_cycles> L1
<cache_impact1> <cache_impact2> <...>
The final lines of the output file are special. They will indicate
the total number of hits, misses, and evictions. The last line will
indicate the total number of simulated cycles that were
necessary to simulate the trace file, as well as the total number
of read and write operations that were directly requested by the
CPU.
These lines should exactly match the following format (with
values given in decimal):
L1 Cache: Hits:<hits> Misses:<misses> Evictions:<evictions>
Cycles:<number of total simulated cycles> Reads:<# of CPU
read requests> Writes:<# of CPU write requests>
In order to illustrate the output file format let’s look at an
example. Suppose we are simulating a direct-mapped cache
operating in write-through mode. Note that the replacement

policy does not have any effect on the operation of a direct-
mapped cache. Assume that the configuration file told us that it
takes 13 cycles to access the cache and 230 cycles to access
main memory. Keep in mind that a hit during a load operation
only accesses the cache while a miss must access both the cache
and the main memory. For this scenario, assume that memory
access is aligned to a single block and does not straddle
multiple cache blocks.
In this example the cache is operating in write-through mode so
a standalone (S)tore operation takes 243 cycles, even if it is a
hit, because we always write the block into both the cache and
into main memory. If this particular cache was operating in
write-back mode, a (S)tore operation would take only 13 cycles
if it was a hit (since the block would not be written into main
memory until it was evicted).
The exact details of whether an access is a hit or a miss is
entirely dependent on the specific cache design (block size,
level of associativity, number of sets, etc). Your program will
implement the code to see if each access is a hit, miss, eviction,
or some combination.
Since the (M)odify operation involves a Load operation
(immediately followed by a Store operation), it is recorded
twice in the output file. The first instance represents the load
operation and the next line will represent the store operation.
See the example below...
==For this example we assume that addresses 04222cac,
04222caf, and 04222ca8 are all in the same block at index 2
==Assume that addresses 047ef249 and 047ef24d share a block
that also falls at index 2.
==Since the cache is direct-mapped, only one of those blocks
can be in the cache at a time.
==Fortunately, address 1ffefffd78 happens to fall in a different
block index (in our hypothetical example).
==Side note: For this example a store takes 243 cycles (even if
it was a hit) because of the write-through behavior.
==The output file for our hypothetical example:

S 04222cac,1 486 L1 miss <-- (243 cycles to fetch the block and
write it to L1) + (243 cycles to update the cache & main
memory)
L 04222caf,8 13 L1 hit
M 1ffefffd78,8 243 L1 miss <-- notice that this (M)odify has a
miss for the load and a hit for the store
M 1ffefffd78,8 243 L1 hit <-- this line represents the Store of
the modify operation
M 04222ca8,4 13 L1 hit <-- notice that this (M)odify has two
hits (one for the load, one for the store)
M 04222ca8,4 243 L1 hit <-- this line represents the Store of
the modify operation
S 047ef249,4 486 L1 miss eviction <-- 486 cycles for miss, no
eviction penalty for write-through cache
L 04222caf,8 243 L1 miss eviction
M 047ef24d,2 243 L1 miss eviction <-- notice that this (M)odify
initially misses, evicts the block, and then hits
M 047ef24d,2 243 L1 hit <-- this line represents the Store of the
modify operation
L 1ffefffd78,8 13 L1 hit
M 047ef249,4 13 L1 hit
M 047ef249,4 243 L1 hit
L1 Cache: Hits:8 Misses:5 Evictions:3
Cycles:2725 Reads:7 Writes:6 <-- total sum of simulated cycles
(from above), as well as the number of reads and writes that
were requested by the CPU
NOTE: The example above is assuming that the cache has a
block size of at least 8 bytes. Simulating a cache with a smaller
blocksize would affect the timing and could also lead to
multiple evictions in a single cache access. For example, if the
blocksize was only 4 bytes it's possible that an 8 byte l oad
might evict 3 different blocks. This happens because the data
might straddle two or more blocks (depending on the starting
memory address).
Sample Testing Information
Some students have asked for additional test files with "known"

results that they can compare against. I've created my own
implementation of the cache simulator and provided students
with the following files (and results).
Note: These files are not an exhaustive representation of the
testing that your cache will undergo. It is your job to
independently test your code and verify proper behavior.
Examples that utilize an L1 cache:
· Sample 1
· sample1_config download
· sample1_trace download
· sample1_trace.out download
· Sample 2
· Sample 3
(572) Examples that utilize at least 2 caches:
· Sample 1
· multi_sample1_config download
· multi_sample1_trace download
· multi_sample1_trace.out download
· Sample 2
· multi_sample2_config download
· multi_sample2_trace download
· multi_sample2_trace.out download
Facts and Questions (FAQ):
· Your "random" cache replacement algorithm needs to be
properly seeded so that multiple runs of the same tracefile will
generate different results.
· I will never test your simulator using a block size that is
smaller than 8 bytes.

· During testing, the cache will not contain more than 512
indexes.
· For our purposes the level of associativity could be as small as
N=1 (direct mapped) or as large as N=64.
· The last line of your output will indicate the total number of
simulated cycles that were necessary to simulate the trace file,
as well as the total number of read and write operations that
were directly requested by the CPU. In other words, this is
asking how many loads and stores the CPU directly requested
(remember that a Modify operation counts as both a Load and a
Store).
· 572 students: For our purposes an L2 cache will always have a
block size that is greater than or equal to the L1 block size. The
L3 block size will be greater than or equal to the L2 block size.
Implementation Details
You may use either the C or the C++ programming language.
Graduate students will have an additional component to this
project.
In our simplified simulator, increasing the level of associativity
has no impact on the cache access time. Furthermore, you may
assume that it does not take any additional clock cycles to
access non-data bits such as Valid bits, Tags, Dirty Bits, LRU
counters, etc.
Your code must support the LRU replacement scheme and the
random replacement scheme. For the LRU behavior, a block is
considered to be the Least Recently Used if every other block in
the cache has been read or written after the block in question. In
other words, your simulator must implement a true LRU
scheme, not an approximation.
You must implement the write-through cache mode. You will
receive extra credit if your code correctly supports the write -
back cache mode (specified in the configuration file).
Acceptable Compiler Versions
The flip server provides GCC 4.8.5 for compiling your work.
Unfortunately, this version is from 2015 and may not support

newer C and C++ features. If you call the program using “gcc”
(or “g++”) this is the version you will be using by default.
If you wish to use a newer compiler version, I have compiled a
copy of GCC 10.3 (released April 8, 2021). You may write your
code using this compiler and you’re allowed to use any of the
compiler features that are available. The compiler binaries are
available in the path:
/nfs/farm/classes/eecs/spring2021/cs472/public/gcc/bin
For example, in order to compile a C++ program with GCC
10.3, you could use the following command (on a single
terminal line):
/nfs/farm/classes/eecs/spring2021/cs472/public/gcc/bin/g++ -
ocache_sim -Wl,-
rpath,/nfs/farm/classes/eecs/spring2021/cs472/public/gcc/lib64
my_source_code.cpp
If you use the Makefile that is provided in the starter code, it is
already configured to use GCC 10.3.
L2/L3 Cache Implementation (required for CS/ECE 572
students)
Implement your cache simulator so that it can support up to 3
layers of cache. You can imagine that these caches are
connected in a sequence. The CPU will first request information
from the L1 cache. If the data is not available, the request will
be forwarded to the L2 cache. If the L2 cache cannot fulfill the
request, it will be passed to the L3 cache. If the L3 cache cannot
fulfill the request, it will be fulfilled by main memory.
It is important that the properties of each cache are read from
the provided configuration file. As an example, it is possible to
have a direct-mapped L1 cache that operates in cohort with an
associative L2 cache. All of these details will be read from the
configuration file. As with any programming project, you
should be sure to test your code across a wide variety of
scenarios to minimize the probability of an undiscovered bug.
Cache Operation
When multiple layers of cache are implemented, the L1 cache
will no longer directly access main memory. Instead, the L1

cache will interact with the L2 cache. During the design
process, you need to consider the various interactions that can
occur. For example, if you are working with three write-through
caches, than a single write request from the CPU will update the
contents of L1, L2, L3, and main memory!
++++++++++++ ++++++++++++ ++++++++++++
++++++++++++ +++++++++++++++
| | | | | | | | | |
| CPU | <----> | L1 Cache | <----> | L2 Cache | <----> | L3
Cache | <----> | Main Memory |
| | | | | | | | | |
++++++++++++ ++++++++++++ ++++++++++++
++++++++++++ +++++++++++++++
Note that your program should still handle a configuration file
that specifies an L1 cache (without any L2 or L3 present). In
other words, you can think of your project as a more advanced
version of the 472 implementation.
572 Extra Credit
By default, your code is only expected to function with write-
through caches. If you want to earn extra credit, also implement
support for write-back caches.
In this situation, you will need to track dirty cache blocks and
properly handle the consequences of evictions. You will earn
extra credit if your write-back design works with simple L1
implementations. You will receive additional extra credit if
your code correctly handles multiple layers of write-back
caches (e.g. the L1 and L2 caches are write-back, but L3 is
write-through) .
Simulator Operation
Your cache simulator will use a similar implementation as the
single-level version but will parse the configuration file to
determine if multiple caches are present.
Input Information
The input configuration file is as shown below. Note that it is
backwards compatible with the 472 format.

The exact length of the input configuration file will depend on
the number of caches that are specified.
3 <-- this line indicates the number of caches in the simulation
(this can be set to a maximum of 3)
230 <-- number of cycles required to write or read a block from
main memory
8 <-- number of sets in L1 cache (will be a non-negative power
of 2)
16 <-- L1 block size in bytes (will be a non-negative power of
2)
4 <-- L1 level of associativity (number of blocks per set)
1 <-- L1 replacement policy (will be 0 for random replacement,
1 for LRU)
1 <-- L1 write policy (will be 0 for write-through, 1 for write-
back)
the L1 cache (consider this to be the access time)
of 2)
2)
1 for LRU)
back)
of 2)
2)
1 for LRU)

back)
Cache Simulator Output
The output file will contain nearly the same information as in
the single-level version (see the general description provided in
the black text). However, the format is expanded to contain
information about each level of the cache.
The general format of each output line is as follows (and can
list up to 2 cache impacts for each level of the cache):
operation address,size <total number_of_cycles> L1
<cache_impact1> <cache_impact2> <...> L2 <cache_impact1>
<cache_impact2> <...> L3 <cache_impact1> <cache_impact2>
<...>
The exact length of each line will vary, depending how many
caches are in the simulation (as well as their interaction). For
example, imagine a system that utilizes an L1 and L2 cache.
If the L1 cache misses and the L2 cache hits, we might see
something such as the following:
L 04222caf,8 53 L1 miss L2 hit
In this scenario, if the L1 cache hits, then the L2 cache will not
be accessed and does not appear in the output.
L 04222caf,8 13 L1 hit
Suppose L1, L2, and L3 all miss (implying that we had to access
main memory):
L 04222caf,8 393 L1 miss L2 miss L3 miss
(M)odify operations are the most complex since they involve
two sub-operations... a (L)oad immediately followed by a
(S)tore.
M 1ffefffd78,8 163 L1 miss eviction L2 miss L3 hit <-- notice
that the Load portion of this (M)odify operation caused an L1
miss, L2 miss, and L3 hit
M 1ffefffd78,8 13 L1 hit <-- this line belongs to the store

portion of the (M)odify operation
The final lines of the output file are special. They will indicate
the total number of hits, misses, and evictions for each specific
cache. The very last line will indicate the total number of
simulated cycles that were necessary to simulate the trace file,
as well as the total number of read and write operations that
were directly requested by the CPU.
These lines should exactly match the following format (with
values given in decimal):
Cycles:<number of total simulated cycles> Reads:<# of CPU
read requests> Writes:<# of CPU write requests>
Project Write-Up
Note: Any chart or graphs in your written report must have
labels for both the vertical and horizontal axis.
Undergraduates (CS/ECE 472)
Part 1: Summarize your work in a well-written report. The
report should be formatted in a professional format. Use images,
charts, diagrams or other visual techniques to help convey your
information to the reader.
Explain how you implemented your cache simulator. You should
provide enough information that a knowledgeable programmer
would be able to draw a reasonably accurate block diagram of
your program.
· What data structures did you use to implement your design?
· What were the primary challenges that you encountered while
working on the project?
· Is there anything you would implement differently if you were
to re-implement this project?
· How do you track the number of clock cycles needed to
execute memory access instructions?
Part 2: There is a general rule of thumb that a direct-mapped
cache of size N has about the same miss rate as a 2-way set
associative cache of size N/2.

Your task is to use your cache simulator to conclude whether
this rule of thumb is actually worth using. You may test your
simulator using instructor-provided trace files (see the sample
trace files section) or you may generate your own trace files
from Linux executables (“wget oregonstate.edu”, “ls”, “hostid”,
“cat /etc/motd”, etc). Simulate at least three trace files and
compare the miss rates for a direct-mapped cache versus a 2-
way set associative cache of size N/2. For these cache
simulations, choose a block size and number of indices so that
the direct-mapped cache contains 64KiB of data. The 2-way set
associative cache (for comparison) should then contain 32KiB
of data. You are welcome to experiment with different block
sizes/number of indices to see how your simulation results are
affected. You could also simulate additional cache sizes to
provide more comparison data. After you have obtained
sufficient data to support your position, put your simulation
results into a graphical plot and explain whether you agree with
the aforementioned rule of thumb. Include this information in
your written report.
Part 3: If you chose to implement any extra credit tasks, be sure
to include a thorough description of this work in the report.
Graduate Students (CS/ECE 572)
Part 1: Summarize your work in a well-written report. The
report should be formatted in a professional format. Use images,
charts, diagrams or other visual techniques to help convey your
information to the reader.
Explain how you implemented your cache simulator. You should
provide enough information that a knowledgeable programmer
would be able to draw a reasonably accurate block diagram of
your program.
· What data structures did you use to implement your multi-
level cache simulator?
· What were the primary challenges that you encountered while
working on the project?
· Is there anything you would implement differently if you were
to re-implement this project?

