asia display electronic equipment -...

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


[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

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


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


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


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.

LED vs OLED efficiency

Source: U.S. Department of Energy

0

75

150

225

300

2015 2016e 2017e 2018e 2019e 2020e 2021e 2022e 2023e 2024e 2025e 2025+

OLED LED

lm/Watt


31 August 2016

6

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


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


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


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


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.


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


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


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


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,


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


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


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


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

2016 2017 2018 2019 Pitch 15um 10um 7um 5um

Applications Wearables Wearables Wearables Wearables AR, VR AR, VR AR, VR NB, tablets NB, tablets Smartphones


Expand LED epi wafer size to 8”

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


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


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"


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


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


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


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.


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


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.


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)

Chinese TV set margin falling

Source: IHS Source: Company reports

40

55

70

85

100

Jul-11 Apr-12 Jan-13 Oct-13 Jul-14 Apr-15 Jan-16

12

15

18

21

24

3Q14 4Q14 1Q15 2Q15 3Q15 4Q15 1Q16 2Q16

TCL Multi. Hisense

OPM %

21


31 August 2016

Potential impact on 2017-18e estimates

_________ 2017e __________ __________ 2018e___________ Difference 2017e Difference 2018e (TWDm) HSBC

estimates With

micro-LED Consensus HSBC

estimates With micro-

LED Consensus vs HSBC

estimates vs

Consensus vs HSBC

estimates vs

Consensus

Rev 27,378 27,678 28,347 28,063 28,863 28,780 1.1% -2.4% 2.8% 0.3% GP 4,492 4,609 2,256 4,922 5,251 2,245 2.6% 104.3% 6.7% 133.9% OP 1,170 1,287 -431 1,567 1,850 -226 10.0% NA 18.1% NA N.I. 1,733 1,835 -461 1,771 2,001 -399 5.9% NA 13.0% NA EPS (TWD) 1.57 1.67 -0.54 1.61 1.81 -0.35 5.9% NA 13.0% NA GM (%) 16.4% 16.7% 8.0% 17.5% 18.2% 7.8% 24bp 869bp 100bp 1,074bp OPM (%) 4.3% 4.6% -1.5% 5.6% 6.4% -0.8% 38bp 617bp 115bp 752bp

Source: Bloomberg, HSBC estimates. NA – Not applicable.

We believe consensus holds a much more bearish outlook than our outlook for earnings (even

without the micro-LED assumption) due to the impact of the structural oversupply. We agree

oversupply is a situation that could last until 2H17; however, we like to highlight that Epistar has

been undergoing strong margin expansion (up 12ppt in 2Q16), due to stabilising prices of

commodity products and a rising mix of high value-add products, such as infrared, outdoor

display and auto lighting. The proliferation of micro-LED can potentially drive a significant mix

change at Epistar and can be a strong share/re-rating catalyst for the company.

Given the number of MOCVD tools required, if the proliferation of micro-LED spreads to high

volume products, such as low-cost iPhones, Epistar will most likely initiate another round of major

capacity expansion. We would like to point out that despite being in a prolonged downcycle since

late 2014, Epistar still manages to generate decent free cash flows. Its net gearing as of 2Q16

was healthy at 11%. We believe the company has the full capacity of making the 100 MOCVD

tool investment (likely around USD250m) required for low-cost iPhones.

Scenario analysis (assume Epistar has a 50% shares in wafers): impact to revenue and GP

Watch VR/AR NB Low-cost iPhone iPhone*

Pitch assumption 15um 10um 7um 5um 5um Annual device shipment 12m 1m? 15m 50m 120m* Resolution assumption 312x319 2560x1440 2560x1600 2560x1440 4k2k Rev as % of Epistar (vs 2018e) 3% 2% 16% 24% 100% Incremental GP (vs 2018e) 7% 5% 29% 43% 181% 6” wafer required per year 46,000 33,500 280,000 415,000 2,000,000 MOCVD required 20 14 80 115 490

Source: HSBC estimates. *Based on the assumption that OLED iPhone will be 50% of the market

Epistar quarterly earnings

(TWDm) 1Q16 2Q16 3Q16e 4Q16e 1Q17e 2Q17e 3Q17e 4Q17e 2015 2016e 2017e

Revenue 6,058 6,555 6,883 6,733 6,192 6,625 7,357 7,204 25,510 26,228 27,378 q-o-q 2% 8% 5% -2% -8% 7% 11% -2% y-o-y -9% -3% 11% 14% 2% 1% 7% 7% -8% 3% 4% GP (444) 292 1,063 1,298 774 981 1,321 1,416 123 2,209 4,492 OP (1,438) (432) 265 490 (41) 165 495 551 (3,511) (1,115) 1,170 N.I. (1,640) (1,750) 833 1,055 293 255 568 617 (3,019) (1,805) 1,733 EPS (TWD)

(1.49) (1.59) 0.76 0.96 0.27 0.23 0.52 0.56 (2.74) (1.64) 1.57

GM -7.3% 4.5% 15.5% 19.3% 12.5% 14.8% 18.0% 19.7% 0.5% 8.4% 16.4% OPM -23.7% -6.6% 3.9% 7.3% -0.7% 2.5% 6.7% 7.6% -13.8% -4.3% 4.3%

Source: Company data, HSBC estimates

LG Display (034220 KS, KRW30,500, Buy, TP KRW35,200)

The Apple Watch is only 1% of LG Display’s revenue. While that doesn’t appear to be a key

product, it is important for its OLED roadmap, particularly in small-sized displays, which deploy

a different technology from white OLED that LG Display is known for. If we assume Apple


31 August 2016

22

Watch sales remain flat, the migration to micro-LED could result in a lower revenue contribution

from the Apple Watch, as LG Display will likely be shipping backplanes only (which we estimate

are c40% vs OLED displays). However, due to the uncertainty of the event, we are not making

any estimate change.

We believe LG Display will likely be the only panel maker involved in an expected micro-LED

Apple Watch as the company has a proven track record and strong relations with Apple. We

believe micro-LED will be a positive catalyst for LG Display shares as it gives the impression

that the company is still very much involved in Apple’s advanced and proprietary technology.

We have seen Samsung being aggressive in expanding OLED capacity to serve growing

demands from internal customers and Chinese handset brands, as well as to prepare for the

major event of Apple migrating its iPhone to OLED displays, which will likely start in 2017,

according to DigiTimes. We expect LG Display will achieve meaningful volume of iPhone OLED

displays by 2018, but Samsung will likely remain as the major supplier as there is a

considerable gap in terms of capacity and flexible OLED technology.

Small-sized OLED capacity comparison: Samsung Display vs LG Display


Management of LG Display stated on 27 July 2016 that the “company will gradually increase its

OLED capacity in the coming years”. While it is unknown that such comment could be indicating

that management is starting to take a view that Apple’s display technology roadmap could

encompass OLED, LCD and micro-LED, in the long run conservative expansion could be a

prudent strategy for LG Display at this point, in our view.

