Huawei's predicament creates uncertainty over demand/investment until year-end, great opportunity for Samsung
We visited major IT H/W companies and research institutions in Japan and Taiwan from May 27 to 31 and also went to Computex 2019 to examine current industry issues. What we learned from our visits is as follows: 1) although Huawei will try to build up its inventory, its orders could stop abruptly in 4Q19 as its inventory of US-made RF parts is expected to be depleted by end-3Q19; 2) Intel (INTC) unveiled its 10nm Ice Lake CPU for notebook and the notebook CPU shortage will likely ease from end-3Q19; hyperscale server demand is expected to increase over 10% HoH in 2H19 as cloud companies upgrade their data centers with Intel Cascade CPUs; and 3) due to murky Chinese demand, Apple’s (NASDAQ:AAPL) iPhone production volume should decline 21.5% YoY in 2H19.
Despite growing uncertainties, Huawei’s plight create opportunity for Samsung to expand M/S
The Huawei ban increases uncertainties in the short term, characterized by: 1) the “order cliff” from Huawei; 2) Chinese boycott against Apple, and 3) delay of 5G infrastructure investment. However, investments in 5G are likely to pick up momentum again from 2020 onwards as Ericsson’s (ERIC), Nokia’s, (NOK) and Samsung’s (OTC:SSNLF) telecom equipment replace Huawei’s. What is certain is that Huawei sanctions will provide a strong catalyst for Samsung’s market share in smartphones, semiconductors, and telecom equipment in the mid to long term.
Our top picks are Samsung and SEMCO, for Hynix, look for bargain-hunting opportunities
We present Samsung and SEMCO (OTC:SMSGF) as our IT H/W sector top picks for they will benefit the most from Huawei’s plight in the mid to long term. As for Hynix (OTC:HXSCL), we advise investors to look for a bargain-hunting opportunity as it is set to benefit from a recovery of data center demand and increase in Chinese market share in the mid to long term.
Key takeaways from visits to Asian IT H/W companies
Huawei obscures demand/investment outlook, opportunity for Samsung
We visited major IT H/W companies and research institutions such as DRAMeXchange, Stone Partners, and InnoResearch in Japan and Taiwan from May 27 to 31 to examine current industry issues such as: 1) uncertainty around Huawei; 2) 2H19 server demand coming from hyperscale data centers; and 3) the 2H19 iPhone production volume outlook.
In Computex 2019, Intel unveiled its 10-nanometer CPU Ice Lake, stressing that notebooks equipped with the said CPU will become available by the end of the year. Ice Lake for servers is also scheduled for release at the beginning of 3Q20 to cash in on 5G- and AI-related demand. Demand for Cascade Lake is expected to increase but because Ice Lake is set for 3Q20 launch, Facebook (FB) and Alibaba (BABA) will likely skip Cascade to directly migrate to Ice Lake. On the other hand, Microsoft (MSFT) and Amazon (AMZN) are expected to place orders for data centers from 3Q19. Our meeting with Aspeed, which has an 80% market share in Baseboard Management Controllers (OTCPK:BMCS) for hyperscale servers also confirmed that its sales are slated to grow at the low 10% level this year. Huawei, currently drowning in uncertainty, has the fourth-largest server market share (6.4%), but over 90% of its servers are for general enterprises. Thus, we believe Huawei’s impact on the hyperscale server market will be limited.
What has been most talked about in our meetings with Taiwanese and Japanese companies was the possibility of change in competitive environment in the industry because of Huawei. In fact, Huawei appears to be placing massive orders to the Taiwanese supply chain to get the most out of the 90-day grace period which expires on Aug. 19. As a result, domestic companies such as Samsung Electronics (Samsung) and SK Hynix (Hynix) are expected to benefit in the short term. However, it has to be noted that all the products with a Huawei mark are finished products, which means that they cannot be produced if a single part is missing. Therefore, from end-3Q when the inventory of US-made parts is expected to be depleted, Huawei will likely stop ordering parts made by Korean companies too, which could lead to an “order cliff.” The 2011 incident could be repeated when Korean IT H/W companies enjoyed a short rally after the 2011 Japan earthquake only to pare their gains later on concerns that they would not be able to procure all the necessary parts.
Additionally, Chinese smartphone demand, which accounts for 27.6% of global smartphone demand, is expected to cool down faster than expected. The wave of patriotic support may boost demand for Huawei and other Chinese smartphones. However, combined with the Apple boycott and Huawei’s likely failure to get the parts it needs from 4Q19 could work to lengthen Chinese customers’ smartphone replacement cycle.
However, in the mid to long term, the Huawei ban should be an opportunity for Korean companies such as Samsung and Samsung Electro-Mechanics (SEMCO). As the awareness that heavy dependence on US semiconductor and parts makers could be a risk grows, the preference for Korean core components may increase. In the short term, the Huawei ban could negatively affect Hynix for its substantial exposure to the company. However, as other Chinese names begin to replace Huawei’s place, their preference for Korean parts suppliers are likely to grow in the mid to long term.
As the trade war leads to the US ban on Huawei, companies with high sales exposure to Huawei and Apple are likely to report lower-than-expected earnings. Of course, 2Q will see those companies’ earnings grow as Huawei moves to fill its inventory during the grace period. However, from end-3Q, the orders from Huawei will likely drop sharply. As such, we present Samsung and SEMCO as our sector top picks as both companies have smaller exposure to Huawei than peers and large exposure to Samsung smartphones. As for Hynix, its high exposure to Huawei makes its earnings outlook uncertain after 3Q but we advise investors to exploit bargain-hunting opportunities as public cloud companies will begin to build data centers again from 3Q. As for LG Innotek, we lower our rating to Marketperform because of uncertainties looming over Huawei and its North American client.