· How do you track the number of clock cycles needed to
execute memory access instructions?
Part 2: Using trace files provided by the instructor (see the
sample trace files section), how does the miss rate and average
memory access time (in cycles) vary when you simulate a
machine with various levels of cache? Note that you can
compute the average memory access time by considering the
total number of read and write operations (requested by the
CPU), along with the total number of simulated cycles that it
took to fulfill the requests.
Research a real-life CPU (it must contain at least an L2 cache)
and simulate the performance with L1, L2, (and L3 caches if
present). You can choose the specific model of CPU (be sure to
describe your selection in your project documentation). This
could be an Intel CPU, an AMD processor, or some other
modern product. What is the difference in performance when
you remove all caches except the L1 cache? Be sure to run this
comparison with each of the three instructor-provided trace
files. Provide written analysis to explain any differences in
performance. Also be sure to provide graphs or charts to
visually compare the difference in performance.
Part 3: If you chose to implement any extra credit tasks, be sure
to include a thorough description of this work in the report.
Submission Guidelines
You will submit both your source code and a PDF file
containing the typed report.
Any chart or graphs in your written report must have labels for
both the vertical and horizontal axis!
For the source code, you must organize your source code/header
files into a logical folder structure and create a tar file that
contains the directory structure. Your code must be able to
compile on flip.engr.oregonstate.edu. If your code does not
compile on the engineering servers you should expect to receive
a 0 grade for all implementation portions of the grade.
Your submission must include a Makefile that can be used to

compile your project from source code. It is acceptable to adapt
the example Makfile from the starter code. If you need a
refresher, please see this helpful page (Links to an external
site.). If the Makefile is written correctly, the grader should be
able to download your TAR file, extract it, and run the “make”
command to compile your program. The resulting executable
file should be named: “cache_sim”.
Grading and Evaluation
CS/CE 472 students can complete the 572 project if they prefer
(and must complete the 572 write-up, rather than the
undergraduate version). Extra credit will be awarded to 472
students who choose to complete this task.
Your source code and the final project report will both be
graded. Your code will be tested for proper functionality. All
aspects of the code (cleanliness, correctness) and report (quality
of writing, clarity, supporting evidence) will be considered in
the grade. In short, you should be submitting professional
quality work.
You will lose points if your code causes a segmentation fault or
terminates unexpectedly.
The project is worth 200 points (100 points for the written
report and 100 points for the the C/C++ implementation).
Extra Credit Explanation
The extra credit is as follows. Note that in order to earn full
extra credit, the work must be well documented in your written
report.
ECE/CS 472 Extra Credit Opportunities
10 points - Implement and document write-back cache support.
30 points - Implement and document the 572 project instead of
the 472 project. All 572 expectations must be met.
ECE/CS 572 Extra Credit Opportunities
10 points - Implement and document write-back cache support
for a system that contains only an L1 cache.
10 points (additional) - Extend your implementation so that it
works with multiple layers of write-back caches. E.g. if a dirty
L1 block is evicted, it should be written to the L2 cache and the

corresponding L2 block should be marked as dirty. Assuming
that the L2 cache has sufficient space, the main memory would
not be updated (yet).
Errata
This section of the assignment will be updated as changes and
clarifications are made. Each entry here should have a date and
brief description of the change so that you can look over the
errata and easily see if any updates have been made since your
last review.
May 13th - Released the project guidelines.
May 15th - Added note about fetch-on-write behavior for all
caches. Added link to the starter code (written in C++).
May 16th - Refined explanation to clarify that "fetch-on-write"
only comes into play when the CPU tries to store a block that is
not currently contained in the cache.
May 31st - Added some additional configuration files that
students can use while they are verifying their cache behavior.
Updated information regarding GCC 10.3 (dropped GCC 11.1
because Valgrind wasn't entirely compatible). Added a "FAQ"
section to clarify other details based on student questions.
June 1st - The last line of the output file reflects the number of
Load and Store operations (with a Modify counting as one of
each) that the CPU directly requested. For clarification, even if
a Load operation affects multiple cache blocks, it still counts as
a single CPU read request. A similar reasoning applies to the
CPU write counter.
June 3rd/4th - Add some additional test configurations that
students can use to check their simulator's functionality.
L 04 286 L1 miss L2 missL 808 286 L1 miss L2 missL 104 53
L1 miss L2 hitL 1026 286 L1 miss L2 missL 26 13 L1 hitS 1805
572 L1 miss eviction L2 miss hitL 1002 13 L1 hitL 420 26 L1
hitL 1002 13 L1 hitL 808 53 L1 miss eviction L2 hitM 1806 53
L1 miss eviction L2 hitM 1806 286 L1 hit L2 hitL 1008 13 L1
hitS 1005 286 L1 hit L2 hitS 1808 286 L1 hit L2 hitL 808 13 L1
hitM 08 53 L1 miss eviction L2 hitM 08 286 L1 hit L2 hitM 04

13 L1 hitM 04 286 L1 hit L2 hitL 2804 286 L1 miss eviction L2
missM 808 13 L1 hitM 808 286 L1 hit L2 hitL1 Cache: Hits:15
Misses:9 Evictions:5L2 Cache: Hits:11 Misses:5
Evictions:0Cycles:3761 Reads:16 Writes:7
finalprojtemplate/.gitignore
# Ignore the build directory
build
# Ignore any executables
bin/*
# Ignore Mac specific files
.DS_Store
finalprojtemplate/Makefile
# Script adapted from https://hiltmon.com/blog/2013/07/03/a-
simple-c-plus-plus-project-structure/
CC :=
/usr/local/classes/eecs/spring2021/cs472/public/gcc/bin/g++
STRIPUTIL = strip
SRCDIR = src
BUILDDIR = build
TARGET = bin/cache_sim
# Handle debug case

DEBUG ?= 1
ifeq ($(DEBUG), 1)
CFLAGS += -g -Wall
else
CFLAGS += -DNDEBUG -O3
endif
SRCEXT = cpp
SOURCES = $(shell find $(SRCDIR) -type f -name
*.$(SRCEXT))
OBJECTS = $(patsubst
$(SRCDIR)/%,$(BUILDDIR)/%,$(SOURCES:.$(SRCEXT)=.o))
LDFLAGS += -Wl,-
rpath,/usr/local/classes/eecs/spring2021/cs472/public/gcc/lib64
LIB += -pthread
INC += -I $(SRCDIR)
$(TARGET): $(OBJECTS)
@echo "Linking..."
@echo " $(CC) $^ -o $(TARGET) $(LIB) $(LDFLAGS)";
$(CC) $^ -o $(TARGET) $(LIB) $(LDFLAGS)
$(BUILDDIR)/%.o: $(SRCDIR)/%.$(SRCEXT)
@mkdir -p $(BUILDDIR)
@echo " $(CC) $(CFLAGS) $(INC) -c -o [email protected]
$<"; $(CC) $(CFLAGS) $(INC) -c -o [email protected] $<
clean:
@echo "Cleaning...";
@echo " $(RM) -r $(BUILDDIR) $(TARGET)"; $(RM) -r
$(BUILDDIR) $(TARGET)
.PHONY: clean
finalprojtemplate/resources/simpletracefile==This is a simple

tracefile for testing purposes L 08 L 808 L 1006 L 08 S 1805 L
1002 L 808 M 1806 L 1008 S 1005 S 1808 L 808 M 08 M 04 L
2804 M 808
finalprojtemplate/resources/testconfig
1
230
8
16
3
1
0
13
finalprojtemplate/src/CacheController.cppfinalprojtemplate/src/
CacheController.cpp/*
Cache Simulator (Starter Code) by Justin Goins
Oregon State University
Spring Term 2021
*/
#include"CacheController.h"
#include<iostream>
#include<fstream>
#include<regex>
#include<cmath>
usingnamespace std;
CacheController::CacheController(CacheInfo ci, string tracefile
){
// store the configuration info
this->ci = ci;
this->inputFile = tracefile;
this->outputFile =this->inputFile +".out";

// compute the other cache parameters
this->ci.numByteOffsetBits = log2(ci.blockSize);
this->ci.numSetIndexBits = log2(ci.numberSets);
// initialize the counters
this->globalCycles =0;
this->globalHits =0;
this->globalMisses =0;
this->globalEvictions =0;
// create your cache structure
// ...
// manual test code to see if the cache is behaving properly
// will need to be changed slightly to match the function prototy
pe
/*
cacheAccess(false, 0);
*/
}
/*
Starts reading the tracefile and processing memory operations
.
*/
voidCacheController::runTracefile(){
cout <<"Input tracefile: "<< inputFile << endl;
cout <<"Output file name: "<< outputFile << endl;
// process each input line
string line;

// define regular expressions that are used to locate commands
regex commentPattern("==.*");
regex instructionPattern("I .*");
regex loadPattern(" (L )(.*)(,)([[:digit:]]+)$");
regex storePattern(" (S )(.*)(,)([[:digit:]]+)$");
regex modifyPattern(" (M )(.*)(,)([[:digit:]]+)$");
// open the output file
ofstream outfile(outputFile);
// open the output file
ifstream infile(inputFile);
// parse each line of the file and look for commands
while(getline(infile, line)){
// these strings will be used in the file output
string opString, activityString;
smatch match;// will eventually hold the hexadecimal addr
ess string
unsignedlongint address;
// create a struct to track cache responses
CacheResponse response;
// ignore comments
if(std::regex_match(line, commentPattern)|| std::regex_match(li
ne, instructionPattern)){
// skip over comments and CPU instructions
continue;
}elseif(std::regex_match(line, match, loadPattern)){
cout <<"Found a load op!"<< endl;
istringstream hexStream(match.str(2));
hexStream >> std::hex >> address;
outfile << match.str(1)<< match.str(2)<< match.str(3)<
< match.str(4);
cacheAccess(&response,false, address, stoi(match.str(4)
));
logEntry(outfile,&response);

}elseif(std::regex_match(line, match, storePattern)){
cout <<"Found a store op!"<< endl;
< match.str(4);
cacheAccess(&response,true, address, stoi(match.str(4))
);
}elseif(std::regex_match(line, match, modifyPattern)){
cout <<"Found a modify op!"<< endl;
// first process the read operation
< match.str(4);
cacheAccess(&response,false, address, stoi(match.str(4)
));
outfile << endl;
// now process the write operation
< match.str(4);
cacheAccess(&response,true, address, stoi(match.str(4))
);
}else{
throw runtime_error("Encountered unknown line format in trace
file.");
}
outfile << endl;
}
// add the final cache statistics
outfile <<"Hits: "<< globalHits <<" Misses: "<< globalMisse
s <<" Evictions: "<< globalEvictions << endl;

outfile <<"Cycles: "<< globalCycles << endl;
infile.close();
outfile.close();
}
/*
Report the results of a memory access operation.
*/
voidCacheController::logEntry(ofstream& outfile,CacheRespons
e* response){
outfile <<" "<< response->cycles;
if(response->hits >0)
outfile <<" hit";
if(response->misses >0)
outfile <<" miss";
if(response->evictions >0)
outfile <<" eviction";
}
/*
Calculate the block index and tag for a specified address.
*/
CacheController::AddressInfoCacheController::getAddressInfo(
unsignedlongint address){
AddressInfo ai;
// this code should be changed to assign the proper index and ta
g
return ai;
}
/*
This function allows us to read or write to the cache.
The read or write is indicated by isWrite.
address is the initial memory address
numByte is the number of bytes involved in the access

*/
voidCacheController::cacheAccess(CacheResponse* response,b
ool isWrite,unsignedlongint address,int numBytes){
// determine the index and tag
AddressInfo ai = getAddressInfo(address);
cout <<"tSet index: "<< ai.setIndex <<", tag: "<< ai.tag << e
ndl;
// your code should also calculate the proper number of cycles t
hat were used for the operation
response->cycles =0;
// your code needs to update the global counters that track the n
umber of hits, misses, and evictions
if(response->hits >0)
cout <<"Operation at address "<< std::hex << address <<"
caused "<< response->hits <<" hit(s)."<< std::dec << endl;
if(response->misses >0)
cout <<"Operation at address "<< std::hex << address <<"
caused "<< response-
>misses <<" miss(es)."<< std::dec << endl;
cout <<"-----------------------------------------"<< endl;
return;
}
finalprojtemplate/src/CacheController.h
/*

Spring Term 2021
*/
#ifndef _CACHECONTROLLER_H_
#define _CACHECONTROLLER_H_
#include "CacheStuff.h"
#include <string>
#include <fstream>
class CacheController {
private:
struct AddressInfo {
unsigned long int tag;
unsigned int setIndex;
};
unsigned int globalCycles;
unsigned int globalHits;

unsigned int globalMisses;
unsigned int globalEvictions;
std::string inputFile, outputFile;
CacheInfo ci;
// function to allow read or write access to the cache
void cacheAccess(CacheResponse*, bool, unsigned
long int, int);
// function that can compute the index and tag
matching a specific address
AddressInfo getAddressInfo(unsigned long int);
// function to add entry into output file
void logEntry(std::ofstream&, CacheResponse*);
public:
CacheController(CacheInfo, std::string);
void runTracefile();