Meanwhile, for the LCD panel market situation in 2H16, we continue to see panel prices rising

at an unprecedented pace as TV set makers are stocking up on panels amidst potential supply

tightness in 2017. The two sizes that are seeing the greatest price hikes are 32” and 43”, which

saw 18% and 10% hikes QTD, respectively. We believe these trends will help LG Display to

meet the positive expectations for 3Q16 margin (we expect 5ppt expansion q-o-q, while

consensus expects 4.5ppt), as these sizes are making up almost 40% of LG Display’s TV

revenue (or 16-18% of overall revenue).

0

45

90

135

180

1Q15 2Q15 3Q15 4Q15 1Q16 2Q16 3Q16e 4Q16e 1Q17e 2Q17e 3Q17e 4Q17e 1Q18e 2Q18e 3Q18e 4Q18e

Samsung LGD

mn 5.2" equivalent per quarter

23


31 August 2016

LG Display: HSBC estimates vs. consensus

__________ 2016e ___________ __________ 2017e ___________ Difference 2016e Difference 2017e (KRWbn) HSBC estimates Consensus HSBC estimates Consensus vs. Consensus vs. Consensus

Sales 24,519 25,390 24,806 25,927 -3.4% -4.3% Gross profit 3,428 3,140 4,323 3,750 9.1% 15.3% Operating profit 1,041 850 1,846 1,329 22.4% 38.9% EBITDA 4,223 4,053 5,185 4,680 4.2% 10.8% Net profit 718 484 1,483 876 48.4% 69.2% FD EPS (KRW) 1,950 1,325 4,025 2,446 47.2% 64.6% Margin (%) Gross margin 14.0% 12.4% 17.4% 14.5% 162bps 297bps Operating margin 4.2% 3.3% 7.4% 5.1% 90bps 232bps EDITDA margin 17.2% 16.0% 20.9% 18.1% 126bps 285bps

Source: Bloomberg, HSBC estimates

LG Display: Quarterly earnings

(KRWbn) 1Q16 2Q16 3Q16e 4Q16e 1Q17e 2Q17e 3Q17e 4Q17e 2015e 2016e 2017e

Revenue 5,989 5,855 6,373 6,302 5,572 5,969 6,702 6,562 28,384 24,519 24,806 q-o-q -20% -2% 9% -1% -12% 7% 12% -2% y-o-y -15% -13% -11% -16% -7% 2% 5% 4% 7% -14% 1% GP 626 610 1,054 1,137 942 1,074 1,291 1,016 4,314 3,428 4,323 OP 40 44 433 524 352 477 649 369 1,626 1,041 1,846 EBITDA 853 832 1,187 1,352 1,173 1,328 1,462 1,222 5,002 4,223 5,185 N.I. 1 (84) 363 438 280 383 526 294 1,048 718 1,483 FD EPS (KRW)

3 (228) 987 1,188 760 1,040 1,428 798 2,845 1,950 4,025

GM 10.5% 10.4% 16.5% 18.0% 16.9% 18.0% 19.3% 15.5% 15.2% 14.0% 17.4% OPM 0.7% 0.8% 6.8% 8.3% 6.3% 8.0% 9.7% 5.6% 5.7% 4.2% 7.4%


Taiwan panel makers – AU Optronics (2409 TT, TWD12.85, Buy, TP TWD14.50) and

Innolux (3481 TT, TWD11.65, Buy, TP TWD14.30)

Some investors may expect Taiwan panel makers, such as AUO, to become suppliers for Apple’s

micro-LED displays as: 1) AUO had a stake in LuxVue before selling all its shares in April 2014,

2) IHS has indicated that some of the R&D is done in Taiwan, and 3) the supply chain could

involve Taiwanese companies, such as Epistar and TSMC. However, we believe the TFT-

backplane for the first expected micro-LED product, the Apple Watch, will be supplied by LG

Display.

We believe the key is still in the backplane technology. The micro-LED backplane likely requires

current-drive technologies, such as LTPS or oxide. However, Taiwan panel makers have not

been the leaders in these technologies, as most resources have been allocated to conventional

a-si technology. With that said, as AUO and INX will each ramp up an advanced Gen 6 LTPS

fab in late 2016 and early 2017, we believe both companies will be gaining experience in

backplane technology.

We don’t see the need to adjust our estimates for AUO and INX on the back of this emerging

technology. However, we do view micro-LED as a positive sentiment driver for both companies

as: 1) AUO and INX have long been scrutinised for their lack of OLED investment and strategy,

and we believe the emergence of another technology choice could alleviate these concerns,

and 2) in terms of display technical barriers and financial commitment, micro-LED is lower than

those for OLED, which puts Taiwan panel makers in a better position to compete. While most

investors may focus on potential iPhone opportunities, other products, such as NB and AR

displays could be more feasible for the relatively under-resourced Taiwan panel makers.


31 August 2016

24

AU Optronics: HSBC estimates vs. consensus

__________ 2016e ___________ __________ 2017e ___________ Difference 2016e Difference 2017e (TWDm) HSBC estimates Consensus HSBC estimates Consensus vs. Consensus vs. Consensus

Sales 321,373 320,846 339,754 321,841 0.2% 5.6% Gross profit 29,480 23,242 46,250 28,335 26.8% 63.2% Operating profit 7,742 1,902 22,851 8,513 307.0% 168.4% EBITDA 48,224 43,466 64,142 46,213 10.9% 38.8% Net profit 5,440 516 18,325 6,735 NA 172.1% FD EPS (TWD) 0.59 -0.00 1.90 0.72 NA 163.7% Margin (%) Gross margin 9.2% 7.2% 13.6% 8.8% 193bp 481bp Operating margin 2.4% 0.6% 6.7% 2.6% 182bp 408bp EBITDA margin 15.0% 13.5% 18.9% 14.4% 146bp 452bp


AU Optronics: Quarterly earnings


Revenue 71,135 80,091 84,654 85,493 81,359 81,050 88,980 88,365 360,346 321,373 339,754 q-o-q -15% 13% 6% 1% -5% 0% 10% -1% y-o-y -25% -13% -6% 2% 14% 1% 5% 3% -12% -11% 6% GP 390 5,564 10,446 13,081 9,399 10,882 14,017 11,952 39,837 29,480 46,250 OP -5,097 116 5,028 7,695 3,704 5,371 8,589 6,474 17,521 7,742 22,851 N.I. -5,477 -572 4,545 6,944 2,900 4,332 7,099 5,280 4,932 5,440 18,325 FD EPS (TWD)

