Disclosures & Disclaimer This report must be read with the disclosures and the analyst certifications in the Disclosure appendix, and with the Disclaimer, which forms part of it. Issuer of report: HSBC Securities (Taiwan) Corporation Limited View HSBC Global Research at: https://www.research.hsbc.com Developed by Apple, micro-LED offers superb image quality at a fraction of the power consumption; production could begin in… …2H17 for the Apple Watch; smartphone adoption feasible by 2019, low-cost iPhone a probable candidate; ideal for VR/AR Potentially large positive impact on display, semiconductor and LED if adoption expands beyond the Apple Watch; preferred stocks: Epistar and LG Display, both rated Buy Micro-LED is a new class of display technology that has superb image quality (equal or better to OLED and LCD on all metrics) and requires only a fraction (10-20%) of the power consumption of OLED and LCD. We believe the technology will be adopted in wearables initially but is scalable to larger applications and will be a highly viable solution for VR (virtual reality) and AR (augmented reality) applications, and transparent displays. Micro-LED displays can be curved but may not be bendable, which remains OLED’s unique strength. Micro-LED is the first major display technology developed by Apple (AAPL US, Not Rated), which oversees the entire supply chain from displays, semiconductors to LEDs. Production could start in 2H17 for an expected 1H18 launch of the Apple Watch 3 (Apple Watch 2 is likely out later this year, according to DigiTimes). Micro-LED’s superb power efficiency enables longer battery life or additional features, such as LTE. We estimate initial costs to be comparable to current OLED displays, which would then drop quickly (+20% per annum) as the LED chip size shrinks. While there are still technical barriers, micro-LED could be feasible for high-volume devices, such as smartphones, by 2019. However, we believe the deciding factor will come down to consumer preference: the functionality of the foldable/roll-able form factor of OLED vs the image quality and power consumption of micro-LED. The low-cost iPhone could be a potential adopter, in our view. Profound long-term impact: While we see a strong likelihood of micro-LED adoption in the Apple Watch, we make no estimate changes to the companies mentioned in this report, as Apple has yet to announce plans for the commercialisation of micro-LED. However, we believe the impact could be profound if micro-LED displays become a reality. For panel makers, it is a new generation display with innovation potential without requiring large capex; for LED makers, it has the potential to overhaul the product mix, with the high value-add portion becoming much more significant. 31 August 2016 Jerry Tsai* Analyst HSBC Securities (Taiwan) Corporation Limited [email protected]+886 2 6631 2863 Ricky Seo* Semiconductor Analyst The Hongkong and Shanghai Bank Corporation Limited, Seoul Securities Branch [email protected]+822 3706 8765 Steven Pelayo Regional Head of Technology Research The Hongkong and Shanghai Banking Corporation Limited [email protected]+852 2822 4391 David Huang* Associate HSBC Securities (Taiwan) Corporation Limited [email protected]+886 2 6631 2865 *Employed by a non-US affiliate of HSBC Securities (USA) Inc, and is not registered/ qualified pursuant to FINRA regulations Asia Display EQUITIES ELECTRONIC EQUIPMENT Asia Rating, target price and valuation summary Mkt cap 3M ADTV Price TP (lcy) Rating Upside/ Company Ticker USDbn USDm (lcy) downside 2016e PB (x) 2017e PB(x) 2016e ROE 2017e ROE LG Display 034220 KS 19.6 30 30,500 35,200 Buy 15.4% 0.8 0.7 5.5% 10.6% Epistar 2448 TT 0.7 15 21.60 30.00 Buy 38.9% 0.4 0.4 -3.3% 3.2% AUO 2409 TT 3.9 33 12.85 14.50 Buy 12.8% 0.6 0.6 2.8% 8.5% INX 3481 TT 3.7 23 11.65 14.30 Buy 22.7% 0.5 0.5 -1.6% 6.6% TPK 3673 TT 0.7 19 60.30 86.00 Buy 42.6% 0.7 0.6 -2.5% 9.4% TSMC 2330 TT 153.0 165 177.00 191.00 Buy 7.9% 3.3 2.9 24.8% 24.4% Source: Bloomberg, HSBC estimates. Priced as of close at 26 August 2016. Currency: TWD for Taiwan stocks; KRW for Korea stock. ADTV – Average daily traded volume. Micro-LED – the next generation display technology
37
Embed
Asia Display ELECTRONIC EQUIPMENT - jrj.com.cnpg.jrj.com.cn/acc/Res/CN_RES/INVEST/2016/8/31/04baf542-98c6-4a30... · (equal or better to OLED and LCD on all metrics) ... Apple defines
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Disclosures & Disclaimer
This report must be read with the disclosures and the analyst certifications in
the Disclosure appendix, and with the Disclaimer, which forms part of it.
Issuer of report: HSBC Securities (Taiwan) Corporation Limited
View HSBC Global Research at:
https://www.research.hsbc.com
Developed by Apple, micro-LED offers superb image quality at a
fraction of the power consumption; production could begin in…
…2H17 for the Apple Watch; smartphone adoption feasible by
2019, low-cost iPhone a probable candidate; ideal for VR/AR
Potentially large positive impact on display, semiconductor
and LED if adoption expands beyond the Apple Watch;
preferred stocks: Epistar and LG Display, both rated Buy
Micro-LED is a new class of display technology that has superb image quality
(equal or better to OLED and LCD on all metrics) and requires only a fraction
(10-20%) of the power consumption of OLED and LCD. We believe the technology
will be adopted in wearables initially but is scalable to larger applications and will be a
highly viable solution for VR (virtual reality) and AR (augmented reality) applications,
and transparent displays. Micro-LED displays can be curved but may not be
bendable, which remains OLED’s unique strength. Micro-LED is the first major
display technology developed by Apple (AAPL US, Not Rated), which oversees the
entire supply chain from displays, semiconductors to LEDs.
Production could start in 2H17 for an expected 1H18 launch of the Apple Watch 3
(Apple Watch 2 is likely out later this year, according to DigiTimes). Micro-LED’s
superb power efficiency enables longer battery life or additional features, such as
LTE. We estimate initial costs to be comparable to current OLED displays, which
would then drop quickly (+20% per annum) as the LED chip size shrinks. While there
are still technical barriers, micro-LED could be feasible for high-volume devices, such
as smartphones, by 2019. However, we believe the deciding factor will come down to
consumer preference: the functionality of the foldable/roll-able form factor of OLED vs
the image quality and power consumption of micro-LED. The low-cost iPhone could
be a potential adopter, in our view.
Profound long-term impact: While we see a strong likelihood of micro-LED
adoption in the Apple Watch, we make no estimate changes to the companies
mentioned in this report, as Apple has yet to announce plans for the
commercialisation of micro-LED. However, we believe the impact could be profound
if micro-LED displays become a reality. For panel makers, it is a new generation
display with innovation potential without requiring large capex; for LED makers, it has
the potential to overhaul the product mix, with the high value-add portion becoming
Source: Bloomberg, HSBC estimates. Priced as of close at 26 August 2016. Currency: TWD for Taiwan stocks; KRW for Korea stock. ADTV – Average daily traded volume.