};
#endif //CACHECONTROLLER
finalprojtemplate/src/CacheSimulator.cppfinalprojtemplate/src/
CacheSimulator.cpp/*
Spring Term 2021
*/
#include"CacheSimulator.h"
#include"CacheStuff.h"
#include"CacheController.h"
#include<iostream>
#include<fstream>
#include<thread>
usingnamespace std;
/*
This function creates the cache and starts the simulator.
Accepts core ID number, configuration info, and the name of
the tracefile to read.
*/
void initializeCache(int id,CacheInfo config, string tracefile){
CacheController singlecore =CacheController(config, tracefile);
singlecore.runTracefile();
}

/*
This function accepts a configuration file and a trace file on t
he command line.
The code then initializes a cache simulator and reads the requ
ested trace file(s).
*/
int main(int argc,char* argv[]){
CacheInfo config;
if(argc <3){
cerr <<"You need two command line arguments. You shoul
d provide a configuration file and a trace file."<< endl;
return1;
}
// determine how many cache levels the system is using
unsignedint numCacheLevels;
// read the configuration file
cout <<"Reading config file: "<< argv[1]<< endl;
ifstream infile(argv[1]);
unsignedint tmp;
infile >> numCacheLevels;
infile >> config.memoryAccessCycles;
infile >> config.numberSets;
infile >> config.blockSize;
infile >> config.associativity;
infile >> tmp;
config.rp =static_cast<ReplacementPolicy>(tmp);
infile >> tmp;
config.wp =static_cast<WritePolicy>(tmp);
infile >> config.cacheAccessCycles;
infile.close();
// Examples of how you can access the configuration file inform
ation
cout <<"System has "<< numCacheLevels <<" cache(s)."<< e

ndl;
cout << config.numberSets <<" sets with "<< config.blockSiz
e <<" bytes in each block. N = "<< config.associativity << endl;
if(config.rp ==ReplacementPolicy::Random)
cout <<"Using random replacement protocol"<< endl;
else
cout <<"Using LRU protocol"<< endl;
if(config.wp ==WritePolicy::WriteThrough)
cout <<"Using write-through policy"<< endl;
else
cout <<"Using write-back policy"<< endl;
// start the cache operation...
string tracefile(argv[2]);
initializeCache(0, config, tracefile);
return0;
}
finalprojtemplate/src/CacheSimulator.h
/*
Spring Term 2021
*/
#ifndef _CACHESIMULATOR_H_

#define _CACHESIMULATOR_H_
#endif //CACHESIMULATOR
finalprojtemplate/src/CacheStuff.h
/*
Spring Term 2021
*/
#ifndef _CACHESTUFF_H_
#define _CACHESTUFF_H_
enum class ReplacementPolicy {
Random,
LRU
};

enum class WritePolicy {
WriteThrough,
WriteBack
};
// structure to hold information about a particular cache
struct CacheInfo {
unsigned int numByteOffsetBits;
unsigned int numSetIndexBits;
unsigned int numberSets; // how many sets are in the cache
unsigned int blockSize; // size of each block in bytes
unsigned int associativity; // the level of associativity (N)
ReplacementPolicy rp;
WritePolicy wp;
unsigned int cacheAccessCycles;
unsigned int memoryAccessCycles;
};

// this structure can filled with information about each memory
operation
struct CacheResponse {
int hits; // how many caches did this memory operation
hit?
int misses; // how many caches did this memory operation
miss?
int evictions; // did this memory operation involve one or
more evictions?
int dirtyEvictions; // were any evicted blocks marked as
dirty? (relevant for write-back cache)
unsigned int cycles; // how many clock cycles did this
operation take?
};
#endif //CACHESTUFF
L 04 246 L1 missL 808 246 L1 missL 104 13 L1 hitL 1026 246
L1 missL 26 13 L1 hitS 1805 492 L1 missL 1002 13 L1 hitL
420 13 L1 hitL 1002 13 L1 hitL 808 13 L1 hitM 1806 13 L1
hitM 1806 246 L1 hitL 1008 13 L1 hitS 1005 246 L1 hitS 1808
246 L1 hitL 808 13 L1 hitM 08 13 L1 hitM 08 246 L1 hitM 04
13 L1 hitM 04 246 L1 hitL 2804 246 L1 missM 808 13 L1 hitM
808 246 L1 hitL1 Cache: Hits:18 Misses:5

L 04 243 L1 missL 808 243 L1 missL 104 243 L1 missL 1026
243 L1 missL 26 13 L1 hitS 1805 486 L1 miss evictionL 1002
13 L1 hitL 420 269 L1 miss hitL 1002 13 L1 hitL 808 243 L1
miss evictionM 1806 243 L1 miss evictionM 1806 243 L1 hitL
1008 13 L1 hitS 1005 243 L1 hitS 1808 243 L1 hitL 808 13 L1
hitM 08 243 L1 miss evictionM 08 243 L1 hitM 04 13 L1 hitM
04 243 L1 hitL 2804 243 L1 miss evictionM 808 13 L1 hitM
808 243 L1 hitL1 Cache: Hits:15 Misses:10
L 0,4 246 L1 miss
L 1f,1 13 L1 hit
L 20,1 17 L1 miss
L1 Cache: Hits:1 Misses:2 Evictions:0
Cycles:276 Reads:3 Writes:0
F R A N K T. RO T H A E R M E L
MHE-FTR-067
1 2 6 0 2 6 1 2 8 X
R E V I S E D : M A R C H 7, 2 0 2 0
MH0067
Professor Frank T. Rothaermel prepared this case from public
sources. Research assistance by Laura Zhang is gratefully
acknowledged. This case is developed for
the purpose of class discussion. This case is not intended to be
used for any kind of endorsement, source of data, or depiction
of efcient or inefcient management.

All opinions expressed, and all errors and omissions, are
entirely the author’s. © Rothaermel, 2020.
Tesla, Inc.
When Henry Ford made cheap, reliable cars, people said, “Nah,
what’s wrong with a horse?”
That was a huge bet he made, and it worked.
— Elon Musk1
January 7, 2020. Shanghai, China. Just one year after Elon
Musk with a cadre of high-ranking Chinese officials
broke ground on a dirt field near Shanghai China to commence
the building of Gigafactory 3, Tesla’s CEO was now
dancing on stage to celebrate the first deliveries of locally made
Model 3s.
Tesla’s stock market valuation crossed the $150 billion
threshold.2 This made the electric vehicle startup more
valuable than GM, Ford, and Chrysler combined and the second
most valuable auto company globally, only behind
Toyota Motor Corp.—but ahead of the Volkswagen Group, the
world’s two largest car manufacturers. To put Tesla’s
stock market valuation in perspective, in 2019, GM and Ford
combined made more than 10 million vehicles while
Toyota and Volkswagen each made over 10 million. In
comparison, Tesla made less than 370,000 cars.
Just a few months earlier, in the summer of 2019, Tesla’s
market cap hit a low point of $32 billion. Back then,
as the company was trying to scale-up production of the Model
3 to meet demand, many had speculated that the
company would soon run out of cash because it kept missing
delivery deadlines and was mired in “production hell”
(as CEO Elon Musk put it).3

After the raucous delivery party in Tesla’s brand-new Shanghai
Gigafactory, Elon Musk sat back and relaxed
as the private jet took off. As the Gulfstream G700 gained
altitude, Elon was trying to catch up on his sleep, but a
couple of things kept him awake. In particular, the Tesla CEO
worried about how the company would continue to
scale-up production profitably, while also launching several
new models. Later in 2020, Tesla plans the delivery of
its Model Y, a smaller compact sport utility vehicle (SUV). And
in 2021, Musk promises the first deliveries of the
futuristic pickup truck—the Cybertruck. Although Elon Musk
was happy that Tesla beat expectations in 2019 by
delivering 367,500 vehicles (Exhibit 1), for 2020 he promises
delivery of over 500,000 vehicles.
Perhaps, even more important, Elon Musk worried about future
demand in the company’s three key markets:
the United States, China, and Europe. Even if Tesla succeeded
to ramp up production, will there be sufficient
For the exclusive use of X. Shang, 2020.
This document is authorized for use only by Xinyu Shang in
AD715 Fall 2020 taught by VLADIMIR ZLATEV, Boston
University from Aug 2020 to Feb 2021.
2
Tesla, Inc.
Copyright © 2020 McGraw-Hill Education. All rights reserved.
No reproduction, distribution, or posting online without the
prior written consent

of McGraw-Hill Education.
demand? How could the electric vehicle (EV) company grow
production by more than 35 percent while being
profitable? Although Tesla had some profitable quarters in the
recent past, on an annual basis, the company is yet
to make a profit (Exhibit 2). Federal tax credits in the United
States for Tesla vehicles have ended, while China is
also reducing tax incentives for electric vehicles. In the
meantime, his phone kept buzzing with reports of some
kind of new coronavirus causing flu-like symptons and
pneumonia, with adverse repercussions on Tesla’s supply
chain and its new Gigafactory in Shanghai.
Another issue that troubled Musk, and which leads him to tweet
often directly with disgruntled customers, is
the perception that although Tesla styles itself as a luxury car
brand, its delivery experience, and customer service
is not up to par. Given a large number of deliveries of Model 3s
in the United States during the second half of
2019, Tesla’s customer service bandwidth and capabilities had
yet to catch up. The EV-company began to develop
a reputation, at least in the United States, for launching
innovative and paradigm-defining vehicles. Yet at the same
time, many Tesla owners and observers considered its customer
service inferior.
As Musk grabbed a low-carb Monster energy drink from the
fridge in his airplane suite, he booted up his Lenovo
laptop, and began to organize his thoughts …
Elon Musk: Engineer Entrepreneur Extraordinaire
In 1989, at the age of 17, Elon Musk left his native South Africa
to avoid being conscripted into the army. Says
Musk, “I don’t have an issue with serving in the military per se

but serving in the South African army suppressing
black people just didn’t seem like a really good way to spend
time.”4 He went to Canada and subsequently enrolled
at Queen’s University in 1990. After receiving a scholarship,
Musk transferred to the University of Pennsylvania.
He graduated in 1995 with bachelor’s degrees in both
economics and physics, and he then moved to California to
pursue a Ph.D. in applied physics and material sciences at
Stanford University.5
After only two days, Musk left graduate school to start Zip2, an
online provider of content publishing soft-
ware for news organizations, with his brother, Kimbal Musk.
Four years later, in 1999, computer-maker Compaq
acquired Zip2 for $341 million (and was, in turn, acquired by
HP in 2002). Elon Musk then moved on to co-found
PayPal, an online payment processor. In 2002, eBay acquired
PayPal for $1.5 billion, netting Musk an estimated
$160 million.
Elon Musk is widely viewed as the world’s premier innovator,
bringing computer science, engineering, and
manufacturing together to solve some of today’s biggest
problems such as sustainable energy and transport as
well as affordable, multi-plenary living in order to avoid future
human extinction. Musk describes himself as “an
engineer and entrepreneur who builds and operates companies to
solve environmental, social and economic chal-
lenges.”6 At one point, Elon Musk led three companies
simultaneously that address the issues that are at the core of
his identity and vision: Tesla (sustainable transport), SolarCity
(decentralized and sustainable energy), and SpaceX
(multi-planetary existence).
Musk indeed has a larger than life profile and has been
described as “Henry Ford and Robert Oppenheimer

in one person,”7 as well as “Tony Stark, the eccentric inventor
better known as Iron Man.”8 In fact, Musk made a
cameo appearance in the Iron Man 2 movie. In line with his
movie avatar, the real-life Elon Musk plans to retire
on Mars.
Elon Musk’s larger-than-life personality frequently spills over
into his Twitter feed. Musk is an avid tweeter and
has had some run-ins with the Security and Exchange
Commission (SEC) in the past, as some of his tweets led
Telsa, Inc.
3 Copyright © 2020 McGraw-Hill Education. All rights
reserved. No reproduction, distribution, or posting online
without the prior written consent
to movements in Tesla stock.9 The SEC claims that some of his
tweets contain material information that has no
basis in fact, and thus crossed the line into securities fraud. The
issue came to the fore when Musk tweeted in the
summer of 2018: “Am considering taking Tesla private at $420.
Funding secured.” The day prior to Musk’s tweet,
Tesla’s share price was $380. The SEC filed securities fraud
charges against Musk claiming that his social media
statements were false and misleading investors.

In the spring of 2019, Musk settled with the SEC whom Musk
called on Twitter the “Short seller Enrichment
Committee”—referring to investors that short Tesla’s stock.
Short-sellers are investors, who borrow shares, sell
them, and then plan to buy them back at a lower price later to
profit from the difference. Essentially, short-sellers
are betting against a company, and through their short-selling
activities put downward pressure on the company’s
share price. Elon Musk and Tesla each had to pay a fine of $20
million to settle with the SEC, without admitting
to wrongdoing. Moreover, Musk had to agree that any of his
future social media statements needed to be reviewed
internally by Tesla prior to posting. Musk also had to step down
from the position as chairman of Tesla’s board of
directors, but he was allowed to continue to serve as CEO.
Yet, the “taking Tesla private” tweet by Elon Musk in the
summer of 2018 marked the beginning of one of the
most challenging years in the company’s brief history. After
missing production and delivery targets in the first two
quarters of 2019, Tesla’s market cap hit a low of $32 billion,
down from $65 billion a year earlier. And, the company
was running low on cash. The short-sellers thought they had
won, and Tesla would go bankrupt.
Rather than going bankrupt, however, in early 2020, Tesla, Inc.
employed about 50,000 people worldwide and
boasted a market capitalization of $150 billion, an appreciation
of more than 6,000 percent over its initial public
offering in 2010. As a consequence, Tesla’s shares
outperformed the broader market by a large margin. The differ -
ence in performance between Tesla and the broader stock
market has been more pronounced since the fall of 2019
as Tesla began to exceed performance expectations in
subsequent quarters (Exhibits 3 and 4).