-0.57 -0.06 0.47 0.72 0.30 0.45 0.74 0.55 0.51 0.59 1.90

Margin (%) GM 0.5% 6.9% 12.3% 15.3% 11.6% 13.4% 15.8% 13.5% 11.1% 9.2% 13.6% OPM -7.2% 0.1% 5.9% 9.0% 4.6% 6.6% 9.7% 7.3% 4.9% 2.4% 6.7%

Source: Company report, HSBC estimates

Innolux: HSBC estimates vs. consensus

__________ 2016e ___________ __________ 2017e ___________ Difference 2016e Difference 2017e (TWDm) HSBC estimates Consensus HSBC estimates Consensus vs. Consensus vs. Consensus

Sales 272,828 266,461 289,138 275,866 2.4% 4.8% Gross profit 15,444 8,831 39,440 18,897 74.9% 108.7% Operating profit -2,934 -10,522 20,258 975 72.1% NA Net profit -3,624 -9,895 15,609 -1,057 63.4% NA EPS (TWD) -0.36 -0.58 1.57 0.07 61.7% NA Margin (%) GM 5.7% 3.3% 13.6% 6.9% 235bp 679bp OPM -1.1% -3.9% 7.0% 0.4% 287bp 665bp Net margin (%) -1.3% -3.7% 5.4% -0.4% 239bp 578bp


Innolux: Quarterly earnings


Revenue 56,417 66,805 71,880 77,726 70,489 70,606 73,110 74,934 364,133 272,828 289,138 q-o-q -31% 18% 8% 8% -9% 0% 4% 2% y-o-y -44% -29% -19% -5% 25% 6% 2% -4% -15% -25% 6% GP -4,037 1,554 6,746 11,181 8,307 9,027 11,102 11,004 46,640 15,444 39,440 OP -8,402 -3,039 2,146 6,362 3,688 4,286 6,255 6,030 22,431 -2,934 20,258 EBITDA 3,046 7,408 11,778 14,479 11,188 10,786 13,155 12,810 76,002 36,408 47,938 N.I. -8,582 -3,476 2,179 6,255 3,685 4,220 3,984 3,720 12,607 -3,624 15,609 FD EPS (TWD)

-0.86 -0.35 0.22 0.63 0.37 0.42 0.40 0.37 1.27 -0.36 1.57

Margin (%) GM -7.2% 2.3% 9.4% 14.4% 11.8% 12.8% 15.2% 14.7% 12.8% 5.7% 13.6% OPM -14.9% -4.5% 3.0% 8.2% 5.2% 6.1% 8.6% 8.0% 6.2% -1.1% 7.0%

Source: Company report, HSBC estimates

25


31 August 2016

Valuation and risks

Epistar (2448 TT) maintain Buy and TP of TWD30.00

We see fair value at a 0.6x (unchanged) on an average of YE16e and YE17e (unchanged)

BVPS of TWD49.51, which gives us a fair value target price of TWD30. We make no estimate

changes in this report. Our target multiple of 0.6x is the mid-cycle multiple during the current

cycle (that began in September 2014). We believe 0.6x also can be justified by our expectations

for the ROE in the coming 12 months (around 6%). Our target price of TWD30 represents our

assessment of the stock’s current value and implies 38.9% upside from the current share price;

we reiterate our Buy rating.

Key downside risks: severe competition from Chinese LED makers, and a worse-than-

expected product mix.

LG Display (034220 KS) maintain Buy and TP of KRW35,200


BVPS of KRW39,111, which gives us a fair value target price of KRW35,200. We male no

estimate changes in this report. Our target multiple of 0.9x is based on the mid-cycle trading

range in 2H14, as we expect LG Display’s OPM to recover to similar levels (8-9%) in the coming

quarters. We believe 0.9x also can be justified by our expectations for the ROE in the coming

months (around 12%). Our target price of KRW35,200 represents our assessment of the stock’s

current value and implies 15.4% upside from the current share price. We reiterate our Buy rating

as we are positive on the company’s long-term outlook.

Key downside risks: a slowdown in size migration, aggressive investment in OLED, which has

yet to be commercially proven for large-sized applications, and high volatility in Apple’s product

cycles.

AU Optronics (2409 TT) maintain Buy and TP of TWD14.50


BVPS of TWD21.77, which gives us a fair value target price of TWD14.50. We make no

estimate changes in this report. Our target multiple of 0.65x is based on the mid-cycle trading

range in 2H14, as we expect AUO’s OPM to recover to similar levels (8-9%) in the coming

quarters. We believe 0.65x is also justified by our expectations for the ROE in the coming

months (around 12%). Our target price of TWD14.50 represents our assessment of the stock’s

current value and implies 12.8% upside from the current share price. With 12.8% of upside, we

rate AUO as a Buy as the company is a leader in ultra large-sized TV panels, while its product

mix is highly exposed to large-sized LCD, where we see supply and demand remaining highly

favourable in the coming year.

Key downside risk: a lack of sales promotions on large-sized TV, which could slow size

migration, and AUO’s ramp-up of the LTPS (low-temperature polycrystalline silicon) fab could

add more uncertainty as the small-sized LCD market segment remains competitive.

Innolux (3481 TT) maintain Buy and TP of TWD14.30

We see fair value at 0.6x (unchanged) YE16/17e BVPS of TWD23.75 (unchanged), which gives us

a fair value target price of TWD14.30. The 0.6x PB multiple is based on the mid-point of its trading

range in 2014-15, when its ROE was similar to what we expect INX to generate in the next four

quarters (7-8%). We make no estimate changes in this report. Our target price of TWD14.30

represents our assessment of the stock’s current value and implies 22.7% upside from the

current share price; we reiterate our Buy rating.

Key downside risks: a slower-than-expected size migration and slow LCD TV demand in

China, where INX has a high market share.


31 August 2016

26

TPK (3673 TT) maintain Buy and TP of TWD86.00

We see fair value at a 10x (unchanged) PE on our 2017e EPS estimate of TWD8.60

(unchanged), which gives us a fair value target price of TWD86. The target PE multiple of 10x is

more or less in line with the average consensus PE multiple and also in line with other Apple

plays. We make no estimate changes in this report. Our fair value target price of TWD86

represents our assessment of the stock’s current value and implies 42.6% upside from the

current share price; we reiterate our Buy rating.

Key downside risks: more severe-than-expected competition from peers, such as GIS (6456 TT,

Not Rated), weak demand for the Apple Watch and iPhone, and an earlier-than-expected adoption

of fingerprint touch displays; technology risks from the Apple Watch adopting a different touch

solution, such as a touch embedded micro-LED display.