Micro-LED – the next generation display technology
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
2
A game-changer quietly on the way
While LCD has been the mainstream display technology for over a decade, it is widely expected
that in the coming years OLED will take market share, especially in advanced mobile displays
(such as smartphones and wearables). OLED has advantages, such as a thin form factor, light
weight, excellent image quality (in terms of colour spectrum, grey scale and refresh rate), and the
highly intriguing potential to be bendable or even rollable. We expect OLED to make up 45% of the
smartphone market by 2018, with the expected iPhone adoption of OLED as the key catalyst.
However, we believe that in 2018 a new technology called micro-LED could emerge as another
choice for advanced display solutions, with supply chain production to commence in 2H17.
While the adoption of micro-LED could be limited to wearable devices (likely the Apple Watch)
in the beginning, we believe the technology has the potential to be adopted in smartphones and
in VR/AR applications in the long term, owing to the outstanding display quality (equal or
superior to OLED and LCD in all aspects), superb power consumption (requires only 10-20% of
the power consumption of OLED and LCD) and cost reduction potential (given production is
mostly done on fully depreciated assets with conventional materials, which can see steep drops
when the chip size shrinks).
What’s micro-LED?
While there is no official definition yet, Apple defines micro-LED as a LED chip of 1-100um in width
(most likely 10um or less). The size is extremely small compare to conventional LED: typically the
smallest LED on the market is around 100um (0.1mm, or the thickness of a human hair). In other
words, most micro-LED chips are less than 10% in width, while micro-LED chips are less than 1%
in terms of area, compared to the smallest LED chips commercially available today.
All display technologies generate images by altering the colour and brightness of pixels (for a FHD
screen with a 1980x1080 resolution, there are around 2m pixels, or 6m sub-pixels, on the screen).
Micro-LED – the next
generation display
technology
Superior display performance but requires only a fraction (10-20%) of
the power consumption of LCD and OLED
Developed by Apple and involves panel, LED, semiconductor and
OSATs; production could start in 2H17 for 2018 wearables
Technically feasible for smartphones by 2019, likely the most ideal
display option for VR and AR, and significant potential for transparent
displays
3
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Micro-LED offers a direct, simplistic approach to create images: tiny LED chips in blue, green and
red are bonded to a TFT-backplane, which controls the brightness of each LED chip to generate
images. The technology requires no backlight component and no colour filters (such as those in
LCD displays), nor does it need encapsulation or the control layers used in OLED displays.
LCD, OLED and micro-LED structures
Source: HSBC
We expect micro-LED technology to have a profound impact on the display supply chain.
It is the first major display technology not developed by a panel maker, but rather by
Apple.
Producing micro-LED not only requires involving panel makers, but also LED chip
makers, semiconductor foundry service providers and OSATs (outsourced
semiconductor assembly and test providers).
The incremental capex for panel makers as well as for most parts of the supply chain
is modest.
Keep in mind that there are several technology companies developing micro-LED technology.
However, due to the lack of scale and resources needed to manage the supply chain, most
companies are focusing on niche products that take advantage of micro-LED’s high brightness.
While Sony (6758 JP, Not Rated) already has a product in the form of a high-end but small volume
commercial display called CLEDIS that uses micro-LED principles, we believe Apple will be the
only company capable of commercialising micro-LED displays for high-volume applications in the
foreseeable future. We believe Apple is the only company that can do this due to its strong
bargaining power in the supply chain and deep knowledge on semiconductors and displays.
Other developers of micro-LED technology
Company X-Celeprint PlayNitride Leti Sony Mikro Mesa
Technology Focus on mass transfer of micro
devices with stamping technique
PixelLED display claims 99% yield in
transferring GaN devices
iLED matrix reached 10um pitch; uses
Quantum dot to realise full colour
Demoed FHD 55’ TV in 2012; launched
CLEDIS large-sized display in May
Developing micro device transfer tools
Source: HSBC
Micro-LED vs OLED and LCD
While LCD, OLED and micro-LED all share the same driving platform – the TFT-backplane,
micro-LED boasts even fewer components, compared to OLED, which is already a bit simpler
than LCD. Simple structures mean light emitted passes through fewer layers of obstructions
before reaching the viewer’s eyes and, as such, consume less energy.
Cover glass
Micro LEDs
TFT-backplane
Substrate
TFT-LCD
Liquid
crystal
Glass
Polarizer
Backlight unit
Glass
Polarizer
Color
Filter
Rigid OLED Curved OLED Foldable/flexible
OLED
Glass
Glass
Glass
Polyimide coating
OLED materials OLED materials
OLED materials
Cover glass
TFT-backplane
Substrate
Color filter
Polarizer
Liquid crystal
“On-cell” touchscreen with
ITO material patterned on
top of encap glass
Polarizer
Cover glass
“ITO film” touchscreen
on top of encap film
Polarizer
Cover glass
“On-cell” touchscreen with
new material patterned on
top of encap film
Polarizer
Colorless PI or plastic
Glass
TFT substrate
Encap glass
OLED materials
Rigid OLED
Polyimide coating
TFT substrate
Encap film
OLED materials
Curved OLED
Polyimide film or coating
TFT substrate
Encap film
OLED materials
Foldable OLEDBacklight
LCD
Substrate
Polarizer
OLED Micro LED
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
4
Performance comparison: LCD vs OLED vs micro-LED
LCD OLED Micro-LED
Mechanism Colour filter/backlight Self-emissive Self-emissive Luminous efficiency Low Medium High Brightness (cd/sqm) 3,000 1,000 100,000 Contrast ratio 1,000:1 10,000:1 1,000,000:1 Colour rendering 75% NTSC 124% NTSC 140% NTSC Viewing angle Ok Excellent Excellent Refresh rate Ms us ns Power consumption relative to LCD NA 30-40% lower vs LCD +90% lower vs LCD Operating temp (Celsius) -20 to 80 -30 to 70 -100 to 120 Flexibility Low High Medium Life span 60k hours 20-30k hours 80-100k hours
Source: IHS, Sony, HSBC
Far more energy efficient vs any available technology, with potential to widen the lead
Micro-LED consumes only 10-20% of the power required for LCD and OLED as light emitted
reaches the viewer’s eyes almost unhindered. This enables the key benefit of this technology:
longer battery life. Based on various studies on power consumption of smartphones, the power
used by display components constitutes 50-60% of the total energy consumption of the device.
In other words, battery life could easily be doubled when micro-LED displays are adopted.