Brief History of Tesla, Inc.
Tesla, Inc. (TSLA) was founded in 2003 in San Carlos,
California with the mission to design and manufacture
all-electric automobiles. Indeed, the company was inspired by
GM’s EV1 electric vehicle program in California in
the late 1990s, which the Detroit automaker shut down in 2003.
For a comparison between all-electric vehicles
(EVs) and plug-in hybrid electric vehicles, see Exhibit 5.
Tesla, Inc. is named after Nikola Tesla, the engineer and
physicist who invented the induction motor and
alternating-current (AC) power transmission for which he was
granted a patent in 1888 by the U.S. Patent and
Trademark Office. The Serbian-born inventor was a
contemporary of Thomas Edison. Indeed, Edison, the prolific
inventor of the light bulb, phonograph, and the moving picture
(movies), was at one-point Tesla’s boss. The two
geniuses fell out with one another and feuded for the rest of
their lives. Edison won the famous “War of Currents”
in the 1880s (DC vs. AC) and captured most of the limelight.
Because Nikola Tesla’s invention of the alternating-
current (AC) electric motor was neglected for much of the 20th
century and he did not receive the recognition he
deserved in his lifetime, Elon Musk is not just commercializing
Tesla’s invention but also honoring Nikola Tesla
with the name of his company. Tesla Inc.’s all-electric motors
and powertrains build on Tesla’s original inventions.
From day one, Elon Musk was also the controlling investor in
the original company, Tesla Motors, Inc., provid-
ing $7 million from his personal funds to get the company
started. Tesla confronted a major cash crunch in 2007,
which put the future viability of the company into question.
Musk stepped up and invested over $20 million in
this round to keep the company afloat, and his dream of

transition to sustainable transport alive. In total, Musk
4
Tesla, Inc.
provided $50 million from his own money to fund Tesla in its
early days because finding any outside funding was
next to impossible for a new car company during the global
financial crisis.10
Tesla’s Secret Strateg y (Part 1)
In a blog entry on Tesla’s website in 2006, Elon Musk explained
the startup’s initial master plan:11
1. Build a sports car.

2. Use that money to build an affordable car.
3. Use that money to build an even more affordable car.
4. While doing the above, also provide zero-emission electric
power generation options.
5. Don’t tell anyone.
To achieve Step 1, Tesla held a design contest for the styling of
its first product—Roadster. Lotus Cars, a British
manufacturer of sports and racing cars, won the contest. Lotus
Cars and Tesla Motors, as it was known then,
jointly engineered and manufactured the new vehicle using the
Lotus Elise platform. In 2006, Time magazine
hailed the Tesla Roadster as the best invention of the year in the
transportation category. In 2007, Musk was named
“Entrepreneur of the Year” by Inc. magazine.
In the same year, however, it became clear that the production
of the Roadster was not scalable. After taking a
closer look at Tesla’s financial situation, Musk found that Tesla
was losing $50,000 on each car sold. Tesla’s CEO
at the time, Martin Eberhard had led investors to believe that
the manufacturing of the Roadster cost $65,000 per
car, which appeared to justify the $92,000 sticker price. Musk
found that it cost Tesla $140,000 just for the parts,
subassemblies, and supplies to make each vehicle and that the
Roadster could not even be built with Tesla’s current
tools. He also discovered major safety issues with the existing
design. Completely taken aback by the messy state of
affairs, Musk commented, “We should have just sent a $50,000
check to each customer and not bothered making
the car.”12
In 2007, Elon Musk fired Martin Eberhard unceremoniously and

took over the engineering himself. Almost
every important system on the initial Roadster, including the
body, motor, power electronics, transmission, battery
pack, and HVAC, had to be redesigned, retooled or switched to
a new supplier. Such dramatic changes were neces-
sary to get the Roadster on the road at something close to the
published performance and safety specifications, as
well as to cut costs to make it profitable.
By 2008, Tesla Motors was finally able to relaunch an improved
version of its Roadster, and thus fulfill the first
step of its initial strategy laid out two years earlier. The
Roadster is a $115,000 sports coupé with faster acceleration
than a Porsche or a Ferrari. Tesla’s first vehicle served as a
prototype to demonstrate that electric vehicles can be
more than mere golf carts.
After selling some 2,500 Roadsters, Tesla discontinued its
production in 2012. The initial Roadster manufactur-
ing process was not scalable (with a maximum production rate
of no more than two cars per day) because it was
a handcrafted car put together at a former Ford dealership near
the Stanford University campus.13 Tesla learned
that it is better to build an electric vehicle from scratch rather
than retrofit a given car platform created for internal
combustion engines (ICE). The Roadster 1 had no more than 7
percent of parts in common with the Lotus Elise.
https://campus.13
https://crisis.10

Telsa, Inc.
As a side project, in 2017, Tesla unveiled Roadster 2, a sports
coupé that set new records for a vehicle to be
driven on public roads: It goes from 0–60 mph in 1.9 seconds
and from 0–100 mph in 4.2 seconds, with top
speeds of well above 250 mph. The base price of the new
Roadster model is $200,000. First customer deliveries are
expected in the second half of 2020.
In Step 2, Tesla focused on its next car: the Model S. This was a
car that Tesla designed from scratch with the
idea to create the best possible EV that is also scalable for mass
production. The Model S is a four-door family
sedan, with an initial base price of $73,500. Depending on the
size of the battery pack (up to 100kWh), the range
of the vehicle is between 210 and 330 miles. Unveiled in 2009,
the Model S appeals to a much larger market than
the Roadster, and thus allows for larger production runs to drive
down unit costs. When deliveries began in 2012,
the Model S received an outstanding market reception. It was
awarded the 2013 Motor Trend Car of the Year and
also received the highest score of any car ever tested by
Consumer Reports (99/100). By the end of 2019, it had sold
more than 300,000 of the Model S worldwide (Exhibits 1 and
5).
In 2012, Tesla unveiled the Model X, a crossover between an
SUV and a family van with futuristic falcon-wing

doors for convenient access to second- and third-row seating.
The Model X has a similar range as the Model S, with
between 250-330 miles per charge. Technical difficulties with
its innovative doors, however, delayed its launch until
the fall of 2015. The initial base price of the Model X was
$80,000, with the signature premium line ranging from
$132,000 to $144,000, thus limiting its mass-market appeal. By
the end of 2019, it had sold more than 150,000 of
the Model X worldwide (Exhibits 1 and 5).
Tesla also completed Step 3 of its master plan. In 2016, the
electric carmaker unveiled its first mass-market
vehicle: the Model 3, an all-electric compact luxury sedan, with
a starting price of $35,000 for the entry-level model
with a range of 250 miles per charge. Many want-to-be Tesla
owners stood in line overnight, eagerly waiting for Tesla
stores to open so that they could put down their $1,000 deposits
in order to secure their spots on the waiting list
for the Model 3—a car they had never even seen, let alone ever
taken for a test drive. As a result of this consumer
enthusiasm, Tesla received more than 500,000 preorders before
the first delivery, and thus $500 million in interest-
free loans. Despite initial difficulties in scaling-up production,
deliveries of the Model 3 began in the fall of 2017.
By the end of 2019, Tesla had delivered more than 450,000 of
the Model 3 globally.
Step 4 of Musk’s initial master plan for Tesla aims to provide
zero-emission electric power generation options.
To achieve this goal, Tesla acquired SolarCity, a solar energy
company, for $2.6 billion in 2016. With the acquisition
of SolarCity, to which Musk is also chairman and an early
investor, Tesla, Inc. is the world’s first fully integrated
clean-tech energy company, combining solar power, power
storage, and transportation. In the process, Tesla’s mis-
sion also changed from “to accelerate the advent of sustainable

transportation” to “accelerate the advent of sustain-
able energy,” thereby capturing the vision of a fully integrated
clean-tech energy company.
Step 5: “Don’t tell anyone”—thus the cheeky title of Elon
Musk’s original blog post: “Tesla’s Secret Strategy.”
Tesla completed an initial public offering (IPO) on June 29,
2010, the first IPO by an American automaker since
Ford in 1956. On the first day of trading, Tesla’s market
capitalization stood at $2.2 billion; less than 10 years later,
it stood at $150 billion. Despite significant future growth
expectations reflected in Tesla’s stock price appreciation,
the company is still losing a significant amount of money: $900
million in 2015; $675 million in 2016; almost $2
billion in 2017; close to $1 billion in 2018; and over $860
million in 2019. Tesla’s revenues in 2019 were $24.5 bil -
lion, up from $21.5 billion in 2018 and $11.8 billion in 2017.
Exhibit 2 provides an overview of Tesla’s key financial
data, 2015–2019.
6
Tesla, Inc.

Tesla’s Secret Strateg y (Part 2)
In 2016, 10 years after Tesla’s initial “secret strategy,” Elon
Musk unveiled the second part of his master plan for
the company (“Master Plan, Part Deux”) to continue the pursuit
of its vision “to accelerate the advent of sustain-
able energy.”14 Again, Tesla’s CEO and co-founder Elon Musk
detailed a set of stretch goals:
1. Create stunning solar roofs with seamlessly integrated battery
storage.
2. Expand the electric vehicle product line to address all major
segments.
3. Develop a self-driving capability that is 10 times safer than
manual via massive fleet learning.
4. Enable your car to make money for you when you aren’t
using it.
In the updated strategy, Step 1 leverages the 2016 acquisition of
SolarCity. Tesla, Inc. has morphed from a
manufacturer of all-electric cars into one of the first fully
integrated sustainable energy companies, combining
energy generation with energy storage, while providing zero-
emission vehicles.
In Step 2, Elon Musk is planning to expand the lineup of Tesla’s
electric vehicles to address all major segments,

including compact SUVs, pickup trucks, and heavy-duty semis.
In 2019, Tesla launched the Model Y, a compact luxury SUV
that is a smaller and much lower-priced version of
the Model X, starting at $39,000 (and a 230-mile range). The
first customer deliveries for the Model Y are planned
for spring 2021. A longer-range (280 miles) and thus a higher-
priced version of the Model Y starting at $60,000
will be available in the fall of 2020. Customer demand for the
Model Y is expected to be even stronger than that for
the Model 3, the compact luxury sedan.
In late 2019, Elon Musk set out to change the paradigm of what
a pickup truck should look like and how it
should perform by unveiling the Cybertruck. Musk emphasized
that the pickup truck concept had not changed in
the past 100 years and that he decided to do something
completely different. Musk reminded the audience that he
does zero market research whatsoever, but rather he starts with
a clean slate by using a first-principles approach to
design products (that is, reducing problems to the fundamental
parts that are known to be true, and start building
from there).
Musk designed a truck of the future the way he thought it
should look and perform. The result is a futuristic-
looking triangular truck using exoskeleton of ultra-hard
stainless steel, which provides more interior room and also
higher passenger safety and great vehicle durability. Customer
deliveries are planned for late 2021, with the base
model of the Cybertruck achieving a range of 250 miles per
charge, and starting at $40,000. The high-end version
of the Cybertruck equipped with the performance tri-motor and
a range of 500 miles per charge starts at $70,000
(Exhibit 7). Demand for the Cybertruck appears to be high
because Tesla received over 250,000 preorders just

within a few days of introducing the futuristic vehicle.
Moving into the pickup truck segment of the car industry is a
bold move by Tesla because pickup trucks account
for roughly one-third of sales of the incumbent carmakers (GM,
Ford, and Fiat Chrysler Automobiles) in the
United States, but generate some two-thirds of profits. Also,
unlike other models, the Cybertruck is unlikely to
cannibalize an existing segment in which Tesla has vehicles in
its lineup. In contrast, sales of Model S and Model
X vehicles fell sharply after the Model 3 was introduced, and
are expected to drop further with the Model Y intro-
duction. No such direct cannibalization is expected with the
Cybertruck as the customer profile of truck buyers is
generally quite different from those buying high-performance
luxury sedans or SUV crossovers.
Telsa, Inc.
Overall, Tesla has made significant improvements along a
number of important performance dimensions in its
vehicle line since customer deliveries for its first mass -
produced car, the Model S, began in 2012. Exhibit 8 details

Tesla’s product improvements over time by comparing the 2012
Model S with the 2021 Cybertruck.
In Step 3, Tesla is aiming to further develop the self-driving
capabilities of its vehicles. The goal is to make self-
driving vehicles 10 times safer than manual driving, and thus
being able to offer fully autonomous vehicles (Exhibit
9).
Fully autonomous driving capabilities are required for Tesla to
fulfill Step 4 of the new master plan: Turn your
car into an income-generating asset. Musk’s goal is to offer an
Uber-like service made up of Tesla vehicles, but
without any drivers. On average, cars are used less than three
hours a day. The idea is that an autonomous-driving
Tesla will be part of a shared vehicle fleet when the owner is
not using their car. This will drastically reduce the
total cost of ownership of a Tesla vehicle, and it will also allow
pretty much anyone to ride in a Tesla because of
the sharing economy.
Tesla’s Manufact uring
When Tesla began selling its first Roadster model in 2008, it
was plagued with both thorny technical problems
and cost overruns. The fledgling startup managed to overcome
these early challenges, in part by forming strategic
alliances. Tesla entered alliances with premier partners in their
respective category: Daimler in car engineering
(2009; discontinued in 2014); Toyota in lean manufacturing
(2010; discontinued in 2016); Panasonic in batteries
(2014; ongoing).
The alliance with Toyota brought other benefits for Tesla
besides learning large-scale, high-quality manufactur-
ing from the pioneer of lean manufacturing in the car industry.