TSMC (2330 TT) maintain Buy and TP of TWD191.00

Our fair value target price of TWD191 is based on a 14x PE multiple applied to our forward

(2H16/1H17e) EPS estimate of TWD13.61. We note the 14x PE multiple is still within the

historical 10-year 10-15x PE range and also translates into a forward 3.1x PB, which is in the

middle of the historical 2.5-3.5x PB range (and don’t forget that TSMC has more fully

depreciated assets still being utilised, so we estimate its “adjusted” book value is likely 15%

higher – and, as such, its PB is similarly lower). Given TSMC’s consistent earnings growth,

20%-plus ROE, bellwether status/scarcity value in Asia’s technology sector and rising dividends,

we think our multiple is justified. Given the above-mentioned factors we rate TSMC a Buy even

though our target price implies only 7.9% upside from the current share price.

Key downside risks: currency headwinds, and post build, end-demand matter most. A weak

global economy may ultimately pressure end-demand in 2H16. We have already seen strong

supply chain growth rates in large-sized LCD drivers and guidance for a strong wireless

communication ramp in 2Q16. We can’t help but worry that excess inventories could be a risk

by the October earnings season. The TWD has depreciated 3-4% over the past 90 days, may

pressure TSMC, given most of its sales are in USD. Lastly, any excessive delay to 10/7nmnm,

higher depreciation or capital intensity, entry into new markets, and/or increased

insourcing/competition are additional risks.

27


31 August 2016

Financial statements

Year to 12/2015a 12/2016e 12/2017e 12/2018e

Profit & loss summary (KRWbn)

Revenue 28,384 24,519 24,806 27,038

EBITDA 5,002 4,223 5,185 5,038

Depreciation & amortisation -3,376 -3,183 -3,338 -3,338

Operating profit/EBIT 1,626 1,041 1,846 1,699

Net interest -128 -135 -155 -175

PBT 1,434 981 1,786 1,659

HSBC PBT 1,434 981 1,786 1,659

Taxation -386 -262 -304 -282

Net profit 1,048 718 1,483 1,377

HSBC net profit 1,048 718 1,483 1,377

Cash flow summary (KRWbn)

Cash flow from operations 2,768 3,144 4,058 3,946

Capex -2,588 -4,250 -4,750 -5,000

Cash flow from investment -2,486 -4,158 -4,667 -4,925

Dividends -179 -179 -184 -184

Change in net debt -131 1,541 609 979

FCF equity 1,021 -581 -187 -589

Balance sheet summary (KRWbn)

Intangible fixed assets 0 0 0 0

Tangible fixed assets 10,546 11,613 13,025 14,686

Current assets 9,532 9,399 10,617 10,994

Cash & others 2,524 1,983 2,874 2,895

Total assets 22,577 21,420 24,070 26,129

Operating liabilities 5,648 5,962 6,327 6,756

Gross debt 4,224 5,224 6,724 7,724

Net debt 1,700 3,241 3,850 4,829

Shareholders' funds 12,705 13,244 14,727 16,104

Invested capital 11,906 13,068 14,441 16,029

Ratio, growth and per share analysis

Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue 7.3 -13.6 1.2 9.0

EBITDA 3.1 -15.6 22.8 -2.8

Operating profit 19.8 -36.0 77.4 -8.0

PBT 15.4 -31.6 82.2 -7.1

HSBC EPS 14.2 -31.5 106.4 -7.1

Ratios (%)

Revenue/IC (x) 2.4 2.0 1.8 1.8

ROIC 11.5 6.1 11.1 9.3

ROE 8.7 5.5 10.6 8.9

ROA 5.0 3.7 7.1 6.1

EBITDA margin 17.6 17.2 20.9 18.6

Operating profit margin 5.7 4.2 7.4 6.3

EBITDA/net interest (x) 39.2 31.3 33.5 28.8

Net debt/equity 13.4 24.5 26.1 30.0

Net debt/EBITDA (x) 0.3 0.8 0.7 1.0

CF from operations/net debt 162.8 97.0 105.4 81.7

Per share data (KRW)

EPS Rep (diluted) 2,844.59 1,949.87 4,025.33 3,738.58

HSBC EPS (diluted) 2,844.59 1,949.87 4,025.33 3,738.58

DPS 514.72 514.72 514.72 514.72

Book value 35,507.18 37,014.44 41,158.23 45,006.84

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 0.8 1.0 1.0 1.0

EV/EBITDA 4.6 5.8 4.9 5.2

EV/IC 1.9 1.9 1.7 1.6

PE* 10.7 15.6 7.6 8.2

PB 0.9 0.8 0.7 0.7

FCF yield (%) 4.8 -2.7 -0.9 -2.8

Dividend yield (%) 1.7 1.7 1.7 1.7

* Based on HSBC EPS (diluted)

Issuer information

Share price (KRW) 30,500.00 Free float 45%

Target price (KRW) 35,200.00 Sector Electronic Equipment

Reuters (Equity) 034220.KS Country Korea

Bloomberg (Equity) 034220 KS Analyst Jerry Tsai

Market cap (USDm) 19,594 Contact +886 2 6631 2863

Price relative

Source: HSBC Note: Priced at close of 26 Aug 2016

18000.00

23000.00

28000.00

33000.00

38000.00

18000.00

23000.00

28000.00

33000.00

38000.00

2014 2015 2016 2017

LG Display Rel to KOSPI INDEX

Financials & valuation: LG Display Buy


31 August 2016

28


Year to 12/2015a 12/2016e 12/2017e 12/2018e

Profit & loss summary (TWDm)

Revenue 25,510 26,228 27,378 28,063

EBITDA 2,166 4,282 6,836 7,517


Operating profit/EBIT -3,515 -1,115 1,170 1,567

Net interest -1,109 -920 -920 -920

PBT -3,625 -1,386 1,969 2,012

HSBC PBT -3,625 -1,386 1,969 2,012

Taxation 303 -286 -236 -241

Net profit -3,019 -1,805 1,733 1,771

HSBC net profit -3,019 -1,805 1,733 1,771

Cash flow summary (TWDm)


Capex -3,925 -3,000 -3,000 -3,000


Dividends -915 0 -866 -866

Change in net debt 712 -1,631 -2,338 -1,561

FCF equity 1,749 1,023 790 2,160

Balance sheet summary (TWDm)

Intangible fixed assets 7,995 7,835 7,678 7,524


Current assets 30,886 31,805 36,009 39,105

Cash & others 8,430 10,061 12,399 13,960

Total assets 83,132 81,393 83,097 83,838


Gross debt 16,696 16,696 16,696 16,696

Net debt 8,266 6,635 4,297 2,736




Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue -8.0 2.8 4.4 2.5

EBITDA -67.7 97.7 59.7 10.0

Operating profit -246.9 - - 34.0

PBT -264.8 - - 2.2

HSBC EPS -266.5 - - 2.2

Ratios (%)

Revenue/IC (x) 0.4 0.5 0.5 0.5

ROIC -5.6 -2.5 1.9 2.5

ROE -5.2 -3.3 3.2 3.2

ROA -2.6 -0.6 3.2 3.2

EBITDA margin 8.5 16.3 25.0 26.8

Operating profit margin -13.8 -4.3 4.3 5.6


Net debt/equity 14.9 12.3 7.7 4.8

Net debt/EBITDA (x) 3.8 1.5 0.6 0.4


Per share data (TWD)

EPS Rep (diluted) -2.74 -1.64 1.57 1.61

HSBC EPS (diluted) -2.74 -1.64 1.57 1.61

DPS 0.00 0.00 0.79 0.80

Book value 50.51 48.87 50.45 51.27

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 1.0 1.0 0.8 0.8

EV/EBITDA 12.1 5.9 3.4 2.9

EV/IC 0.5 0.5 0.4 0.4

PE* NM NM 13.7 13.4

PB 0.4 0.4 0.4 0.4

FCF yield (%) 9.8 5.5 4.2 11.1

Dividend yield (%) 0.0 0.0 3.6 3.7

* Based on HSBC EPS (diluted). NM – Not meaningful.