Smartphone power consumption breakdown during video playback
Smartphone power consumption breakdown during web browsing
Source: NICTA, University of South Wales Source: NICTA, University of South Wales
While OLED materials are improving, LED still offers the best efficiency among all light source
materials (LCD uses conventional LED as a light source, but the complex structure blocks most
light). Some studies show micro-LED’s efficiency could be as much as 4x vs regular LEDs, as
the very small size allows the refraction of light to be focused in a single direction; as such,
there is no need for a lens, which is needed for regular LEDs, that blocks light. Furthermore,
micro-LED also requires much lower driving current vs regular LEDs.
0
120
240
360
480
Display GSM CPU RAM Graphics
Power (mW)
0
120
240
360
480
Display GSM CPU RAM Graphics
Power (mW)
5
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Superb energy efficiency of micro-LED compared to LCD and OLED can be improved
For example, if the energy efficiency of OLED and LED rises by 20%, the increased luminance
from OLED is still subject to more layers of light objections; therefore, the gain in efficiency at
the display level could be much less than 20% (the incremental gain for LCD could be even less
due to more hindrance). Meanwhile, almost all efficiency gains from micro-LED can reach the
viewer’s eyes.
Best possible picture quality
While OLED is known for vivid colour, sharp contrast, and a refresh speed that is 100x faster
than that of LCD, micro-LED performs better in all these areas. We believe the ability to control
each individual pixel, plus the fast on-off characteristic of LED, enables micro-LED displays to
achieve the strong performance. Meanwhile, micro-LED also offers a wide viewing angle (close
to 180 degrees), which allows the technology to be adopted for large-sized displays.
Sunlight readability has been the strong area for LCD, but micro-LED is even better due to its
superb brightness (30x brighter than LCD; 100x brighter than OLED), making it an ideal display
technology for wearable devices.
Significant potential of integrating touch, 3D touch and fingerprint functions
We believe micro-LED can incorporate touch sensing, 3D touch sensing, and fingerprint
sensing by planting infrared LED chips alongside micro-LED chips. Infrared LED would be able
to detect finger movement, as well as the ridges and valleys of the fingerprint. We believe the
potential for micro-LED incorporating these features is high, given that there is no need to plant
infrared LED at all pixels, but just selected pixels. This would simply complex processes and the
supply chain that are required to realise these functions. For example, we counted 4-5 types of
suppliers involved in completing the 3D touch component in iPhones.
Highly durable with a long life
As LED is a solid material, micro-LED is capable of withstanding impacts. Meanwhile, the LED’s
life can exceed 100,000 hours, 3-4x longer compared to OLED materials. In addition, there are
no burn-in issues to be considered.
Best technology to realise transparent displays
Due to extremely high emitting efficiency, micro-LED pixels can be made small but still generate
enough brightness to render an image. At below 10um, the chip’s width is only 10% of a normal
human hair and, as such, invisible to the human eye. If we consider the simple structure, we
believe micro-LED is the display technology most likely to be used in transparent displays.
Some flexibility in the form factor, but OLED is still likely the best
LuxVue, which is fully owned by Apple after the 2014 acquisition, has filed patents on making
flexible micro-LED displays. This means that displays can be made into a curved shape, which
allows the device to fit on the wrist, or have slanted edges, such as the latest Samsung (005930
KS, KRW1,645,000, Buy) Galaxy phones.
However, it is unknown whether micro-LED can be used for foldable displays, which can be
realised by OLED by 2017, in our view. Based on our understanding of the structure, it is
unlikely that micro-LED can withstand the repeat bending/rolling motion of the display. As such,
we believe OLED will still be the technology of choice when it comes to foldable displays.
Any risk in commercialisation?
While we believe micro-LED is a technology that can be ready for commercialisation in 2017-18, we
like to point out that the supply chain is complex and lengthy compared with that for typical displays;
meanwhile, every process is critical and, as such, managing all parts of the supply chain effectively
will be a challenge. We don’t rule out that there is a likelihood for delays. Also, if one or both of the
key technical targets, which are chip size reduction and transfer capacity increase, can’t be met, the
development of micro-LED could be suspended or even terminated, as the costs to produce such
display solutions would be exceedingly high to make commercialisation infeasible.
7
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Who are the players?
In 2014, Apple acquired LuxVue, which holds a wide range of micro-LED patents from forming
micro-LED structures to chip transfer. The acquisition provides the foundation of Apple’s
micro-LED technology and considerable resources have continued to be invested since 2014.
Producing micro-LED displays requires the collaboration of display, LED, semiconductor and
transfer of micro-LED chips. The latter is a completely new process where LuxVue likely holds
the most comprehensive and advanced technologies among all tech firms researching the
subject globally.
List of key LuxVue patents
Publish date Patent
5/23/2013 Light emitting diode structure 5/23/2013 Micro device array 8/15/2013 Method of transferring and bonding an array of micro devices 6/19/2014 Smart pixel lighting and display microcontroller 1/29/2015 Adhesive wafer bonding with controlled thickness variation 4/2/2015 Method and structure for receiving a micro device 7/30/2015 Micro device transfer head array 11/19/2015 Flexible display and formation with sacrificial release layer 12/3/2015 Interactive display panel with emitting and sensing diodes 1/7/2016 Mass transfer tool
Source: United States Patent and Trademark Office, HSBC
The major processes for micro-LED include:
LED epi-wafer: Depositing thin epitaxy layers on the substrate, using MOCVDs, which are
the same tools used in making typical LEDs. We believe LED makers with a large
capacity scale and strong experience in four-element LED (which emits the red light
needed for red pixels), such as Epistar (Buy) and Osram (OSR GR, EUR47.32, Hold),
could be the major suppliers.
Production could start in
2H17
Micro-LED could be adopted in the Apple Watch in 1H18; supply
chain may start production in 2H17
Production requires collaboration across industries; if this happens,
we believe potential beneficiaries include Epistar, TSMC, LG Display
and OSATs
Modest capex is required as existing capacity can be used; there is
considerable cost reduction potential with shrinkage of the chip size
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
8
LED chip process: The conductive pads (for electrical connection) and metal layers (for
protection and insulation) are added to the epi-wafer. Tiny LED chips (micro-LEDs) are
formed and bonded to a carrier substrate. We believe TSMC (Buy) is in the best position to
provide the chip process as it holds more 6” wafer capacity than any foundry.
Chip transferring: Micro-LEDs are lifted off the carrier substrate and transferred to the
TFT-backplane (likely supplied by LG Display (Buy), which has a long standing
relationship with Apple), which can drive and control the brightness of each micro-LED
chip individually. The transfer of the micro-LED chips is done in batches, likely c100,000
chips at once. It takes several iterations to fill up the surface of the backplane. We believe
OSATs will be performing the transfer process.