It enabled Tesla to buy the former New United
Motor Manufacturing Inc. (NUMMI) factory in Fremont,
California in 2010. The NUMMI factory was created as
a joint venture between Toyota and GM in 1984. Toyota sold
the factory to Tesla in the aftermath of GM’s Chapter
11 bankruptcy in 2009.
The NUMMI plant was the only remaining large-scale car
manufacturing plant in California and is a mere 25
miles from Tesla’s Palo Alto headquarters. Tesla manufactures
both the Model S and the Model X in its Fremont,
California factory. The Model S and Model X share the same
vehicle platform and about 30 percent in parts.
Unlike more traditional car manufacturers that outsource
components and tend to rely heavily on third-party
suppliers, Tesla is largely a vertically integrated company.
Tesla put a value chain together that relies on integration
from upstream research and development, as well as battery and
electric vehicle manufacturing to downstream bat-
tery software; hardware chip and software design for full self-
driving (FSD) capability and autopilot; and provides
a dense network of supercharging stations (complimentary for
all Model S and Model X owners), as well as sales
and service—all done in-house. Tesla’s proprietary
supercharging station network allows seamless coast-to-coast
travel with any Tesla vehicle.
Gigafactor y 1 (Giga Nevada)
Tesla also made serval multi-billion-dollar commitments when
building super large manufacturing facilities,
dubbed “gigafactories.” Elon Musk describes a gigafactory as
the machine that builds the machine. Tesla Gigafactory
1 (or, Giga Nevada) is a 1,000-acre facility near Reno, Nevada,
with an option to expand by an additional 2,000

acres. The new factory required a $5 billion investment and
produces an estimated 50 GWh/year of battery packs
8
Tesla, Inc.
with a full capacity of 150 GWh/year. Giga Nevada’s current
output can produce 500,000 battery packs per year,
with an estimated 1.5 million battery packs upon final
completion, in late 2020.
Lithium-ion batteries are the most critical and the most
expensive component for electric vehicles, and each car
requires a battery pack. Tesla also uses battery packs for its
Powerwall (residential use) and Powerpack (commer-
cial use) product offerings, which are rechargeable lithium-ion
battery stations that allow for decentralized energy
storage and use. The production of batteries in Gigafactory 1 is
done in collaboration with Panasonic of Japan, a
global leader in battery technology.
Gigafactor y 2 (Giga New York)

Giga New York is photovoltaic (PV) factory by Tesla’s
subsidiary, SolarCity, and located in Buffalo, NY.
Gigafactory 2 produces Tesla’s Solar Roof, among other solar
panel products. Tesla’s Solar Roof is a system of
interconnected solar cells that mimic the look of more
traditional roofing systems. Basically, the entire roof func-
tions like one large solar panel and the energy generated can be
stored in the Tesla Powerwall and used for residen-
tial consumption.
Gigafactor y 3 (Giga Shanghai)
Requiring an investment of $2 billion, Giga Shanghai is Tesla’s
first production facility outside the United
States. Rather than import vehicles from the United States to its
main international markets in China and Europe,
Tesla has begun to build factories in the respective overseas
markets.
The design idea behind Giga Shanghai is to co-locate a car
manufacturing facility and battery pack production.
Tesla’s new factory in China was completed in record time—it
took less than 12 months from breaking ground in
January 2019 until locally produced Model 3s were rolling off
the production line. Benefiting from China’s lifting
of restrictions for foreign producers in the auto industry, Tesla
is the sole owner and operator of Giga Shanghai,
without any local joint-venture partner as previously required.
Sole ownership of Gigafactory 3 enables Tesla to
protect its proprietary technology and know-how.
The production target for its new Gigafactory is to produce
500,000 Model 3s per year for the Chinese market,
rather than import vehicles from the United States. Foreign
imports into the Chinese markets of the same model

(Model 3) are more expensive due to taxes and tariffs, but also
due to higher cost of manufacturing in the United
States. Some estimates put the production cost of a Model 3
made in China at $10,000 (or some 30 percent) lower
than producing the same Model 3 built in the United States.15
Gigafactor y 4 (Giga Berlin)
In late 2019, Elon Musk announced building Giga Berlin
(Gigafactory 4) in Germany. Giga Berlin places Tesla
in the heart of the leading European carmakers, who are located
in Germany including the Volkswagen Group
(owner of the VW, Audi, Porsche, and several other brands),
BMW, and Daimler, the maker of the Mercedes line of
cars, among other models. Tesla committed an investment of
$4.5 billion and hopes to have Giga Berlin completed
by the end of 2021. Tesla is using the same template as in Giga
Shanghai by producing both battery packs and
vehicles in the same location. The plan is to produce 150,000
Model Ys annually by that time, and when reaching
full capacity of 500,000 units per year.
https://States.15
Telsa, Inc.

In 2020, Elon Musk announced that Tesla plans to have a
gigafactory on each continent to produce vehicles and
battery packs for the local markets. Musk also stoked
speculation of building Gigafactory 5 in the United States, in
particular, in Austin, Texas (“Giga Texas”).16
Tesla’s Business Model
Tesla’s business model differs from traditional car
manufacturers along several dimensions, including:
Direct-to-consumer sales via online website (www.tesla.com)
and company-owned delivery centers. Tesla does
not use car dealers, and therefore, it is running into legal
obstacles in some U.S. states. In 2019, Tesla announced
that car ordering is online only (to bring Model 3 costs down
further), only to later reverse that decision and keep
a number of physical show rooms open. The result of this
reversal is raising vehicle prices by an estimated 3 per -
cent.17 However, orders in physical stores are also made on the
Tesla website. The EV-startup company facilitates
online ordering by providing limited choices in colors, interiors,
wheels, and a few other customization options.
This makes online ordering simple, but also reduces
manufacturing complexity as the number of permutations are
somewhat limited, unlike more traditional carmakers.
“No-haggle” policy. Tesla does not discount vehicles. Prices of
vehicles are displayed on the website and are non-
negotiable. The no-discount policy holds even for Musk family
members as the CEO reiterated in a widely circu-
lated company memo after a family member had requested a
discount. Musk replied, “Go to Tesla.com, buy the car

online, and the price you see there is the family discount.”18
Low marketing expenses. For the first few years, Tesla’s
marketing expenses were zero dollars because Musk
believed that all of its resources should be focused on
improving the company’s product. Initial Tesla “marketing”
was word of mouth. Tesla also does not use any paid celebrity
endorsements that are common in the car industry.
Yet, the Tesla brand has a cult following not unlike Apple in its
early days; indeed, several marketing professionals
have created pro-bono Tesla high-end ad campaigns and
uploaded them on YouTube to show their enthusiasm for
the company’s vision. Indeed, media coverage is high on Tesla.
Although no longer zero in 2020, Tesla’s marketing
expenses are considered low by industry standards.
Social network. Rather than a simple communications device,
Tesla’s website (www.tesla.com) is a social media
platform where users can connect with each other and Tesla
itself. In addition, the website contains Elon Musk’s
blog where the CEO frequently shares company and technology
updates, or addresses concerns. Also, Elon Musk’s
Twitter account has over 31 million followers.
Software platform. All Tesla cars are linked through a 4G Wi -Fi
connection, with the base program provided
complimentary. The Tesla fleet is akin to a network of
autonomous vehicles. This allows Tesla’s car software, such
as the autopilot, to learn from other vehicles’ experiences, as
well as consider real-time road and weather condi-
tions. Moreover, by early 2020, Tesla remains the only car
manufacturer globally that provides over-the-air (OTA)
software updates for its owners; and thereby incrementally
upgrading the vehicles through continued new software
releases. Software tweaks may cover fundamental issues such as
improving the autopilot or more mundane things

such as viewing YouTube videos in HD while waiting at a
supercharger.
Vehicle service model. While OTA software updates have
upended much of the traditional service model done by
car dealers, Tesla is also changing other aspects of the vehicle
service model. Given the large number of vehicles
on the road (the average car in the United States is 11 years old,
and many cars are driven for 20 years), the legacy
carmakers make a tidy profit by selling spare parts to car
dealers, body shops, and other car repair facilities to keep
www.tesla.com
https://Tesla.com
www.tesla.com
https://Texas�).16
10
Tesla, Inc.
their large fleet of vehicles operating. Tesla has produced less
than 800,000 vehicles in total; more importantly, EVs

need a lot less service than traditional ICE cars.
Open-source innovation. Tesla shares its patents with any
interested party to help set a new standard in the car
industry. Elon Musk’s goal is to accelerate the transiti on to
sustainable transport through electrification of vehicles.
Moreover, he believes that an open-source mindset at Tesla will
encourage continued innovation and helps in
attracting the best engineering talent globally. The Tesla CEO
believes that patents tend to hinder innovation rather
than help to spur it. In a 2014 blog post, Musk wrote:
Tesla Motors was created to accelerate the advent of sustainable
transport. If we clear a path to the
creation of compelling electric vehicles, but then lay
intellectual property landmines behind us to inhibit
others, we are acting in a manner contrary to that goal. Tesla
will not initiate patent lawsuits against
anyone who, in good faith, wants to use our technology. … Our
true competition is not the small trickle
of non-Tesla electric cars being produced, but rather the
enormous flood of gasoline cars pouring out
of the world’s factories every day. … We believe that Tesla,
other companies making electric cars, and
the world would all benefit from a common, rapidly-evolving
technology platform.19
Tesla’s Diversif ication
Besides expanding its electric vehicle product line to all major
segments (Step 2, Master Plan 2), including a
pickup truck, a commercial semi-truck, and a new roadster,
Tesla is diversified into other business activities.
Tesla Energy. Tesla is leveraging its technological expertise in
batteries, thermal management, and power

electronics that were originally developed for its EVs in energy
storage. Combining these various technologies
allows Tesla Energy to offer decentralized energy generation
(through Solar Roof), energy storage (Powerwall and
Powerpack), and energy use. Revenues of Tesla Energy were
$1.6 billion in 2018, up from $181 million in 2016.
Tesla provides energy generation via its innovative Solar Roof
that looks like regular shingles but cost less, all
things considered, and last longer. For residential consumers,
Tesla offers its Powerwall that allows customers to
store the solar energy captured on their roofs for later use.
Energy generation, therefore, becomes decentralized.
This implies that consumers can generate and use energy
without being dependent on any utility and can sell back
excess energy to utilities. The idea is that consumers will
generate not only energy for the use of their Tesla cars but
also enough to cover the energy needs of their entire house.
Tesla Energy is offering a similar solution on a larger
scale for commercial use. Decentralized energy generation and
storage (Powerpack) is of interest to commercial
customers in order to avoid power outages and to use renewable
energy.
Tesla Artificial Intelligence. Tesla has considerable expertise in
artificial intelligence (AI), derived not only from
its large drove of data but also from its acquisition in 2019 of
DeepScale, a machine learning startup with a focus on
computer vision. To provide autonomous-driving capabilities of
its vehicles (Level 5 in Exhibit 9), Tesla integrates
hardware and software. In particular, it developed its own AI-
chip that affords full self-driving (FSD) capabilities
(“autonomy”).
Elon Musk credits autonomy and electrification as the two
technological discontinuities that allowed Tesla to

enter the car industry, and to scale-up to mass production.20
Tesla’s fleet of EVs is a network of interconnected
autonomous vehicles that have accrued 18 billion miles of real -
world driving. Tesla applies machine learning algo-
rithms (i.e., artificial neural networks) to these data to
constantly improve and upgrade its autopilot (“massive
fleet learning”; Step 3, Master Plan 2). Alphabet’s Waymo,
Tesla’s closest competitor in autonomous driving has
accrued a total of 20 million miles in real-world driving.
https://production.20
https://platform.19
Telsa, Inc.
Tesla Insurance. Leveraging vast amounts of fine-grained data,
the EV company now offers Tesla Insurance,
which offers micro-targeted pricing for car insurance to Tesla
owners. Lower insurance premiums correlate with
more use of the autonomous-driving mode, a safer way of
driving (Step 3, Master Plan 2). Given that all vehicles

are connected, Tesla has a log of each driver’s trips, including
how many miles driven, and when and where. Tesla
also knows whether the driver obeys the speed limit and other
rules of the road. Given that Tesla has complete
transparency on how the vehicle is being used, it can fine-tune
the respective car insurance premium to a much
greater degree than traditional car insurers.
Tesla and the Compet it ion
Tesla’s competition can be grouped around three categories:
1. the new-vehicle market overall, that is, EVs and ICE cars
combined;
2. EV segment only; and
3. geography.
In the overall market for new cars, including both traditional
ICE cars and EVs, the market share for electric
vehicles remains small, with 2 percent in the United States, 1.5
percent in Europe, and 5 percent in China. In the
EV-only segment, however, Tesla is a leader with an 18 percent
market share globally (in 2019). Indeed, Tesla’s
Model 3 was the most-sold EV worldwide (in 2018), and the
more expensive Model S and Model X coming in
at fourth and fifth, respectively (Exhibit 10). At the same time,
demand for the Nissan Leaf (EV) and the Toyota
Prius (PHEV), both popular in the early 2010s, has declined. In
terms of geography, the United States, China, and
Europe are Tesla’s most important markets.
United States
Exhibit 11 shows annual new vehicle sales (ICE cars and EVs

combined) in the United States over time. In
the past five years, Americans purchased some 17 million new
vehicles per year. The average sales prices of new
vehicles have increased to $36,700, driven in large part by
strong demand for SUVs and pickup trucks. At the same
time, gas prices in the United States have remained low,
hovering around $2.50 a gallon (Exhibit 12).
Over the past decade, the legacy carmakers in the United States
(GM, Ford, Fiat Chrysler) have dedicated on
average no more than 10 percent of their total R&D spending on
electrification and autonomy. Much of the profits
in the industry were derived from strong demand in SUVs and
pickup trucks (with traditional ICEs).
Although Tesla’s market share in the total new-vehicle market
remains small, it is the leader in EV car sales in
the United States. Tesla’s Model 3 alone, a luxury compact
sedan, holds some 60 percent market share in the EV
segment. There are some questions whether the demand for
Tesla Model 3 will remain strong because Tesla no
longer qualifies for any federal credits. When first enacted, the
federal tax credit for the purchase of new electric
and plug-in hybrid electric vehicles of $7,500 was to phase out
gradually once a car manufacturer sold more than
250,000 EVs and PHEVs (Exhibit 5). In 2009, the limit per car
manufacturer was reduced to 200,000. Both, Tesla
as well as GM have crossed that threshold.
General Motors Co. (GM) is the largest of the legacy carmakers
(in terms of the number of cars sold) in the
United States, with some 17 percent market share and revenues
of $138 billion in 2019. GM’s market cap stood
at some $50 billion (in early 2020). In 2019, GM sold 7.7
million vehicles globally, with 2.9 million vehicles in the