Issuer information

Share price (TWD) 21.60 Free float 70%

Target price (TWD) 30.00 Sector Semiconductors

Reuters (Equity) 2448.TW Country Taiwan

Bloomberg (Equity) 2448 TT Analyst Jerry Tsai

Market cap (USDm) 744 Contact +886 2 6631 2863

Price relative


9.40

19.40

29.40

39.40

49.40

59.40

69.40

79.40

9.40

19.40

29.40

39.40

49.40

59.40

69.40

79.40

2014 2015 2016 2017

Epistar Corp Rel to TAIWAN WEIGHTED INDEX

Financials & valuation: Epistar Corp Buy

29


31 August 2016


Year to 12/2015a 12/2016e 12/2017e 12/2018e


Revenue 360,346 321,373 339,754 336,356

EBITDA 64,373 48,223 64,142 65,709



Net interest -4,221 -1,268 -940 -950

PBT 7,599 7,743 21,611 22,229

HSBC PBT 7,599 7,743 21,611 22,229

Taxation -2,756 -2,312 -3,206 -3,304

Net profit 4,932 5,725 18,325 18,844

HSBC net profit 4,932 5,725 18,325 18,844



Capex -33,440 -46,000 -34,800 -31,800


Dividends -4,812 -3,368 -1,632 -5,498

Change in net debt -30,445 -582 -22,253 -26,199

FCF equity 39,922 467 19,912 26,872


Intangible fixed assets 3,295 3,229 3,164 3,101


Current assets 161,992 136,391 132,083 141,868

Cash & others 79,095 54,677 46,930 53,129

Total assets 425,554 449,122 439,625 440,031


Gross debt 109,239 84,239 54,239 34,239

Net debt 30,144 29,562 7,309 -18,891




Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue -11.7 -10.8 5.7 -1.0

EBITDA -17.7 -25.1 33.0 2.4

Operating profit -21.0 -55.8 195.2 1.4

PBT -62.0 1.9 179.1 2.9

HSBC EPS -72.0 16.1 220.1 2.8

Ratios (%)

Revenue/IC (x) 1.8 1.5 1.5 1.5

ROIC 5.5 2.6 8.4 8.8

ROE 2.4 2.8 8.5 8.2

ROA 1.7 1.5 4.4 4.5

EBITDA margin 17.9 15.0 18.9 19.5



Net debt/equity 14.7 14.3 3.3 -8.0

Net debt/EBITDA (x) 0.5 0.6 0.1 -0.3

CF from operations/net debt 266.5 169.3 805.3 -


EPS Rep (diluted) 0.51 0.59 1.90 1.96

HSBC EPS (diluted) 0.51 0.59 1.90 1.96

DPS 0.35 0.17 0.57 0.50

Book value 21.26 21.48 23.21 24.58

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 0.4 0.4 0.3 0.3

EV/EBITDA 2.2 2.9 1.8 1.3

EV/IC 0.7 0.6 0.5 0.4

PE* 25.1 21.6 6.7 6.6

PB 0.6 0.6 0.6 0.5

FCF yield (%) 36.7 0.4 18.6 25.3

Dividend yield (%) 2.7 1.3 4.4 3.9


Issuer information


Target price (TWD) 14.50 Sector Electronic Equipment




Price relative


6.70

8.70

10.70

12.70

14.70

16.70

18.70

20.70

6.70

8.70

10.70

12.70

14.70

16.70

18.70

20.70

2014 2015 2016 2017

AU Optronics Rel to TAIWAN WEIGHTED INDEX

Financials & valuation: AU Optronics Buy


31 August 2016

30


Year to 12/2015a 12/2016e 12/2017e 12/2018e


Revenue 364,133 272,828 289,138 287,403

EBITDA 76,002 36,409 47,938 60,412


Operating profit/EBIT 22,431 -2,933 20,258 22,732

Net interest -2,091 -1,500 -2,000 -1,800

PBT 16,652 -2,432 18,613 18,932

HSBC PBT 16,652 -2,432 18,613 18,932

Taxation -4,045 -1,191 -3,004 -2,272

Net profit 12,607 -3,623 15,609 16,660

HSBC net profit 12,607 -3,623 15,609 16,660



Capex -24,511 -34,438 -35,000 -30,000


Dividends -6,947 0 -9,955 -9,955

Change in net debt -52,593 -1,063 -8,075 -24,130

FCF equity 45,354 -720 7,934 26,340




Current assets 138,867 121,750 131,642 157,673

Cash & others 52,643 33,706 41,781 65,910

Total assets 387,442 387,438 408,866 431,643


Gross debt 59,991 39,991 39,991 39,991

Net debt 7,348 6,286 -1,790 -25,919




Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue -15.1 -25.1 6.0 -0.6

EBITDA -14.7 -52.1 31.7 26.0

Operating profit -20.4 -113.1 - 12.2

PBT -26.1 -114.6 - 1.7

HSBC EPS -41.8 -128.7 - 6.7

Ratios (%)

Revenue/IC (x) 1.7 1.4 1.4 1.4

ROIC 8.5 -2.2 8.2 9.8

ROE 5.5 -1.6 6.6 6.6

ROA 3.3 -0.4 4.3 4.3

EBITDA margin 20.9 13.3 16.6 21.0

Operating profit margin 6.2 -1.1 7.0 7.9


Net debt/equity 3.2 2.7 -0.7 -9.9

Net debt/EBITDA (x) 0.1 0.2 0.0 -0.4

CF from operations/net debt 1034.3 579.2 - -

Per share data (CUR)