Major processes and likely candidates
Process/component LED epi-wafer Chip process TFT-backplane Transfer
Candidates Epistar, Osram TSMC LGD OSATs Reasons Combination of
production scale and technology; leaders in
fur-element (red) LEDs
Leading technology; capacity of 6”, the likely
size for initial micro-LED production
Decent backplane technology; strong
relationship with Apple
Tools to be designed by LuxVue and operated by
OSATs; strong know-how in testing and packaging,
as well as micro device bonding, are likely
required
Source: HSBC estimates
Micro-LED process flow chart
Source: HSBC estimates
While the manufacturer of the transfer tool is unknown at this point, we don’t see tool
manufacturing as a bottleneck for the Apple Watch, given that fewer than 500 transfer tools will
be needed, according to our calculations. Meanwhile, foundries, such as TSMC, could be
providing a key component called transfer head (which picks up micro-LEDs from the substrate
and then places the chips on the backplane), which is subject to replacement due to wear and
tear. Each transfer head is made up of numerous silicon electrodes that are manufactured with
a semiconductor process.
Apple Watch could be the first adopter of micro-LED
1H18 the likely time for the product, but the supply chain could start to get moving in 2H17
The first generation Apple Watch started to ship in April 2015. The replacement cycle for the
Apple Watch seems to be longer than the typical one-year cycle for smartphones. The second
generation Apple Watch will likely be announced in 2H16 (according to DigiTimes), implying an
18-month replacement cycle.
LED epitaxy wafer process are Chip processes to form Micro
performed in MOCVD, by LED are done by semi foundry
Epistar, Osram…etc like TSMC
Micro LEDs are transferred to
TFT-Backplane, likely by test
and packaging providers
TFT-Backplanes are produced
by LTPS. Most likely by LGD
9
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
As such, we believe the third generation Apple Watch could be launched sometime in 1H18. We
believe reducing the power consumption of the display will allow Apple to add key features, such
as LTE connectivity, and could boost the lacklustre sales of the Apple Watch. Given micro-LED is
a new technology and one that requires elaborated processes, we believe the supply chain could
start to commence production in 2H17 if the Apple Watch is indeed adopting micro-LED.
We believe the technology could be ready for commercialisation in 2H17
The capacity requirements of the supply chain are highly dependent on: 1) the micro-LED chip
size (bigger means much more LED/chip foundry capacity is needed), and 2) the transfer rate of
chips from carrier substrate to TFT-backplane.
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
10
Based on our analysis of LuxVue’s patent filings, we believe these key performance
metrics, which can make the technology feasible for production, can be achieved by 2H17:
1) 15um pitch of the micro-LED chips (this refers to the width of a chip, say 10um, plus the
spacing between the chips, say 5um), and 2) 100,000 chips per transfer.
As our analysis shows, achieving these metrics assumes sufficient LED and chip foundry
capacity is available, while capex on chip transfer stays manageable, which includes: 1) c40
MOCVDs, or 2-3% of global MOCVD capacity, 2) c13,000 6” wafer monthly foundry capacity, or
c7% of TSMC’s Fab 2 (the 6” fab) capacity, and 3) less than 400 transfer tools, which should
cost less than USD100m.
How micro-LED displays are made
Source: HSBC
LED epitaxy layers are deposited on substrate. Process done in the MOCVD.
Metal conductive layer
Metal layers are added, then Micro LED structures are formed by etching
This and next three processes are likely done in semi-conductor foundry
Dielectric layer
A dielectric layer is formed to protect the Micro LED structure.
This provides insulation so adjacent chips don't get disturbed during transfer
Turn over the substrate, then bond Micro LED to a carrier substrate
The adhesive layer will melt pass certain temperature, such as 200 degree Celsius
Grind away the LED substrate layer.
Adhesive layer will be heated so LEDs can be de-bonded from carrier substrate
A transfer tool with many transfer heads, is than put over the Micro LEDs array.
Transfer heads can be selectively charged to pick up certain LEDs by static force
Some adhesive will be attached to the bottom of LED
Transfer head then move the selected LED to the destinations on TFT-backplane
At least three transfers are needed for R (red), G (green), and B (blue) subpixels
Transfer head These subpixels form a pixel
One pixel (RGB) TFT-backplane
LED epitaxy layer
Substrate
Adhesive layer
Carrier substrate
Transfer tool
11
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Scenario analysis based on pitch assumptions: enough LED and semiconductor capacity for the Apple Watch
Micro-LED pitch assumptions 10um pitch 15 um pitch 20 um pitch
Total pixels (312 x 390) per Watch 121,680 121,680 121,680 Total sub-pixels (pixels x 3) per Watch 365,040 365,040 365,040 No. of micro-LED chips provided by each 4” wafer 63m 28m 15m No. of micro-LED chips provided by each 6” wafer 165m 73m 41m No. of Watch can be made from each 4” wafer 170 76 40 No. of Watch can be made from each 6” wafer 440 195 110 Assume at peak monthly run-rate of 2.5m Watches Monthly chip capacity needed, assume 4” wafer 15k wafers 33k wafers 62k wafers Monthly chip capacity needed, assume 6” wafer 6k wafers 13k wafers 23k wafers No. of MOCVD needed using 4” wafer 19 MOCVDs 42 MOCVDs 80 MOCVDs No. of MOCVD needed using 6” wafer 16 MOCVDs 36 MOCVDs 63 MOCVDs
Source: HSBC estimates
Our key assumptions are:
While current Apple Watch demand is around 10-13m units annually, we believe a monthly
run-rate of 2.5m micro-LED panels is likely needed for Apple to go ahead, given the higher
component built needed for the initial product launch and assembly yield losses.
We assume a higher resolution for all Apple Watches (312x390; there is another resolution
of 272x340) to ensure there is no supply constraint.
Two shifts per day for MOCVD running time.
Transfer tool calculation
Transfer tool calculation, based on a 2.5m watches peak monthly run-rate 75k per transfer 100k per transfer 150k per transfer
Assume 1 min per transfer 440 tools 330 tools 220 tools Assume 2 mins per transfer 880 tools 660 tools 440 tools Assume 5 mins per transfer 2,200 tools 1,650 tools 1,100 tools
Source: HSBC estimates
Assuming the number of LEDs that can be moved per transfer is indeed reaching 100,000, each
Apple Watch will take four transfers to complete. We estimate each transfer should take no more
than 1-2 minutes (most likely under 1 minute; see reasons below). This suggests only 3-400 tools
are needed, a manageable investment that should not exceed USD100m, in our view.
Based on the patents filed by Apple, the pick-ups of micro-LED chips are done with electrostatic
force; as such, we believe the transfer time will be kept short, otherwise the delicate power
control will need to be maintained for a long period of time. In addition, when LED chips are
lifted off from the carrier substrate, they also pick up some lingering adhesive (which is melted
during de-bonding). This lingering adhesive can be used to re-attach the chips onto the
TFT-backplane, before it is dried. This is another reason the transfer is unlikely to take a long
time as the adhesive could dry.
Strong cost reduction potential
Based on the 15um pitch/100,000 chip per transfer assumptions, we believe the cost of a
micro-LED display should be comparable to the OLED display cost of USD15 (display only, not
including touch, 3D touch and cover glass). However, we like to point out that the micro-LED
cost reduction is highly intriguing.