12
Tesla, Inc.
United States. GM vehicle sales are declining in the United
States after weaker demand for some pickup trucks and
SUVs. GM’s net income was $6.7 billion in 2019.
GM’s efforts in the electrification of vehicles have been met
with mixed results. In the 1990s, GM was a leader
in electric car technology with its innovative EV1. To respond
to environmental regulation in California, GM
launched the fully electric car in 1996 with a range of up to 105
miles per charge. The EV1 was the first electric
vehicle designed and produced by a mass car manufacturer
making about 1,100 EV1. When California’s zero-
emission mandate was revoked, GM recalled and destroyed its
EV1 cars and terminated its electric-vehicle program
in 2002. Ever since GM has been lagging in electrification. The
Detroit automaker did not have a suitable car in its
lineup to compete against the popular Toyota Prius (PHEV) or
the Nissan Leaf (EV). The launch of the Chevy Volt
(PHEV) was delayed by over a decade because GM had to start

its electric-vehicle program basically from scratch.
In 2017, GM introduced the Chevy Bolt, an all-electric vehicle
with a 230-mile range on a single charge and start-
ing price of $37,500. Yet, the Chevy Volt is not selling well,
with 16,500 units sold in 2019, down 9 percent from
18,000 in 2018 and down 30 percent from 21,500 sold in 2017.
GM CEO Mary Barra believes that electrification autonomy are
of great importance to GM’s future. In 2020,
GM committed $20 billion to develop electric and autonomous
vehicles, planning to have at least 20 EVs in its
lineup by 2023. GM also announced (in 2020) to have achieved
a breakthrough in battery technology, which allows
EVs to travel over 400 miles on a single charge.21
To make progress in autonomy, in 2016, GM acquired Cruise, a
startup focused on self-driving car technology,
for $1 billion. In the same year, moreover, GM invested $500
million to form an equity alliance with Lyft, the
second-largest ride-hailing company in the United States after
Uber. GM wants to be in the mobile transportation
and logistics space because the age-old private car ownership
model is likely to shift in favor of fleet ownership
and management. Consumers will rent a car for a specific ride,
rather than own a car as a fixed asset. Private cars
in the United States are used no more than 5-10 percent of the
time, and sit idle for most of the day. Car owners
have the fixed costs of purchasing a car, buying insurance, and
maintaining the car. Lyft, in turn, has an alliance
with Waymo (a subsidiary of Alphabet, the parent company of
Google), one of the leaders in an autonomous car
technology venture.
Ford Motor Co. (F) is the second-largest incumbent carmaker in
the United States, with some 14 percent market
share and revenues of $155 billion (2019). Ford’s market cap

stood at some $30 billion (in early 2020). In 2019,
Ford sold close to 5 million vehicles globally, with 2.4 million
vehicles in the United States. Ford’s vehicle sales,
however, are declining in the United States in the SUV segment,
which more than offset gains in the market for
pickup trucks. In 2018, Ford announced that it will stop making
most of the sedans in its lineup in order to focus
on pickup trucks and SUVs, and thus basically exiting a market
segment that is created with its iconic Model T
first sold in 1908.
Ford derives two-thirds of its sales from the U.S. market and a
bit over 20 percent from Europe. In recent years,
Ford lost money in all of its non-U.S. operations; only North
America was profitable consistently because of its
popular F-series pickup trucks. Both GM and Ford are faced
with weakening demand for their vehicles and rising
labor costs in the United States.
Since 2016, Ford has invested some $12 billion into
electrification. In 2019, the Detroit automaker made a high
profile move into electrification by introducing the Ford
Mustang Mach-E, with customer deliveries commencing
in 2020. Depending on the size of the battery pack, the Mach-E
has a range of 240 to 300 miles per charge. Given
that the torque of an electric motor is available instantaneous,
the Mustang Mach-E is expected to accelerate much
more quickly than the Porsche Macan, one of the most popular
compact luxury SUVs. The Mach-E will thus com-

https://charge.21
Telsa, Inc.
pete head-on with the Tesla Model Y, which has somewhat
favorable specs in terms of range, performance, and
price.
When launching its new compact luxury sedan, Elon Musk had
initially planned to use the name Model E for
what is now known as the Model 3. With the Models S and X
already on the market and the Models E and Y
next in line, Musk got a kick out of the notion that Tesla lineup
of the first four vehicles would read S-E-X-Y. Ford
thwarted Tesla’s plans, however, because it had trademarked
Model E for vehicles in 2001, and refused to sell it
to Tesla. Now, Ford was using its iconic Mustang brand, as well
as its trademarked Model E for its new EV: the
Mustang Mach-E, signaling the importance of this all-electric
vehicle to Ford’s future. Given the small number of
EV vehicles sold by Ford, buyers of the Mustang Mach-E will
be able to claim a $7,500 tax credit; thus effectively
lowering the starting price of $44,000.
In the field of autonomy, Ford acquired a majority stake in Argo
AI, an artificial intelligence startup. To further
the development of its autonomous-driving technology, Ford
plans to invest $1 billion into Argo AI over the next
five years.

Rivian Automotive, Inc. In 2019, Ford invested $500 million
into the EV-startup Rivian. Other investors in
Rivian include Amazon.com, which invested $700 million, also
in 2019. The Michigan-based EV company is best
known for its R1T pickup truck and R1S SUV. Rivian’s niche
focus is on “adventure vehicles” that will provide off-
road capabilities and an overall ruggedness lacking in today’s
commercial vehicles, combined with a 400-mile range
per battery charge. The EV startup was founded in 2009 by RJ
Scaringe, who graduated from the Massachusetts
Institute of Technology with a doctorate in mechanical
engineering.
It is not clear, however, whether Rivian will morph into a stand-
alone car designer and mass-manufacturer of
electric vehicles such as Tesla or whether it will become more
of an EV platform company because of its propri-
etary “skateboard platform” design, which features a large
battery under the floor of a chassis upon which original
equipment manufacturer (OEM) car brands can build different
models. These vehicles would be zero-emission and
fully electric because they are powered by a drivetrain and four
electric motors, one for each wheel, all developed by
Rivian. The R1T pickup truck and R1S SUV might serve as
prototypes for such an EV platform strategy. Rivian’s
CEO RJ Scaringe also indicated that the company would license
its technology to other companies. Rivian did also
announce, however, that it starts planning to sell its all -electric
pickup truck by late 2020, and is accepting customer
deposits on its website (https://rivian.com). It also purchased a
former Mitsubishi Motors plant in Illinois, which
has the capacity to produce 250,000 vehicles a year.
Fiat Chrysler Automobiles (FCA). With GM and Ford, Fiat
Chrysler makes up the Big Three American car

manufacturers. FCA has a 12 percent market share and $122
billion in revenues (2019). Close to 70 percent of
FCA’s revenues are in North America, mostly from its iconic
Jeep line of vehicles and the Ram pickup trucks. In
2019, Fiat Chrysler sold some 4.4 million vehicles globally,
with 2 million vehicles in the United States. FCA’s
market cap stood at some $26 billion (in early 2020).
In December 2019, FCA announced that it intends to merge with
European PSA Group, maker of the Peugeot
brand of cars. The newly combined company would be the third-
largest car company globally, just behind the
Volkswagen Group and Toyota, but ahead of Nissan-Renault. As
an integrated company, FCA and Peugeot would
be selling close to 9 million cars and with revenues of some
$190 billion.
In early 2020, FCA also announced a joint venture to develop
electric vehicles with Foxconn, an electron-
ics OEM, best known for its assembly of Apple iPhones in
China. In 2016, Foxconn bought Sharp, a Japanese
electronics company, in order to offer its products under the
Sharp brand. FCA is not only a latecomer to vehicle
https://rivian.com
https://Amazon.com
14

Tesla, Inc.
electrification and autonomy, but continues to invest less in
these two new technological discontinuities than other
legacy carmakers.
China
In terms of units, China, with some 26 million new vehicles
sold in 2019, is by far the largest car market glob-
ally. In 2019, however, new vehicle sales in China were down
by 8.2 percent from 28 million cars sold in 2018.
Moreover, the Chinese car market is highly competitive with a
large number of smaller car manufacturers offering
low-priced basic vehicles combined with super-demanding
customers. Also notable is that while luxury SUVs and
sedans such as Porsche, Audi, or BMW are priced similarly as
in the United States, some 25 percent of the Chinese
new car market consists of local brands that sell new, low -end
cars for less than $12,000 (a market segment that
does not exist in the United States or Europe).22
The Chinese authorities consider EV manufacturing a
strategically important industry that will help to achieve
the China 2025 industrial plan, providing a road map for global
leadership in a number of important high-tech sec-
tors. Some estimates are that China spent more than $60 billion
to jump-start domestic EV production, including
research-and-development funding, financing for battery-
charging infrastructure, and tax exemptions. China is the
largest EV market globally. In 2019, 1.3 million EVs were sold

in China, in comparison to 330,000 EVs sold in the
United States (Exhibit 13). The Chinese market is almost four
times the size of the EV market in the United States,
albeit the average price point for new vehicles is generally
lower.
There are a large number of EV manufacturers in China. Most
of them focus on the low end of the market, in
particular, to help consumers take advantage of EV-quotas and
other tax incentives that the Chinese authorities put
in place to foster domestic production of all-electric vehicles
and to reduce air pollution. Most domestic EV produc-
ers are unlikely to survive without government support and
other incentives. Given that some of the tax incentives
granted for the purchase of all-electric vehicles have been
reduced by the Chinese authorities recently, demand for
EVs in China has fallen by 4 percent in 2019. Previously, the
year-over-year growth rate for new EV sales in China
ranged from 50 percent to 367 percent from 2013 to 2018
(Exhibit 14).
The Chinese market provides significant growth opportunities
for Tesla. In the luxury EV segment, in particu-
lar, Tesla’s main Chinese competitor is NIO, which delivered
20,000 EVs in 2019. Unlike Tesla, NIO outsources
procurement of battery packs and focuses mainly on design and
marketing. NIO EVs have an appealing design, and
the company provides an upscale customer experience (“NIO
houses”). NIO is also active in Formula-E racing.
However, some EV-industry experts consider NIO to be some
five years behind Tesla in technology development,
however.24
One challenge that Tesla and other EV makers face in China
(and Europe) is the fact that in the United States,
most Tesla owners tend to charge their vehicles at home in their

own garages. Given the long-range of Tesla vehicles,
commuters in the United States begin each day with a full
charge, negating the need to charge at work. In contrast,
in China (and Europe) most people live in apartment complexes
in more densely populated areas, often with no
private parking options. This makes the need for a dense public
charging station network much more pressing and
can limit the EV adoption rate in China (and Europe). Tesla is
building its own proprietary supercharging network
globally.
In terms of U.S. competitors, only GM plays a role in the
overall Chinese car market. Prior to its bankruptcy
filing in 2009, GM was mainly focused on the U.S. domestic
market. Now, close to 60 percent of GM’s revenues
come from outside the United States. The Chinese market is
becoming increasingly important to GM’s perfor-
mance, already accounting for greater than 40 percent of total
revenues. This number has risen steadily in the past
https://however.24
https://Europe).22
Telsa, Inc.

few years. Unlike some of the other U.S. carmakers, GM
entered the Chinese market earlier. In 1997, GM formed a
joint venture with Shanghai Automotive Industrial Corp.
(SAIC), one of the “big four” Chinese carmakers and one
of the largest companies worldwide. In 2018, GM sold more
cars in China than it did in the United States. GM’s
market share is 14 percent. GM’s China operation has been
cost-competitive from their entry into the market.
In contrast, Ford’s current presence in China is negligible, with
less than 2 percent market share in 2019, down
from 5 percent in 2016. FCA’s position in China is even weaker
with less 1.8 percent market share in 2019. The
hope is that FCA’s joint venture with the electronics company
Foxconn will allow FCA to penetrate the Chinese
market more in the future.
Europe
Around 15 million new cars are sold a year in Europe. The
European car market is one of the most fragmented
in highly developed economies, with a number of smaller local
competitors such as Alfa Romeo, Fiat, Peugeot,
Renault or Opel, while at the same it is also home to some of
the world’s best-known car brands such as Porsche,
Audi, and VW (all part of the Volkswagen Group), Daimler, and
BMW. Following losses over many years, GM has
exited the European car market altogether. It stopped production
of cars in Russia in 2015, and sold its European
subsidiaries, Opel and Vauxhall, in 2017. As of 2020, Europe
also has the smallest percentage of EV vehicles on
the road (1.5 percent) among the big three car markets (Europe,
China and the United States), in part, due to the
lack of public charging options and perceptions that ICE cars
are superior.