EPS Rep (diluted) 1.27 -0.36 1.57 1.67

HSBC EPS (diluted) 1.27 -0.36 1.57 1.67

DPS 0.70 0.00 1.00 1.00

Book value 23.33 22.97 24.53 26.21

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 0.3 0.4 0.4 0.3

EV/EBITDA 1.5 3.1 2.2 1.3

EV/IC 0.6 0.6 0.5 0.4

PE* 9.2 NM 7.4 7.0

PB 0.5 0.5 0.5 0.4

FCF yield (%) 42.7 -0.7 7.5 25.2

Dividend yield (%) 6.0 0.0 8.6 8.6


Issuer information


Target price (TWD) 14.30 Sector Computers & Peripherals




Price relative


7.70

9.70

11.70

13.70

15.70

17.70

19.70

7.70

9.70

11.70

13.70

15.70

17.70

19.70

2014 2015 2016 2017

Innolux Corporation Rel to TAIWAN WEIGHTED INDEX

Financials & valuation: Innolux Corporation Buy

31


31 August 2016


Year to 12/2015a 12/2016e 12/2017e 12/2018e


Revenue 121,364 99,782 105,346 112,720

EBITDA -8,652 5,807 10,399 12,345


Operating profit/EBIT -18,354 -995 3,796 5,741

Net interest -358 -774 -700 -400

PBT -21,019 -787 3,596 4,741

HSBC PBT -21,019 -787 3,596 4,741

Taxation 943 -140 -719 -1,185

Net profit -20,007 -780 3,024 3,556

HSBC net profit -20,007 -780 3,024 3,556



Capex -5,852 -3,200 -3,300 -3,299


Dividends -181 0 0 0

Change in net debt -637 -7,789 -8,223 -9,290

FCF equity -22,600 5,313 6,318 7,869




Current assets 61,034 60,667 65,353 71,685

Cash & others 16,393 19,182 22,405 26,695

Total assets 115,228 127,745 130,074 134,186


Gross debt 59,419 54,419 49,419 44,419

Net debt 43,026 35,237 27,015 17,724




Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue -6.3 -17.8 5.6 7.0

EBITDA -205.5 - 79.1 18.7

Operating profit -12768.2 - - 51.2

PBT -3752.9 - - 31.8

HSBC EPS -7672.0 - - 17.6

Ratios (%)

Revenue/IC (x) 1.7 1.6 1.9 2.3

ROIC -24.1 -1.8 5.4 8.7

ROE -51.5 -2.5 9.4 10.0

ROA -14.1 0.5 3.2 3.3

EBITDA margin -7.1 5.8 9.9 11.0

Operating profit margin -15.1 -1.0 3.6 5.1

EBITDA/net interest (x) - 7.5 14.9 30.9

Net debt/equity 136.5 114.6 80.0 47.5

Net debt/EBITDA (x) -5.0 6.1 2.6 1.4



EPS Rep (diluted) -57.82 -2.22 8.60 10.11

HSBC EPS (diluted) -57.82 -2.22 8.60 10.11

DPS 0.00 0.00 0.00 0.00

Book value 91.11 87.44 96.04 106.15

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 0.5 0.6 0.4 0.3

EV/EBITDA - 9.5 4.5 3.0

EV/IC 0.9 0.9 0.9 0.8

PE* NM NM 6.9 5.8

PB 0.6 0.7 0.6 0.6

FCF yield (%) -114.5 27.0 32.1 40.1

Dividend yield (%) 0.0 0.0 0.0 0.0


Issuer information


Target price (TWD) 86.00 Sector Electronic Equipment



Market cap (USDm) 645 Contact +8862 6631 2863

Price relative


24.00

74.00

124.00

174.00

224.00

274.00

324.00

24.00

74.00

124.00

174.00

224.00

274.00

324.00

2014 2015 2016 2017

TPK Holding Rel to TAIWAN WEIGHTED INDEX

Financials & valuation: TPK Holding Buy


31 August 2016

32


Year to 12/2015a 12/2016e 12/2017e 12/2018e


Revenue 843,497 927,620 1,014,239 1,054,809

EBITDA 546,314 607,867 676,481 696,478



Net interest 30,381 7,390 7,100 3,164

PBT 354,189 374,908 420,413 422,352

Taxation -43,873 -51,359 -58,964 -59,129

Net profit 310,326 323,602 361,489 363,264

HSBC net profit 310,326 323,602 361,489 363,264



Capex -257,517 -336,727 -350,000 -369,183


Dividends -116,686 -155,580 -181,510 -208,062

Change in net debt -144,115 -13,973 -87,374 -60,758

FCF equity 277,324 197,283 272,903 271,330


Tangible fixed assets 875,780 996,669 1,088,894 1,180,786

Current assets 746,744 778,104 855,694 709,336

Cash & others 586,163 581,205 656,578 709,336

Total assets 1,657,518 1,832,528 2,012,154 1,958,364


Gross debt 254,957 236,026 224,026 216,026

Net debt -331,206 -345,179 -432,552 -493,310

Shareholders' funds 1,222,634 1,387,546 1,579,156 1,579,156

Invested capital 856,434 984,612 1,079,038 971,815


Year to 12/2015a 12/2016e 12/2017e 12/2018e

Y-o-y % change

Revenue 10.6 10.0 9.3 4.0

EBITDA 9.7 11.3 11.3 3.0

Operating profit 8.7 13.5 12.5 1.4

PBT 16.5 5.8 12.1 0.5

HSBC EPS 16.7 4.3 11.7 0.2

Ratios (%)

Revenue/IC (x) 1.0 1.0 1.0 1.0

ROIC 33.7 34.5 34.4 35.2

ROE 27.4 24.8 24.4 23.0

ROA 18.0 18.2 18.5 18.2

EBITDA margin 64.8 65.5 66.7 66.0


Net debt/equity -27.1 -24.9 -27.4 -31.2

Net debt/EBITDA (x) -0.6 -0.6 -0.6 -0.7


EPS Rep (diluted) 11.97 12.48 13.94 13.97

HSBC EPS (diluted) 11.97 12.48 13.94 13.97

DPS 4.50 6.00 7.00 8.00

Book value 47.15 53.51 60.90 60.72

Valuation data

Year to 12/2015a 12/2016e 12/2017e 12/2018e

EV/sales 5.0 4.5 4.0 3.8

EV/EBITDA 7.7 6.9 6.0 5.8

EV/IC 4.9 4.2 3.8 4.1

PE* 14.7 14.1 12.7 12.6

PB 3.7 3.3 2.9 2.9

FCF yield (%) 6.1 4.4 6.1 6.0

Dividend yield (%) 2.5 3.4 4.0 4.5


Issuer information


Target price (TWD) 191.00 Sector Semiconductors


Bloomberg (Equity) 2330 TT Analyst Steven Pelayo

Market cap (USDm) 143,991 Contact +852 2822 4391

Price relative


90.00

110.00

130.00

150.00

170.00

190.00

90.00

110.00

130.00

150.00

170.00

190.00

2014 2015 2016 2017

Taiwan Semiconductor Rel to TAIWAN WEIGHTED INDEX

Financials & valuation: Taiwan Semiconductor Buy

33


31 August 2016

Disclosure appendix

Analyst Certification

The following analyst(s), economist(s), and/or strategist(s) who is(are) primarily responsible for this report, certifies(y) that the

opinion(s) on the subject security(ies) or issuer(s) and/or any other views or forecasts expressed herein accurately reflect their

personal view(s) and that no part of their compensation was, is or will be directly or indirectly related to the specific

recommendation(s) or views contained in this research report: Jerry Tsai, Ricky Seo, Steven Pelayo and David Huang