Micro-LED material cost could see a significant drop if the chip size shrinks, while the transfer
speed picks up. A 10um pitch would imply 7um of chip dimension (vs 10um for 15um); as such,
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
12
there will be a reduction of the chip size by 50%. Accordingly, the number of wafers, as well as
the LED epitaxy level and semiconductor foundry capacity required, will be reduced
significantly. Meanwhile, we believe our assumption of 1-2 minutes per transfer could be
shortened. We believe the magnitude of chip size reduction and transfer process improvement,
which lead to a 30% reduction in total display costs, can be achieve every 12-18 months.
Micro-LED display costs
(USD) 15um pitch/10k per transfer 10um pitch/2x transfer speed
LED Epi 2.88 1.50 Chip process 3.15 1.55 Transfer 4.00 2.50 Backplane 6.50 6.50 Total 16.53 12.05
Source: HSBC estimates
Currently, touch, 3D touch and cover glass for the Apple Watch cost USD7-8. As micro-LED
offers strong potential to embed these touch functions directly in the display, we believe the
potential for bringing the overall costs (display plus touch) meaningfully below the current costs
of USD23-24 is high.
Any potential losers?
We believe one company that needs to be watched is TPK (Buy), the current sole touch and 3D
touch component supplier for the Apple Watch. The company has provided the glass-on-glass
type touch solution for the current generation and will likely provide a touch-on lens-like solution
for the new models to be launched in 2H16.
We believe TPK still has a +50% likelihood of retaining the touch and 3D touch business for the
Apple Watch even if the device adopts micro-LED, as the focus will be on commercialising the
display technology (while integration will add complexity as an entry barrier). However, in the
longer term, we believe integration will be the future technology direction. The Apple Watch
generates 8-9% of TPK’s revenue in 2016, according to our estimates.
While some investors would point out that the introduction of micro-LED displays could diminish
LG Display’s value-add for the Apple Watch, we don’t see it as a negative event for the
company. The revenue contribution from the Apple Watch is modest for LG Display and being
the first to be involved in Apple’s proprietary display technology will likely bolster investor
confidence about LG Display’s positioning within the Apple supply chain, in our view.
13
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
We believe micro-LED displays could be adopted in the Apple Watch in 1H18 (display
production could start in 2H17) and become another option of advanced display technology
besides LCD and OLED. Naturally, the next question will be: Can micro-LED be adopted in the
iPhone, or other devices?
We believe that if there is demand for a display technology with superb power savings but
lacking the ability to be foldable, or rollable, micro-LED can be a highly suitable solution. In our
view, the technology would be considered proprietary to Apple, which is likely the only company
capable of managing the supply chain that spans a number of sectors in the foreseeable future.
Based on our analysis, by overcoming additional technical barriers (on top of what is needed to
be resolved for the Apple Watch adoption), micro-LED could be eventually ready for the high-
volume production required for the iPhone. Our best guess for the earliest timeline would be
2019. However, considering the low technical entry barriers and capacity requirements (which
would be substantial to realise the mainstream iPhone model adoption of micro-LED), a lower
volume, lower specification (QuadHD vs 4k2k) iPhone could be a possibility. We still assume
OLED will be the mainstream technology for the iPhone in the foreseeable future.
Adoption by the iPhone: achievable, given time but requires a lot of capacity
The production scale for the micro-LED commercialisation on the iPhone is far greater than for
wearables, by a factor of 60x, due to the much higher resolution (we assume in 2-3 years
premium phones will feature 4k2k resolution and, as such, more pixels per screen) and volume
vs the Apple Watch. Assuming sometime in the future Apple utilises OLED for half of the
mainstream iPhone models, while using a different technology for the other half (say around
120m units annually), our scenario analysis shows that micro-LED could be feasible, provided
the following technology barriers and bottlenecks are overcome.
Reduce the micro-LED chip by a factor of 10 vs the Apple Watch
Based on our calculations, a 15 um pitch (10um for LED chip plus 5um for the spacing with the
adjacent chip) is sufficient to commercialise micro-LED on wearable devices. To realise micro-
LED in iPhones, a pitch of no more than 5um is needed (3um chip plus 2um spacing). The chip
A viable technology for
smartphones
Potentially feasible by 2019 but need more breakthroughs in chip
size and transfer capacity; requires sizable LED capacity
OLED still remains the best option for foldable/rollable displays;
low-cost iPhone could be a strong possibility for micro-LED
Unique attributes make micro-LED a high value-add to VR and AR
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
14
size at 5um will be around 1/10th
of that for wearables. While this seems to be a challenging
task, we believe the micro-LED chip size is being shrunk by 50% every 12 months or so. As
such, we believe the 5um target could be reached by 2019.
Micro-LED pitch (chip width plus spacing) development timeline and applications
Our analysis shows at least more than 60% of micro-LEDs need to be produced on an 8” wafer,
due to the constraints of the 6” semiconductor foundry capacity. According to our estimates,
200,000 of 6” and 200,000 of 8” wafers (both on a monthly basis) will provide enough chips for
120m iPhones. Keep in mind that 20,000 6” wafers per month will be all of TSMC’s 6” capacity,
and 200,000 of 8” wafer capacity per month will be 40% of TSMC’s 8” capacity. While these
requirements seem to be high, we believe securing capacity is still possible, as these are no
advanced fabs (90% of TSMC revenue is generated from the 12” fab). Our cost analysis below
also shows that decent prices can be paid for the wafers.
Scenario analysis based on pitch assumptions: 4k2k smartphones
2 um pitch 5 um pitch 10 um pitch
Total pixels (4k x 2k) 8m 8m 8m Total sub-pixels (pixels x 3) 24m 24m 24m No. of micro-LED chips provided by each 4” wafer 1,570m 250m 63m No. of micro-LED chips provided by each 6” wafer 4,000m 660m 165m No. of micro-LED chips provided by each 8” wafer 7,100m 1,140m 285m No. of handset can be made from each 4” wafer 65.4 10.4 2.6 No. of handset can be made from each 6” wafer 166.7 27.5 6.9 No. of handset can be made from each 8” wafer 295.8 47.5 11.9 Assume at peak monthly run-rate of 14m handsets Monthly chip capacity needed, assume 4” wafer 214k wafers 1,344k wafers 5,333k wafers Monthly chip capacity needed, assume 6” wafer 84k wafers 509k wafers 2,036k wafers Monthly chip capacity needed, assume 8” wafer 47k wafers 295k wafers 1,179k wafers No. of MOCVD needed using 4” wafer 225 MOCVDs 1,400 MOCVDs 5,555 MOCVDs No. of MOCVD needed using 6” wafer 175 MOCVDs 1,060 MOCVDs 4,240 MOCVDs No. of MOCVD needed using 8” wafer 155 MOCVDs 980 MOCVDs 3,930 MOCVDs
Source: HSBC estimates
We believe the main challenge is that larger-sized wafers could make achieving the high uniformity
requirements difficult. Currently, globally, only less than 3% of LEDs are produced on 8” wafers
due to technical barriers, such as warping and the lack of incentives to upgrade equipment.