The Volkswagen Group is the world’s largest car manufacturer
by volume (over 10 million vehicles in 2019).
Following the diesel emissions scandal (“Dieselgate”) in which
VW engineers installed so-called defeat devices
in all its smaller (2.0 liter) TDI engines beginning with the
model year 2009, the Volkswagen Group was in a
crisis. These defeat devices were software codes contained in
the car’s onboard computer. The computer was pro-
grammed to detect when the car was being tested for emissions
by assessing a host of variables, including whether
the car was moving or stationary, whether the steering wheel
was being touched, what the speed and duration of
the engine run were, and so forth. This sophisticated defeat
device allowed the vehicles to pass the required and
rigid U.S. emissions tests. In reality, however, the vehicles
equipped with TDI engines actually exceeded the limits
for pollutants by up to 40 times during use. Between 2009 and
2015, VW sold 500,000 TDI vehicles equipped with
defeat devices in the United States and a total of 11 million
worldwide. Dieselgate cost VW $25 billion in fines and
legal settlements, not to mention the loss of reputation.
With a new top management team and Dieselgate as a catalyst
for a strategic transfor mation, the Volkswagen
Group not only streamlined its production to become more cost-
competitive but perhaps, more importantly, it led
to a strong commitment to electrification and autonomy. VW is
dedicating $30 billion by 2023 to produce fully
connected EVs, as well as building out a network of public
charging stations.25 The Volkswagen Group announced
a commitment to producing some 20 million EVs by 2025 or
some 20 percent of its current total production. By
2030, VW plans for its lineup of vehicles to be 40 percent fully
electric. Some of the first full-electric vehicles avail-
able are the Audi e-tron (2018) and the Porsche Taycan (2019).

The VW Group plans to launch some 70 new EV
models by 2028. The conglomerate targets to be fully CO2-
neutral by 2050.
Abstracting from VW’s luxury brands such as Porsche and
Audi, the vast number of vehicles in VW’s lineup is
more complementary to Tesla’s higher-end vehicles such as the
high-volume, lower-priced Volkswagen brand. And
with Giga Berlin, Tesla made a significant strategic
commitment to Germany and Europe, allowing it to plan for a
capacity of 500,000 locally-produced vehicles, mostly Models S
and Y.
https://stations.25
16
Tesla, Inc.
VW’s strong strategic commitment to electrification and
autonomy could be a tipping point for their widespread
adoption in Europe. Uncertainties remain concerning consumer
reception as well as government regulations in
Germany and across the European Union. Globally, carmakers

have committed a combined $225 billion to electri-
fication and autonomy by 2023, with the largest number of new
EVs expected in China and Europe.26
Coming Home
As the Gulfstream G700 was touching down in Palo Alto’s
private airport, Elon Musk woke up from his slumber
and wondered how he would scale-up production profitably
given that the company had several new vehicles in
development and was planning to build new gigafactories across
the globe. It was clear to him that demand would
need to remain strong in Tesla’s three key markets: the United
States, China, and Europe.
While scaling-up production of Model 3 was “hell,” he just had
promised 500,000 vehicles to be produced in
2020. Scaling-up customer service and avoiding long wait times
for repairs after a fender bender, for instance,
appeared to be even more difficult. Musk also was impatient in
bringing about the transition to all-electric vehicles
to replace internal combustion engines, and thus to promote
sustainable mobility . . . in the meantime, his phone
kept buzzing, and be began to read more about the new
coronavirus and the resulting production slow-down in Giga
Shanghai, the place he had just left so ebullient a mere 12 hours
earlier . . .
https://Europe.26

Telsa, Inc.
EXHIBIT 1 Tesla Total Vehicle Deliveries, 2012–2019*
450,000
400,000
367,500 66
350,000
300,000
250,000
245,200 200
200,000
150,000
103,100
100,000 76,200 0
50,000
2,600
22,400 32,000
50,000
0

2012 2013 2014 2015 2016 2017 2018 2019
* dotted trendline.
Source: Depiction of publicly available data.
17
18
Tesla, Inc.
EXHIBIT 2 Tesla Financial Data, 2015–2019 ($ millions,
except EPS data)
Fiscal Year 2015 2016 2017 2018 2019
Cash and short-term investments 1,219.54 3,498.74 3,523.24
3,878.17 6,514.00
Receivables–total 168.96 499.14 515.38 949.02 1,324.00

Inventories–total 1,277.84 2,067.45 2,263.54 3,113.45 3,552.00
Property, plant, and equipment–total
(net)
5,194.74 15,036.92 20,491.62 19,691.23 20,199.00
Depreciation, depletion, and
amortization (accumulated)
422.59 947.10 1,636.00 1,887.79 2,154.00
Assets–total 8,067.94 33,664.08 8,067.94 33,664.08 34,309.00
Accounts payable 916.15 1,860.34 2,390.25 3,404.45 3,771.00
Long-term debt 2,040.38 6,053.86 9,486.25 9,454.06 11.634.00
Liabilities–total 6,961.47 17,117.21 23,420.71 23,981.97
26,842.00
Stockholders’ equity–total 1,083.70 4,752.91 4,237.24 4,923.24
6,618.00
Sales (net) 4,046.03 7,000.13 11,758.75 21,461.27 24,578.00
Cost of goods sold 3,122.52 5,400.88 9,536.26 17,419.25
18,355.00
Selling, general, and
administrative expense
922.23 1,432.19 2,476.50 2,834.49 2,646.00
Income taxes 13.04 26.70 31.54 57.84 110.00

Income before extraordinary items -888.66 -674.91 -1,961.40 -
976.09 -862.00
Net income (loss) -888.66 -674.91 -1,961.40 -976.09 -862.00
Earnings per share (basic)
excluding extraordinary items
-6.93 -4.68 -11.83 -5.72 -4.92
Earnings per share (diluted)
excluding extraordinary items
-6.93 -4.68 -11.83 -5.72 -4.92
Source: Tabulation of publicly available data.
Telsa, Inc.
EXHIBIT 3 Tesla’s (Normalized) Stock Performance since
Initial Public Offering vs. Dow Jones Industrial
Average (DJIA), June 29, 2010–February 14, 2020

Tesla Price % Change Sep 28 ‘17 1.32K%
Dow Jones Industrials Level % Change Sep 28 ‘17 126.8%
8.00K%
6.42K%
197.8%
6.00K%
4.00K%
2.00K%
0.00K%
2012 2014 2016 2018 2020
20
Tesla, Inc.

EXHIBIT 4 Tesla’s Market Capitalization and Key Events, June
2019 – February 2020
+$100
bn
Tesla Inc Market Cap
Tesla Cap
160.00B
Tesla Cap
Market
$32bn
(6/19),
52-week
low
Q3 results,
Tesla beats
expectations
Q4 results,
Tesla beats
expectations
Giga Shanghai
opens
Market
$135bn
(2/20)

134.84B
120.00B
80.00B
40.00B
Jul ‘19 Sep ’19 Nov ‘19 Jan ’20
Telsa, Inc.

EXHIBIT 5 All-Electric Vehicles vs. Plug-in Hybrids
Electric Vehicles (EVs). All-electric vehicles use batteries as
the sole power source to supply
the electric energy needed for propulsion. Leveraging the fact
that electric motors can also act
as generators, electric vehicles utilize regenerative braking to
save a significant portion of the
energy expended during acceleration, thus increasing the energy
efficiency of the vehicle. Pure
electric vehicles have a higher torque over a larger range of
speeds during acceleration compared
with internal combustion engines (ICE). Running and servicing
costs of EVs are significantly lower
than its gasoline-based counterparts, because electric motors
and powertrains have relatively
few moving pieces, compared with the hundreds of precision-
engineered parts necessary for an
internal combustion engine. All-electric vehicles are also quiet
and zero-emission.
The battery in all-electric vehicles remains the most expensive
part of the car and is subject to
deterioration over its lifetime. All-electric vehicles tend to be
heavy, and thus wear out tires much
more quickly. Given a limited energy-to-weight ratio, the
driving range of electric vehicles remains
somewhat limited, although has been improving over time. The
cost of lithium-ion battery packs
(which Tesla is using) is expected to come down over time due
to economies of scale and an
estimated 18% learning curve. This implies that each time the

output doubles, per unit cost ($/kWh)
falls by 18% percent.
Moreover, EVs need to rely on a network of charging stations to
overcome range anxiety by
consumers; many mass-market electric vehicles cannot drive as
far on one charge as gasoline-
powered cars can with a full tank of gas. Gas stations can be
found pretty much on any corner in
cities and every couple of miles on highways. As of 2019, most
EVs such as the entry-level Tesla
Model 3 or the Nissan Leaf have a range of 200 miles or less
per charge; about one half of the
range of an average ICE car.
Plug-in Hybrid Electric Vehicles (PHEV). Plug-in hybrid
electric vehicles rely on hybrid
propulsion, which combines an electric motor with an internal
combustion engine. PHEVs
attempt to combine the advantages of pure electric vehicl es but
to avoid the range-restriction
problem by using an additional gasoline-powered internal
combustion engine. PHEVs contain
a battery that stores electricity for the electric motor and can be
recharged. Because
the battery shares the propulsion load, hybrid engines are
significantly smaller than their
traditional gasoline counterparts, reducing vehicle weight. The
most popular PHEV globally
is the Toyota Prius with Toyota over 10 million cars sold since
first introduced in 1997.
Tesla’s CEO Elon Musk is a strong opponent of hybrid vehicles
because he believes that PHEVs
combine the disadvantages of both electric and gasoline-
powered vehicles, more than offsetting

the advantages that each type offers.* Musk argues that hybrids
are “bad electric cars” because
they must carry around an additional engine and drive train,
adding weight, cost, and additional
parts to maintain and repair. As such, the Prius in EV-mode
only has a range of no more than
25 miles. Musk also criticizes the combustion engines as too
small, anemic, and inherently less
efficient than full-size engines. Moreover, the combination of
these technologies in a single-vehicle
adds to the technological complexity, which increases cost,
error rates, and maintenance.
*Malone, M. (2009), “Uber Entrepreneur: An Evening with Elon
Musk,” April 8 http//bit.ly/2uv2HTI.
Source: Courtesy of F.T. Rothaermel.
Tesla, Inc.
EXHIBIT 6 Tesla Total Vehicle Deliveries by Model, 2016–
2019

450,000
400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000
0
Model S/X
Model 3
2019
Delivery Goal
360,000
2016 2017 2018 2019
22

Telsa, Inc.
EXHIBIT 7 Tesla Cybertruck Models (2021)
Cybertruck Models Range (mi) 0-60 (s) Top Speed
(mph)
Payload
(lbs)
Tow
Rating (lbs)
Price
Single Motor RWD 250+ 6.5 110 3,500 7,500 $39,900
Dual Motor AWD 300+ 4.5 120 3,500 10,000 $49,900
Tri Motor AWD 500+ 2.9 130 3,500 14,000 $69,900

24
Tesla, Inc.
EXHIBIT 8 Tesla Product Improvements over Time: Comparing
2012 Model S with 2021 Cybertruck
Cybertruck Models 2021 Cybertruck
Tri Motor
2012 Model S Performance 2021 Cybertruck vs
2012 Model S
Price $69,900 $92,400 24% cheaper
Acceleration (0-60 mph) < 2.9 second 4.4 second 34% faster
acceleration
Acceleration (quarter mile) 10.8 seconds 12.6 seconds 14%
faster quarter mile
Total Range 500+ miles 265 miles 89% more range
Range-Adjusted Cost 7.1 miles/$1k 2.9 miles/$1k 150% more
range per dollar

Supercharging Capacity 250 kW+ 120kW 108% higher charging
rate
Storage 100 ft^3 26ft^3 280% more storage
Seats (adults) 6 5 +1 Seat
Powertrain All-wheel drive Rear-wheel drive +AWD
Automation Full Self-Driving Hardware No Autopilot/FSD
Hardware +FSD
Telsa, Inc.
EXHIBIT 9 The Six Stages of Automation (Autonomous
Vehicles)
Level 0: No Automation. A human control all the critical
driving functions.

Level 1: Driver Assistance. The vehicle can perform some
driving functions, often with a single feature such as cruise
control.
The driver maintains control of the vehicle.
Level 2: Partial Automation. The car can perform one or more
driving tasks at the same time, including steering and
accelerating, but still requires the driver to remain alert and in
control.
Level 3: Conditional Automation. The car drives itself under
certain conditions but requires the human to intervene upon
request
with sufficient time to respond. The driver isn’t expected to
constantly remain alert.
Level 4: High Automation. The car performs all critical driving
tasks and monitors roadway condiions the entire trip and does
not
require the human to intervene. Self-driving is limited to certain
driving lcoations and environments.
Level 5: Full Automation. The car drives itself from departure
to destination. The human is not needed; indeed, human
intervention would introduce more errors than fully automated
driving. The car is as good or better than a human and sterring
wheels and pedals are potentially no longer needed in a vehicle.
Source: Adapted from definitions provided by U.S. National
Highway Traffic Safety Administration.