Important disclosures

Equities: Stock ratings and basis for financial analysis

HSBC believes an investor's decision to buy or sell a stock should depend on individual circumstances such as the investor's

existing holdings, risk tolerance and other considerations and that investors utilise various disciplines and investment horizons

when making investment decisions. Ratings should not be used or relied on in isolation as investment advice. Different

securities firms use a variety of ratings terms as well as different rating systems to describe their recommendations and

therefore investors should carefully read the definitions of the ratings used in each research report. Further, investors should

carefully read the entire research report and not infer its contents from the rating because research reports contain more

complete information concerning the analysts' views and the basis for the rating.

From 23rd March 2015 HSBC has assigned ratings on the following basis:

The target price is based on the analyst’s assessment of the stock’s actual current value, although we expect it to take six to 12

months for the market price to reflect this. When the target price is more than 20% above the current share price, the stock will

be classified as a Buy; when it is between 5% and 20% above the current share price, the stock may be classified as a Buy or a

Hold; when it is between 5% below and 5% above the current share price, the stock will be classified as a Hold; when it is

between 5% and 20% below the current share price, the stock may be classified as a Hold or a Reduce; and when it is more

than 20% below the current share price, the stock will be classified as a Reduce.

Our ratings are re-calibrated against these bands at the time of any 'material change' (initiation or resumption of coverage,

change in target price or estimates).

Upside/Downside is the percentage difference between the target price and the share price.

Prior to this date, HSBC’s rating structure was applied on the following basis:

For each stock we set a required rate of return calculated from the cost of equity for that stock’s domestic or, as appropria te,

regional market established by our strategy team. The target price for a stock represented the value the analyst expected the

stock to reach over our performance horizon. The performance horizon was 12 months. For a stock to be classified as

Overweight, the potential return, which equals the percentage difference between the current share price and the target price,

including the forecast dividend yield when indicated, had to exceed the required return by at least 5 percentage points over the

succeeding 12 months (or 10 percentage points for a stock classified as Volatile*). For a stock to be classified as Underweight,

the stock was expected to underperform its required return by at least 5 percentage points over the succeeding 12 months (or

10 percentage points for a stock classified as Volatile*). Stocks between these bands were classified as Neutral.

*A stock was classified as volatile if its historical volatility had exceeded 40%, if the stock had been listed for less than 12

months (unless it was in an industry or sector where volatility is low) or if the analyst expected significant volatility. However,

stocks which we did not consider volatile may in fact also have behaved in such a way. Historical volatility was defined as the

past month's average of the daily 365-day moving average volatilities. In order to avoid misleadingly frequent changes in rating,

however, volatility had to move 2.5 percentage points past the 40% benchmark in either direction for a stock's status to change.


31 August 2016

34

Rating distribution for long-term investment opportunities

As of 30 August 2016, the distribution of all independent ratings published by HSBC is as follows:

Buy 43% (25% of these provided with Investment Banking Services)

Hold 41% (26% of these provided with Investment Banking Services)

Sell 16% (18% of these provided with Investment Banking Services)

For the purposes of the distribution above the following mapping structure is used during the transition from the previous to

current rating models: under our previous model, Overweight = Buy, Neutral = Hold and Underweight = Sell; under our current

model Buy = Buy, Hold = Hold and Reduce = Sell. For rating definitions under both models, please see “Stock ratings and basis

for financial analysis” above.

For the distribution of non-independent ratings published by HSBC, please see the disclosure page available at

http://www.hsbcnet.com/gbm/financial-regulation/investment-recommendations-disclosures.

Information regarding company share price performance and history of HSBC ratings and target prices in respect of long-term

investment opportunities for the companies that are the subject of this report is available from www.hsbcnet.com/research.

To view a list of all the independent fundamental ratings disseminated by HSBC during the preceding 12-month period, please

see the disclosure page available at www.research.hsbc.com/A/Disclosures.

HSBC & Analyst disclosures

Disclosure checklist

Company Ticker Recent price Price date Disclosure

AU OPTRONICS 2409.TW 12.85 30-Aug-2016 4, 6 TAIWAN SEMICONDUCTOR 2330.TW 176.50 29-Aug-2016 5, 6 TPK HOLDING 3673.TW 59.10 29-Aug-2016 1, 4, 5, 6

Source: HSBC

1 HSBC has managed or co-managed a public offering of securities for this company within the past 12 months.

2 HSBC expects to receive or intends to seek compensation for investment banking services from this company in the next 3

months.

3 At the time of publication of this report, HSBC Securities (USA) Inc. is a Market Maker in securities issued by this

company.

4 As of 31 July 2016 HSBC beneficially owned 1% or more of a class of common equity securities of this company.

5 As of 31 July 2016, this company was a client of HSBC or had during the preceding 12 month period been a client of

and/or paid compensation to HSBC in respect of investment banking services.


and/or paid compensation to HSBC in respect of non-investment banking securities-related services.


and/or paid compensation to HSBC in respect of non-securities services.

8 A covering analyst/s has received compensation from this company in the past 12 months.

9 A covering analyst/s or a member of his/her household has a financial interest in the securities of this company, as

detailed below.

10 A covering analyst/s or a member of his/her household is an officer, director or supervisory board member of this

company, as detailed below.

11 At the time of publication of this report, HSBC is a non-US Market Maker in securities issued by this company and/or in

securities in respect of this company

12 As of 25 August 2016, HSBC beneficially held a net long position of more than 0.5% of this company’s total issued share

capital, calculated according to the SSR methodology.

13 As of 25 August 2016, HSBC beneficially held a net short position of more than 0.5% of this company’s total issued share

capital, calculated according to the SSR methodology.

35


31 August 2016

HSBC and its affiliates will from time to time sell to and buy from customers the securities/instruments, both equity and debt

(including derivatives) of companies covered in HSBC Research on a principal or agency basis.

Analysts, economists, and strategists are paid in part by reference to the profitability of HSBC which includes investment

banking, sales & trading, and principal trading revenues.

Whether, or in what time frame, an update of this analysis will be published is not determined in advance.

Economic sanctions imposed by the EU and OFAC prohibit transacting or dealing in new debt or equity of Russian SSI entities.