Meanwhile, 6” has become the mainstream size for LED wafers. LED makers need to gain more
experience with 8” wafers before meeting the high quality requirements of micro-LED.
15
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Global sapphire market shares (area) by substrate diameter 2016
Source: U.S. Department of Energy
Transfer technology needs to improve by a factor of 10x
For us, this would be the main challenge. We believe transfer capacity needs to be boosted by
as much as 10-20x from the 100,000 chips per transfer that can be achieved when the Apple
Watch ramps up to 1-2m chips per transfer. We believe the density of the electrodes on the
transfer head could be the main challenge, as we believe 50,000-100,000 electrodes could be
needed per cm2 (vs around 10,000 per cm
2 for the Apple Watch transfer head). This would
require delicate control of power management and distribution of the transfer head. Once the
capacity is achieved, the time to complete the chip transfer for a micro-LED smartphone display
will be 10-20 minutes, implying 5,000-10,000 tools are needed. This could be an investment of
USD0.5-1bn.
Support from the LED supply chain
We estimate that the number of MOCVD tools required is 1,000 or USD1bn in terms of
investment. While that does not sound like an exceedingly high amount of capex by “Apple
standards,” we believe it could be challenging to convince LED makers to allocate that level of
resources, given LED makers in general have been experiencing difficult times in recent years.
However, the bigger issue is that LED chip makers would be reluctant to make investments if
chip sizes continue to shrink, as the number of MOCVD tools needed would be drastically
reduced. As such, we believe that some sort of financial support or commitment to expand
micro-LED to other applications beyond wearables and smartphones is needed.
We see micro-LED as a highly viable option for low-cost iPhones
Apple has been offering low-cost smartphone models to capture emerging market opportunities.
Given micro-LED stands out in terms of cost-down potential, energy efficiency and durability, we
see micro-LED as a highly viable alternative to lower cost iPhone models.
If we relax the resolution requirement from 4k2k to Quad HD (2560x1440) and still assume the
5um pitch can be achieved, many hurdles could be lowered.
There is likely enough 6” foundry capacity to produce all the micro-LEDs needed (so there
is no need to move to 8” LED epi wafers; thus higher uniformity can be ensured). While the
235,000 wafers shown is slightly above TSMC’s 6” capacity of c200,000, a 14m handset
monthly run-rate is unlikely needed, as low-cost iPhones volume tends to be lower than that
of the mainstream models.
3%
47%39%
11%
8"
6"
4"
2"
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
16
Transfer capacity improvement over the Apple Watch needed could be only 3-4x compared
to the 10x with the 4k2k resolution.
MOCVD investment is half of what is needed for 4k2k.
Scenario analysis based on pitch assumptions: QuadHD smartphones
2 um pitch 5 um pitch 10 um pitch
Total pixels (QuadHD 2560x1440) 3.7m 3.7m 3.7m Total sub-pixels (pixels x 3) 11.1m 11.1m 11.1m No. of micro-LED chips provided by each 4” wafer 1,570m 250m 63m No. of micro-LED chips provided by each 6” wafer 4,000m 660m 165m No. of micro-LED chips provided by each 8” wafer 7,100m 1,140m 285m No. of handset can be made from each 4” wafer 141.4 22.5 5.7 No. of handset can be made from each 6” wafer 360.4 59.5 14.9 No. of handset can be made from each 8” wafer 639.6 102.7 25.7 Assume at peak monthly run-rate of 14m handsets Monthly chip capacity needed, assume 4” wafer 99k wafers 622k wafers 2,467k wafers Monthly chip capacity needed, assume 6” wafer 39k wafers 235k wafers 942k wafers Monthly chip capacity needed, assume 8” wafer 22k wafers 136k wafers 545k wafers No. of MOCVD needed using 4” wafer 105 MOCVDs 650 MOCVDs 2,570 MOCVDs No. of MOCVD needed using 6” wafer 85 MOCVDs 490 MOCVDs 1,960 MOCVDs No. of MOCVD needed using 8” wafer 75 MOCVDs 455 MOCVDs 1,815 MOCVDs
Source: HSBC estimates
What will it come down to?
We believe micro-LED holds strong potential in terms of cost reduction. For smartphone
micro-LED displays, where material and transfer costs as a percentage of total cost are even
higher than those for the wearables (70-80% vs 50-60%), such an advantage will be even more
pronounced. Keep in mind that micro-LED costs could fall by a significant increment every time
the chip size shrinks or when the transfer capacity improves. As such, the potential cost
reduction will likely be greater than the 10-15% annual declines of LCD and OLED costs. Also
keep in mind the ability to fully integrate touch, 3D touch or even fingerprint sensors could
provide more cost advantages.
Cost of micro-LED displays for smartphones
(USD) 5um pitch Micro LED for 4k2k smartphone
5um pitch Micro LED for qHD smartphone
LED Epi 17.88 10.37 Chip process 19.66 11.40 Transfer 10.00 5.50 Backplane 11.50 9.50 Total 59.04 36.77
Source: HSBC estimates
However, we believe costs might not be the only concern. We believe the adoption will still
hinge on the how foldable/rollable features are valued by consumers. We believe if Apple sees
that most iPhones buyers demand such features, OLED will likely be the choice of technology
since it is less likely micro-LED could be made foldable. We believe Samsung will retain its
foldable technology leadership for many years to come; however, the company will unlikely put
in any effort to make micro-LED foldable, given its vast investment in OLED.
17
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
Virtual reality? Augmented reality? Why not the best of both worlds?
Meanwhile, we believe that the significant potential of micro-LED could be realised in both VR
(virtual reality) and AR (augmented reality) applications. Micro-LED is the most feasible
technology for transparent displays, given the extremely high brightness and simple structure
that can be made invisible to the eye. In such cases, the image can make up a fraction (AR) or
the entire surface (VR) of a transparent screen. Meanwhile, current display technology can
either be VR (solid display blocking a person’s entire vision) or AR (projection onto a lens), but
not together. Keep in mind that the refresh rate of micro-LED is far above that of any display
technologies available today, so there is also the advantage of making images life-like.
Apple’s CEO Tim Cook discussed AR as an “extremely interesting and sort of a core
technology”, and the company is investing considerable resources in the area (CNBC, 10
August 2016). Apple now has over 200 employees working on AR projects.
Scalable to IT and TV products; oxide LCD is the key
As micro-LED offers the best picture quality, the technology can also be adopted in TV sets; in
fact, Sony demonstrated a 55” micro-LED TV (Sony calls it Crystal LED) in 2012, which drew
strong interest for its superb image quality. However, the product was never commercialised.