Tesla, Inc.
EXHIBIT 10 Electric Vehicle Sales Globally by Model (2018)*
Tesla Model 3
BAIC EC-Series
Nissan Leaf
Tesla Model S
Tesla Model X
BYD Zin PHEV
JAC iEV E/S
BYD e5
Toyota Prius PHEV
Mitsubishi Outlander PHEV
Renault Zoe
BMW 530e

Chery eQ EV
0 20,000 40,000 60,000 80,000 100,000 120,000 140,000
160,000
* includes Electric Vehicles (EVs) and Plug-in Hybrid Electric
Vehicles (PHEV)
26
Telsa, Inc.
EXHIBIT 11 Vehicle Sales in the United States, 1978–2019 (in
1,000 units per year)
19
80
19
85

19
90
19
95
20
00
20
05
20
10
20
15
18,000
17,000
16,000
15,000
14,000
13,000
12,000
11,000
10,000

9,000
8,000
19
78
19
79
19
81
19
82
19
83
19
84
19
86
19
87
19
88
19
89
19
91

19
92
19
93
19
94
19
96
19
97
19
98
19
99
20
01
20
02
20
03
20
04
20
06

20
07
20
08
20
09
20
11
20
12
20
13
20
14
20
16
20
17
20
18
20
19

27
Tesla, Inc.
EXHIBIT 12 Average Annual Gasoline Price in the United
States (dollars/gallon), 2000–2020*
$4.00
$3.50
$3.00
$2.50
$2.00
$1.50
$1.00
$0.50

$0.00
2000 ‘01 ’02 ‘03 ’04 ‘05 ’06 ‘07 ’08 ‘09 ’10 ‘11 ’12 ‘13 ’14 ‘15
’16 ‘17 ’18 ‘19 ’20
* dotted trendline.
28
Telsa, Inc.
EXHIBIT 13 Electric Vehicle Sales in China (by units), 2011–
2019
1,400,000
China Sales
2011 2012 2013 2014 2015 2016 2017 2018 2019
1,200,000

1,000,000
800,000
600,000
400,000
200,000
0
1.21 million
U.S. Sales
329,528
30
Tesla, Inc.

EXHIBIT 14 Electric Vehicle Sales Growth Rate in China
(percentage change from the previous year),
2013–2019
400%
367%
-50%
2013 2014 2015 2016 2017 2018 2019
50%
200%
67%
123%
62%
-4%
350%
300%
250%
200%
150%
100%

50%
0%
Telsa, Inc.

5
10
15
20
25
Endnotes
1 As quoted in “Biography Elon Musk” http://bit.ly/2vNrIK6.
2 On February 4, 2020, Tesla’s stock market valuation crossed
the $150 billion threshold.
3 As quoted in: T. Higgins and S. Pulliman (2018, June 27),
“Elon Musk Races to Exit Tesla’s Production Hell,” The
Wall Street Journal.
4 M. Belfore, (2007), “Chapter 7: Orbit on a Shoestring,” in
Rocketeers (New York: HarperCollins), 166–195.

This section draws on J. Davis (2009, Sept. 27), “How Elon
Musk Turned Tesla Into the Car Company of the Future,”
Wired Magazine; and M. Malone (2009), “Uber Entrepreneur:
An Evening with Elon Musk,” Churchhill Club event.
April 7, 2009, http://bit.ly/2uv2HTI.
6 E. Howell (2019, Aug. 20), “Elon Musk: Revolutionar y
Private Space Entrepreneur,” http://bit.ly/37Jur4n .
7 M. Malone (2009), “Uber Entrepreneur: An Evening with Elon
Musk,” Churchhill Club event. April 7, 2009, http://
bit.ly/2uv2HTI.
8 G. Fowler, “Being Elon Musk, Tony Stark of SXSW,” The
Wall Street Journal, March 9, 2013, http:// on.wsj.
com/161hD1P.
9 D. Michaels and T. Higgins (2019, April 26), “Musk, SEC
Reach Deal to End Court Fight Over Tesla CEO’s Tweets,”
The Wall Street Journal.
Sony Pictures Classics (2011), “Revenge of the Electric Car,”
Film Documentary.
11 E. Musk, “The Secret Tesla Motors Master Plan (Just
Between You and Me),” Tesla, August 2, 2006, http://bit.
ly/29Y1c3m .
12 M. Malone (2009), “Uber Entrepreneur: An Evening with
Elon Musk,” Churchhill Club event. April 7, 2009, http://
bit.ly/2uv2HTI and J. Davis (2009, Sept. 27), “How Elon Musk
Turned Tesla into the Car Company of the Future,” Wired
Magazine.
13 E. Musk in interview with Third Row Tesla Podcast, January
30, 2020, http://bit.ly/2vTNv2B.

14 E. Musk, “Master Plan, Part Deux,” Tesla Blog Entry, July
20, 2016, http://bit.ly/29QwI0X.
S. Hanley (2019, May 30). “Tesla Model 3 Made in China May
Cost $10,000 Less Than US-Built Cars,” CleanTech-
nica http://bit.ly/2vSO2lD .
16 E. Musk in interview with Third Row Tesla Podcast, January
17 C. Atiyeh (2019, March 11), “Tesla Reverses Course on
Store Closings, Will Raise Vehicle Prices Instead.” Car and
Driver http://bit.ly/2vQG7Fc .
18 D. Muoio (2017, Sept. 28), “Elon Musk won’t give family
members early access or discounts for a Tesla — including
his own mother,” Business Insider.
19 Excerpt from E. Musk (June 12, 2014): “All Our Patents
Belong to You,” https://www.tesla.com/blog/all-our-patent-
are-belong-you.
E. Musk in interview with “Third Row Tesla Podcast,” January
21 M. Colias (2020, Mar. 4), “GM Aims to Convince Wall
Street Skeptics Its Future Is Electric,” The Wall Street Journal.
22 A. Potter of Piper Jafray in interview with Rob Maurer’s
“Tesla Daily Podcast,” December 13, 2019
http://bit.ly/2OEJ6r3.
23 J. Whalen (2020, Jan. 17), “The next China trade battle could
be over electric cars,” The Washington Post.

24 A. Potter of Piper Jafray in interview with Rob Maurer’s
“Tesla Daily Podcast,” December 13, 2019 http://bit.
ly/2OEJ6r3.
Volkswagen Group Press Release (March 12, 2019),
“Volkswagen plans 22 million electric vehicles in ten years (by
2030)” http://bit.ly/38fK7NQ.
26 M. Colias (2020, Mar. 4), “GM Aims to Convince Wall
Street Skeptics Its Future Is Electric,” The Wall Street Journal.
http://bit.ly/38fK7NQ
http://bit
http://bit.ly/2OEJ6r3
http://bit.ly/2vTNv2B
https://www.tesla.com/blog/all-our-patent
http://bit.ly/2vQG7Fc
http://bit.ly/2vSO2lD
http://bit.ly/29QwI0X
http://bit
http://bit.ly/37Jur4n
http://bit.ly/2uv2HTI
http://bit.ly/2vNrIK6Structure BookmarksTesla, Inc. Elon
Musk: Engineer Entrepreneur Extraordinaire Brief History of
Tesla, Inc. Tesla’s Secret Strategy (Part 1) Tesla’s Secret
Strategy (Part 2) Tesla’s Manufacturing Gigafactory 1 (Giga
Nevada) Gigafactory 2 (Giga New York) Gigafactory 3 (Giga
Shanghai) Gigafactory 4 (Giga Berlin) Tesla’s Business Model

Tesla’s Diversification Tesla and the Competition United States
China Europe Coming Home EXHIBIT 1 Tesla Total Vehicle
Deliveries, 2012–2019* EXHIBIT 2 Tesla Financial Data,
2015–2019 ($ millions, except EPS data) EXHIBIT 3 Tesla’s
(Normalized) Stock Performance since Initial Public Offering
vs. Dow Jones Industrial Average (DJIA), June 29, 2010–
February 14, 2020 EXHIBIT 4 Tesla’s Market Capitalization
and Key Events, June 2019 – February 2020 EXHIBIT 5 All-
Electric Vehicles vs. Plug-in Hybrids EXHIBIT 6 Tesla Total
Vehicle Deliveries by Model, 2016–2019 EXHIBIT 7 Tesla
Cybertruck Models (2021) EXHIBIT 8 Tesla Product
Improvements over Time: Comparing 2012 Model S with 2021
Cybertruck EXHIBIT 9 The Six Stages of Automation
(Autonomous Vehicles) EXHIBIT 10 Electric Vehicle Sales
Globally by Model (2018)* EXHIBIT 11 Vehicle Sales in the
United States, 1978–2019 (in 1,000 units per year) EXHIBIT
12 Average Annual Gasoline Price in the United States
(dollars/gallon), 2000–2020* EXHIBIT 13 Electric Vehicle
Sales in China (by units), 2011–2019 EXHIBIT 14 Electric
Vehicle Sales Growth Rate in China (percentage change from
the previous year), 2013–2019 Endnotes
SHORT PAPER TITLE 2Full Title of the Paper
Author Names
University
Running head: SHORT PAPER TITLE 2
Abstract
What is the problem? Outline the objective, problem statement,
research questions and hypotheses. What has been done?
Explain your method. What did you discover? Summarize the
key findings and conclusions. What do the findings mean?
Summarize the discussion and recommendations. What is the
problem? Outline the objective, problem statement, research
questions and hypotheses. What has been done? Explain your
method. What did you discover? Summarize the key findings

and conclusions. What do the findings mean? Summarize the
discussion and recommendations. What is the problem? Outline
the objective, problem statement, research questions and
hypotheses. What has been done? Explain your method. What
did you discover? Summarize the key findings and conclusions.
What do the findings mean? Summarize the discussion and
recommendations. What is the problem? Outline the objective,
problem statement, research questions and hypotheses. What has
been done? Explain your method. What did you discover?
Summarize the key findings and conclusions. What do the
findings mean? Summarize the discussion and
recommendations. What is the problem? Outline the objective,
problem statement, research questions and hypotheses. What has
been done? Explain your method. What did you discover?
Summarize the key findings and conclusions. What do the
findings mean? Summarize the discussion and
recommendations.
Keywords: medical, bio, innovation, engineering
Table of Contents
Abstract 2
List of Figures 3
List of Tables 3
Heading 1 4
Heading 2 4
Heading 3. 4
Heading 4. 4
Heading 5. 4
Reference list 7
Appendix A 8
Appendix B 10
List of Figures
Figure 1. Example figure body text6
Figure A1. Example figure appendix9
Figure A2. Example figure appendix9

Figure B1. Example figure appendix10
List of Tables
Table 1 Example table body text6
Table A1 Example table appendix9Heading 1
This research aims to gain insight into the relationship between
smartphones and students’ attention in classrooms. This chapter
further discusses the research method, the sampling method and
the data analysis procedure. Heading 2
In addition to an extensive literature review, 40 interviews were
conducted for this study. The goal of conducting interviews was
to find out how students looked at the use of smartphones in the
classroom.
Heading 3. A non-probability sample was used to gather
participants for this research. The driving factors behind this
decision were cost and convenience.Heading 4. Students who
participated in this study were recruited through posts on the
school’s Facebook page. As an incentive, students who
participated were granted an exemption for writing an essay.
Heading 5.
Participants were selected based on their age and gender to
acquire a representative sample of the population. Furthermore,
students had to share additional demographic information.
Table 1 Example table body text
College
New students
Graduating students
Change
Undergraduate
Cedar University
110
103

+7
Elm College
223
214
+9
Maple Academy
197
120
+77
Pine College
134
121
+13
Oak Institute
202
210
-8
Graduate
Cedar University
24
20
+4
Elm College
43
53
-10
Maple Academy
3
11
-8
Pine College
9
4

+5
Oak Institute
53
52
+1
Total
998
908
90
Figure 1. Example figure body text
(Author, 2018)
References
AuthorLastName, FirstInitial., & Author LastName, FirstInitial.
(Year). Title of article. Title of Journal, Volume(Issue), Page
Number(s). https://doi.org/number
Appendix A
Table A1 Example table appendix
College
New students

Graduating students
Change
Undergraduate
Cedar University
110
103
+7
Elm College
223
214
+9
Maple Academy
197
120
+77
Pine College
134
121
+13
Oak Institute
202
210
-8
Graduate
Cedar University
24
20
+4
Elm College
43

53
-10
Maple Academy
3
11
-8
Pine College
9
4
+5
Oak Institute
53
52
+1
Total
998
908
90
Source: Fictitious data, for illustration purposes only
Figure A1. Example figure appendix
(Author, 2018)
Figure A2. Example figure appendix
(Author, 2018)
Appendix B
Figure B1. Example figure appendix
(Author, 2018)
Series 1 Category 1 Category 2 Category 3
Category 4 4.3 2.5 3.5 4.5 Series 2 Category 1
Category 2 Category 3 Category 4 2.4
4.4000000000000004 1.8 2.8 Series 3 Category 1
Category 2 Category 3 Category 4 2 2 3
5

4.4000000000000004 1.8 2.8 Series 3 Category 1
5
4.4000000000000004 1.8 2.8 Series 3 Category 1
5
4.4000000000000004 1.8 2.8 Series 3 Category 1
5

Tesla, Inc.
1. Within a few short months between the summer of 2019 and
early 2020, Tesla added more than $100
billion to its market cap. Do you believe that Tesla’s stock
market valuation making it the second most
valuable car company globally is rational or do you think it is a
hyped-up overvaluation?
2. What business model is Tesla pursuing? How is Tesla’s
business model different from traditional car
manufacturers?
3. Historically, the automotive industry in the United States has
been identified by high barriers to entry.
How was Tesla able to enter the automotive mass-market
industry?
4. What type of innovation strategy is Tesla pursuing? Tie your
explanation to Elon Musk’s “Master Plan,
Part 1.”
5. In which stage of the industry life cycle is the electric
vehicle industry? What core competencies are
the most important at this stage of the industry life cycle? What
are the strategic implications for the
future development of this industry?
6. Apply the Crossing-the-Chasm framework to explain some of
the challenges Tesla is facing and
provide some recommendations on how to overcome them.
7. Evaluate Elon Musk’s “Master Plan, Part 2” and assess if
Tesla can gain and sustain a competitive
advantage.

ECECS 472572 Final Exam ProjectRemember to check the errata

ECECS 472572 Final Exam ProjectRemember to check the errata

More Related Content

Similar to ECECS 472572 Final Exam ProjectRemember to check the errata (20)

More from EvonCanales257 (20)

Recently uploaded (20)

ECECS 472572 Final Exam ProjectRemember to check the errata