This report does not constitute advice in relation to any securities issued by Russian SSI entities on or after July 16 2014 and as

such, this report should not be construed as an inducement to transact in any sanctioned securities.

For disclosures in respect of any company mentioned in this report, please see the most recently published report on that

company available at www.hsbcnet.com/research. In order to find out more about the proprietary models used to produce this

report, please contact the authoring analyst.

Additional disclosures

1 This report is dated as at 31 August 2016.

2 All market data included in this report are dated as at close 26 August 2016, unless a different date and/or a specific time of

day is indicated in the report.

3 HSBC has procedures in place to identify and manage any potential conflicts of interest that arise in connection with its

Research business. HSBC's analysts and its other staff who are involved in the preparation and dissemination of Research

operate and have a management reporting line independent of HSBC's Investment Banking business. Information Barrier

procedures are in place between the Investment Banking, Principal Trading, and Research businesses to ensure that any

confidential and/or price sensitive information is handled in an appropriate manner.

4 You are not permitted to use, for reference, any data in this document for the purpose of (i) determining the interest

payable, or other sums due, under loan agreements or under other financial contracts or instruments, (ii) determining the

price at which a financial instrument may be bought or sold or traded or redeemed, or the value of a financial instrument,

and/or (iii) measuring the performance of a financial instrument.

Production & distribution disclosures

1 This report was produced and signed off by the author on 30 Aug 2016 11:18 GMT.

2 In order to see when this report was first disseminated please see the disclosure page available at

https://www.research.hsbc.com/R/34/GWVhs9P


31 August 2016

36

Disclaimer

Legal entities as at 1 July 2016

‘UAE’ HSBC Bank Middle East Limited, Dubai; ‘HK’ The Hongkong and Shanghai Banking Corporation Limited, Hong

Kong; ‘TW’ HSBC Securities (Taiwan) Corporation Limited; 'CA' HSBC Bank Canada, Toronto; HSBC Bank, Paris Branch;

HSBC France; ‘DE’ HSBC Trinkaus & Burkhardt AG, Düsseldorf; 000 HSBC Bank (RR), Moscow; ‘IN’ HSBC Securities and

Capital Markets (India) Private Limited, Mumbai; ‘JP’ HSBC Securities (Japan) Limited, Tokyo; ‘EG’ HSBC Securities Egypt

SAE, Cairo; ‘CN’ HSBC Investment Bank Asia Limited, Beijing Representative Office; The Hongkong and Shanghai

Banking Corporation Limited, Singapore Branch; The Hongkong and Shanghai Banking Corporation Limited, Seoul

Securities Branch; The Hongkong and Shanghai Banking Corporation Limited, Seoul Branch; HSBC Securities (South

Africa) (Pty) Ltd, Johannesburg; HSBC Bank plc, London, Madrid, Milan, Stockholm, Tel Aviv; ‘US’ HSBC Securities (USA)

Inc, New York; HSBC Yatirim Menkul Degerler AS, Istanbul; HSBC México, SA, Institución de Banca Múltiple, Grupo

Financiero HSBC; HSBC Bank Australia Limited; HSBC Bank Argentina SA; HSBC Saudi Arabia Limited; The Hongkong

and Shanghai Banking Corporation Limited, New Zealand Branch incorporated in Hong Kong SAR; The Hongkong and

Shanghai Banking Corporation Limited, Bangkok Branch

Issuer of report


13th Floor, 333 Keelung Road, Sec. 1,

Taipei, Taiwan

Telephone: + 886 2 2722 8458

Fax: + 886 2 2722 2056

Website: www.research.hsbc.com

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).

© Copyright 2016, HSBC Securities (Taiwan) Corporation Limited, ALL RIGHTS RESERVED. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, on any

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]

Europe

Analyst Nicolas Cote-Colisson +44 20 7991 6826 [email protected]

Analyst Antonin Baudry +33 1 56 52 43 25 [email protected]

Analyst Christopher Johnen +49 211 910 2852 [email protected]

Analyst Dominik Klarmann, CFA +49 211 910 2769 [email protected]

Analyst Sebastian Grabert +49 211 910 1096 [email protected]

Analyst Luigi Minerva +44 20 7991 6928 [email protected]

Analyst Olivier Moral +33 1 5652 4322 [email protected]

Analyst Adam Fox-Rumley +44 20 7991 6819 [email protected]

Analyst Dhiraj Saraf, CFA +91 80 3001 3773 [email protected]

Americas

Analyst Christopher A Recouso +1 212 525 2279 [email protected]

Analyst Ronny Berger, CFA 44 20 7991 2750 [email protected]

Analyst Sunil Rajgopal +1 212 525 0267 [email protected]

Global Emerging Markets (GEMs)

Analyst Hervé Drouet +44 20 7991 6827 [email protected]

Emerging Europe, Middle East & Africa (EMEA)

Analyst Ziyad Joosub +27 11 676 4223 [email protected]

Analyst Eric Chang +971 4 423 6554 [email protected]

Asia

Analyst Yogesh Aggarwal +91 22 2268 1246 [email protected]

Analyst Vivek Gedda +91 22 6164 0693 [email protected]

Analyst Vikas Ahuja +91 22 3396 0690 [email protected]

Analyst Neale Anderson +852 2996 6716 [email protected]

Analyst Angela Tay +65 6658 0612 [email protected]

Analyst Joyce Chen +8862 6631 2862 [email protected]

Analyst Jenny Lai +8862 6631 2860 [email protected]

Analyst Carrie Liu +8862 6631 2864 [email protected]

Analyst Bruce Lu +8862 6631 2861 [email protected]

Analyst Steven C Pelayo +852 2822 4391 [email protected]

Analyst Ricky Seo +822 37068777 [email protected]

Analyst Rajiv Sharma +91 22 2268 1239 [email protected]

Analyst Jerry Tsai +8862 6631 2863 [email protected]

Analyst Chi Tsang +852 2822 2590 [email protected]

Analyst Terry Chen +852 2996 6635 [email protected]

Analyst Jena Han +822 3706 8772 [email protected]

Analyst Will Cho +822 3706 8765 [email protected]

Analyst John Liu +852 2822 4392 [email protected]

Associate Aric Hui +852 2822 3165 [email protected]

Associate Qin Wang +852 2822 4393 [email protected]

Associate Wayne Wang +852 2914 9935 [email protected]

Associate Kenneth Shim +822 3706 8779 [email protected]

Associate David Huang +886 2 66312865 [email protected]

Specialist Sales

Gareth Hollis +44 20 7991 5124 [email protected]

Tarun Viswanathan +44 20 7991 7843 [email protected]

Kubilay Yalcin +49 211 9104880 [email protected]

Myles McMahon +852 2822 4676 [email protected]

Global Telecoms, Media & Technology Research Team

asia display electronic equipment -...

Documents