Sony did launch a line of commercial displays based on Crystal LED in May 2016. It is not clear
how the transfer of chips is performed, but we suspect the transfer works on a different principle
compared to that of Apple/LuxVue.
We believe for large-sized applications, the hurdle would be the TFT-backplane. The LTPS-
backplane is scalable only to certain sizes (we believe less than 20”), while the costs are high.
We believe oxide LCD, which is a rising backplane technology, will be feasible as scalability is
not an issue and costs, in theory, would be much lower than for LTPS.
Lastly, we also believe it is actually quite sensible to apply micro-LED displays in IT products, such
as notebooks (NB) and tablets, given the energy efficiency and durability (IT product life cycles are
longer than those of wearables and smartphones) are both critical for these devices. We believe
the transfer rate and capacity requirements will be considerably lower than for smartphones.
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
18
Tiny chips, profound impact
The basis for our belief that micro-LED will be developed into the display solution for the Apple
Watch are: 1) there is an intensifying effort of patent filings, which are becoming comprehensive
regarding all aspects of the production, and 2) the technology is highly suitable for wearables
and allows significant improvements in the function and performance of the Apple Watch, which
has not achieved as strong sales as other Apple products. However, as Apple so far has not
announced commercialisation targets and plans for micro-LED, we are not making any estimate
changes for any of the potential beneficiaries in our coverage universe. We do, however,
discuss the potential impact on the sector and companies below.
We like to point out that, while Apple closely guards any information on micro-LED, we believe
continuing patent filing efforts could be a sign that Apple is becoming more serious about
adopting the technology.
A highly positive long-term trend for panel makers
Although some investors may point out that as value-add from panel makers seems to play a less
significant role in micro-LED than in other display technologies, we believe it is a positive trend.
While micro-LED is a new generation of technology that offers many strong value-adds and
innovative uses, such as transparent displays, it requires limited additional capex for panel
makers.
In the future, we believe OLED and LCD will exist in the market place along with micro-LED,
as each of the three display technologies brings something to the table. We believe the
increasingly diverse product mix is good for panel makers, as the trend reduces effective
capacity and the likelihood of oversupply.
Keep in mind that the backplane is one of the most value-added processes of a display panel.
While shipping backplanes only (rather than the display module or cell) will lower the revenues
for panel makers, margins should rise with a growing share of micro-LED, in our view.
Many panel makers have been investing considerable resources, or have announced plans to
develop flexible OLED displays. At one point, the rising adoption of micro-LED may cause some
panel makers to review their investment strategies, especially if it is been proven that micro-LED
can be adopted beyond wearables. However, in the next 2-3 years, we don’t expect the robust
Tiny chips, profound impact
Micro-LED allows panel makers technology migration without high
capex; semiconductor and OSATs better utilising trailing edge capacity
Consumes vast LED capacity; plus strong potential for product mix
improvements; reduces much exposure to commoditised products
No estimate changes; potential boost to Epistar’s 2017-18e earnings
of 6% and 13%; positive share price driver for LGD, AUO and INX
19
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
investment in OLED to be influenced by micro-LED, which is likely perceived as Apple’s
proprietary technology.
A potential game-changer for the LED industry
The LED industry is still in the process of resolving the excessive supply situation, which is
rooted in: 1) robust expansion in 2011-14, and 2) the fast shrinking chip sizes, which means
much more chips can be produced from each MOCVD. We believe the LED industry’s recovery
needs to happen in two ways: 1) end-demand for the mass market, such as general lighting,
needs to continue growing at a rapid pace to digest excessive supply, and 2) we need to see
more new applications with a high entry barrier while consuming meaningful capacity (at least
5%). We have seen auto lighting, infrared and outdoor display emerging, but the share of these
high value-add applications needs to grow beyond the current 20-25% of the market.
We believe micro-LED could be an application that provides a significant boost to new
application growth. We expect at the initial stage micro-LED will consume 2% of global
upstream capacity on the Apple Watch adoption, but the ratio could increase to 10% on VR, AR
and IT adoption, and 25-30% when adopted in low-cost iPhones. If micro-LED is adopted in
mainstream iPhone models, the capacity required would be c50% of the existing capacity.
Note the margin difference between commoditised and high value-add LED chips is wide, with a
margin difference of 30-40ppt being common (despite costs to produce being not so different).
As such, the margin expansion potential is vast if the product mix shifts from commoditised to
high value-add, which can result in better bargaining power for LED makers and boost margins
for all products.
Semiconductors and OSATs
While we expect TSMC to play a key role in the micro-LED chip process, we believe the initial
impact is limited due to TSMC’s sheer size. Compared to its estimated 2017 revenue, the
annualised contribution from the Apple Watch is only 0.3-0.4%. However, if the technology is
eventually adopted in smartphones (say low-cost iPhones), the contribution could reach more
than 3%. As micro-LED utilises existing capacity, it is difficult to determine if the technology
could generate incremental sales for TSMC. However, we believe the emergence of a highly
advanced technology that utilises fabs that were previously thought as nearing the end of life
would be viewed as an incrementally positive to investors.
It is less clear when it comes to which OSAT will be providing the transfer process for micro-
LED to the TFT-backplane. Compared to the combined estimated 2017 revenue of Advanced
Semiconductor Engineering (ASE; 2311 TT, TWD39.05, Hold) and Siliconware Precision
Industries (SPIL; 2325 TT, TWD47.70, Hold), the micro-LED transfer process will likely account
for 1% of total sales.
Update on the LCD market
At the end of August, the blended LCD panel price was up 8% QTD (or a 7.5% rise for the QTD
average vs the 2Q16 average). This marks the biggest price hike in the history by far and
suggests 3Q16 panel price hikes will be above our current expectations, which was set in June
2016 (up to a 3% hike in 3Q16). Chinese TV set makers have been eager to procure panels,
due to concerns about supply shortages in 2017. This development has given panel makers
strong bargaining power. We expect stable pricing for the seasonally weak 4Q, which we
believe is achievable if the inventory can remain at mostly healthy levels.
EQUITIES ELECTRONIC EQUIPMENT
31 August 2016
20
However, we like to point out two key areas that deserve close attention: 1) supply of 32” TV
panels, as 32” panel prices have been strong (+18% QTD), can lure some panel makers to
increase supply in this size, which is seeing structural end-demand declines as consumers
worldwide are demanding larger TV sets, and 2) TV set makers’ margins, as we have seen
some Chinese TV set makers’ OPM drop to the low end of range in 2Q16 (even the cost of
panels in 2Q16 should have been low), likely due to fierce competition. If TV set makers start to
struggle financially, the panel procurement policy could turn conservative.
Epistar – the early beneficiary; a catalyst for LG Display; positive for AU Optronics and Innolux, as micro-LED eases concerns about a lack of OLED exposure
Epistar (2448 TT, TWD21.60, Buy, TP TWD30.00)
We believe Epistar should be one of the first beneficiaries from micro-LED, given the potential
revenue impact from the Apple Watch could be meaningful (partly due to the company’s smaller
revenue scale). Based on our estimates, if the Apple Watch’s volume remains 12-13m units
annually (we believe LED makers need to produce enough wafers for 15-20m displays to account
for the yield losses throughout the various stages of production), micro-LED could potentially make
up 3% of Epistar’s revenue, assuming the company captures 50% of the market share.
While the contribution may seem modest, we like to remind investors that: 1) this is assuming
no demand growth, even if the Apple Watch’s functionality could be improved from the superb
energy efficiency of micro-LED, 2) development efforts for other products could also contribute
meaningfully in 2018, and 3) margin could be high due to high entry barriers. We believe
product mix improvement holds the key to Epistar’s earnings recovery, given the margin gap
between the conventional and high value-add segments (micro-LED will likely be a high value-
add, high-margin product).
We maintain our estimates for Epistar in this report. However, we believe the adoption of micro-
LED in the Apple Watch could boost our 2017 EPS estimate by 6% and our 2018 EPS estimate
by 13%, as we believe the product mix change could lead to a 1ppt gain in margin (if we
assume a 30% GM, which is in line with company’s high value-add products). The potential
2018 earnings increase from a 1% GM increase seems sizable due to the high operating nature
of the LED chip making.
Blended large-sized LCD panel price index (July 2011 = 100)
This document has been issued by HSBC Securities (Taiwan) Corporation Limited in the conduct of its Taiwan regulated business for the information of its institutional and professional
customers. It is not intended for and should not be distributed to retail customers in Taiwan. This recommendation material is for reference only. Investors should carefully consider their own
investment risk. Investment results are the responsibility of the individual investor. HSBC Securities (Taiwan) Corporation Limited is regulated by the Securities and Futures Bureau. All enquires
by recipients in Taiwan must be directed to your HSBC contact in Taiwan. If this report is received by a customer of an affiliate of HSBC, it's provision to the recipient is subject to the terms of
business in place between the recipient and such affiliate. This document is not and should not be construed as an offer to sell or the solicitation of an offer to purchase or subscribe for any
investment. HSBC has based this document on information obtained from sources it believes to be reliable but which it has not independently verified; HSBC makes no guarantee,
representation or warranty and accepts no responsibility or liability as to its accuracy or completeness. Expressions of opinion are those of the Research Division of HSBC only and are subject to
change without notice. From time to time research analysts conduct site visits of covered issuers. HSBC policies prohibit research analysts from accepting payment or reimbursement for travel
expenses from the issuer for such visits. HSBC and its affiliates and/or their officers, directors and employees may have positions in any securities mentioned in this document (or in any related
investment) and may from time to time add to or dispose of any such securities (or investment). HSBC and its affiliates may act as market maker or have assumed an underwriting commitment
in the securities of companies discussed in this document (or in related investments), may sell them to or buy them from customers on a principal basis and may also perform or seek to perform
investment banking or underwriting services for or relating to those companies. HSBC Research may not be distributed to the public media or quoted or used by the public media without the
express written consent of HSBC Securities (Taiwan) Corporation Limited. Reports written by Taiwan-based analysts on non-Taiwan listed companies are not considered recommendations to
buy or sell securities under Taiwan Stock Exchange Operational Regulations governing securities firms recommending trades in securities to customers and as such HSBC Securities (Taiwan)
Corporation Limited may not execute transactions for clients in these securities/instruments. In the UK this report may only be distributed to persons of a kind described in Article 19(5) of the
Financial Services and Markets Act 2000 (Financial Promotion) Order 2005. The protections afforded by the UK regulatory regime are available only to those dealing with a representative of
HSBC Bank plc in the UK. HSBC Securities (USA) Inc. accepts responsibility for this research report prepared by its foreign affiliate. All U.S. persons receiving this report and wishing to effect
transactions in any security discussed herein should do so with HSBC Securities (USA) Inc. in the United States and not with the foreign affiliate, the issuer of this report. Note, however, that
HSBC Securities (USA) Inc. is not distributing this report and has not contributed to or participated in its preparation. In Korea, this publication is distributed by either The Hongkong and
Shanghai Banking Corporation Limited, Seoul Securities Branch ("HBAP SLS") or The Hongkong and Shanghai Banking Corporation Limited, Seoul Branch ("HBAP SEL") for the general
information of professional investors specified in Article 9 of the Financial Investment Services and Capital Markets Act (“FSCMA”). This publication is not a prospectus as defined in the FSCMA.
It may not be further distributed in whole or in part for any purpose. Both HBAP SLS and HBAP SEL are regulated by the Financial Services Commission and the Financial Supervisory Service
of Korea. In Singapore, this publication is distributed by The Hongkong and Shanghai Banking Corporation Limited, Singapore Branch for the general information of institutional investors or other
persons specified in Sections 274 and 304 of the Securities and Futures Act (Chapter 289) (“SFA”) and accredited investors and other persons in accordance with the conditions specified in
Sections 275 and 305 of the SFA. This publication is not a prospectus as defined in the SFA. It may not be further distributed in whole or in part for any purpose. The Hongkong and Shanghai
Banking Corporation Limited Singapore Branch is regulated by the Monetary Authority of Singapore. Recipients in Singapore should contact a "Hongkong and Shanghai Banking Corporation
Limited, Singapore Branch" representative in respect of any matters arising from, or in connection with this report. In Australia, this publication has been distributed by The Hongkong and
Shanghai Banking Corporation Limited (ABN 65 117 925 970, AFSL 301737) for the general information of its “wholesale” customers (as defined in the Corporations Act 2001). Where distributed
to retail customers, this research is distributed by HSBC Bank Australia Limited (AFSL No. 232595). These respective entities make no representations that the products or services mentioned in
this document are available to persons in Australia or are necessarily suitable for any particular person or appropriate in accordance with local law. No consideration has been given to the
particular investment objectives, financial situation or particular needs of any recipient. This publication is distributed in New Zealand by The Hongkong and Shanghai Banking Corporation
Limited, New Zealand Branch incorporated in Hong Kong SAR.
In Japan, this publication has been distributed by HSBC Securities (Japan) Limited. It may not be further distributed in whole or in part for any purpose.
In Canada, this document has been distributed by HSBC Bank Canada and/or its affiliates. Where this document contains market updates/overviews, or similar materials (collectively deemed
“Commentary” in Canada although other affiliate jurisdictions may term “Commentary” as either “macro-research” or “research”), the Commentary is not an offer to sell, or a solicitation of an offer
to sell or subscribe for, any financial product or instrument (including, without limitation, any currencies, securities, commodities or other financial instruments).
form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of HSBC Securities (Taiwan) Corporation Limited. MCI (P) 094/06/2016,
MCI (P) 085/06/2016 and MICA (P) 021/01/2016
[525402]
Global
Analyst, Global Sector Head Stephen Howard +44 20 7991 6820 [email protected]