Insufficient supply from Huawei to slow down 5G investment
The US ban on Huawei seems to have been motivated by the need to ban telecom equipment out of China, as in the case with ZTE (OTCPK:ZTCOF). Telecom equipment for cell towers is closely related to the aviation and defense industries. The rise of 5G and AI technologies will help produce more powerful and destructive unmanned weapons, and the technologies used in these weapons are highly related with telecom equipment technologies. In fact, in the spring of 2001 before the September 11th terrorist attacks, a US fighter was shot down in Iraq. The US government believes Iraq does not have the technology to build a state-of-the-art air defense system on its own and identified China’s Huawei as a collaborator. Huawei has been US’ watch list since then. As for servers, US dependence on Huawei is negligible. Some medium-sized US firms use enterprise network equipment such as switches and routers but these are not significant enough to be a threat to national security. If we decide to be more logical, Lenovo (OTCPK:LNVGY) is a bigger threat than Huawei in the enterprise network equipment and servers. In the end, the focus of US sanctions on Huawei is telecom equipment, which accounts for 45% of Huawei’s total sales and the core of future AI network.
The reason that Huawei's telecom equipment is so competitive is because Huawei's growth was backed by the Beijing’s support of the country’s telecommunication infrastructure as well as the development of China’s own telecommunication technology (e.g., TD-SCDMA). In the case of Huawei and ZTE in particular, Beijing generated sales for them by actively investing in telecommunication infrastructure in Africa, and providing Huawei and ZTE's handsets at inexpensive prices. This way, both companies were able to obtain solid sales base in the global market. Then why do many global telcos opt to use Huawei’s equipment? It is because Huawei's products are highly price competitive. Plenty of telecom equipment companies have gone bankrupt during the transition to IMT-2000 and broadband Internet. Nortel Networks (the world’s biggest optical transport equipment maker), Lucent Technologies of the US, and Alcatel of France have all gone bankrupt. Nokia Networks (NOK) expanded its market share by acquiring Alcatel-Lucent. While Ericsson and Nokia claim the number two and three positions in the global telecom equipment market, their profitability is highly volatile. Samsung also suffers from thin margins in this space, although it has seen some improvements in 2018.
Telecom equipment providers tend to sell their equipment at expensive prices during periods of infrastructure building such as 3G and 4G because this is the only time they can enjoy massive earnings growth. This increases capex burdens for telcos. Huawei was able to gain a huge market share by offering its equipment at competitive prices to major telcos like Nokia and Ericsson when these telcos were withdrawing from the handset business and were highly dependent on telecom equipment. Huawei was able to significantly increase its presence in the telecom equipment market with attractive price offerings, while enjoying higher profitability in handsets and servers.
It may be difficult for a company to survive with only telecom equipment since the R&D burden is huge and demand is highly volatile. Telcos consider Huawei’s telecom equipment as the most cost effective. Above all else, Huawei has established a virtuous cycle where its stable product mix helps reduce the volatility of the business. It's clear that the Chinese government’s support enabled the company to establish such a virtuous cycle. However, at this stage, it's unclear to what extent the government has lent support to Huawei since the company is unlisted.
Huawei's sales have grown 25.8% over the past four years, and its 2018 operating margin was 10%. Since telecom equipment accounts for 42.4% of total sales, the telecom equipment division is estimated to have posted an operating margin of more than 10%, which is higher than second-tier companies. With Huawei’s global market share of telecom equipment rising to 31%, Washington’s sanctions against Huawei seems to be a little late, and the pain caused by Huawei’s departure could be especially hurtful for telcos. The fact that Huawei’s equipment was more price competitive vs. Ericsson’s and Nokia’s offerings could slow down telcos’ 5G investment. It could take time for Ericsson, Nokia, and Samsung to expand production capacity in the telecom equipment industry. Furthermore, some core parts need to clear the approval process. In all, the Huawei ban is expected to slow down telcos’ 5G infrastructure investment both in terms of replacement periods and replacement costs.
Among Huawei’s base station equipment, RF parts such as power amplifier seem to be highly dependent on Qorvo (QRVO) and Skyworks (SWKS). Qorvo is a US-based company that was created through the merger of RF Micro Device and TriQuint, and has a top-notch technology in 5G RF parts for base stations. Skyworks also is a highly competitive US firm in RF components for base stations such as amplifiers and attenuators. Broadcom, which was acquired by Avago, also has a top-tier technology in RF components. Although Huawei says that it has one year’s worth of inventory for base station parts, the important point is that the company makes finished products, meaning even if a single part is missing, the rest of the parts are useless. It's unclear as to whether Huawei has devised a plan B. What's certain is that even if it finds RF parts companies that can replace its US counterparts, uncertainties will still remain in terms of produc(GOOt performance and test period.
Meanwhile, Huawei procures most smartphone PMIC and RFIC from subsidiary Hisilicon. The foundry of these RF parts is believed to be provided by Taiwan's Win Semiconductors. As for Hisilicon’s AP, TSMC (TSM) is believed to provide the foundry service.
Although Huawei is the fourth-largest server company in the world, Huawei's share of hyperscale servers for data centers is negligible, and most of its sales are generated by Tencent. As such, even if Huawei fails to procure server CPUs from Intel, its impact on the market will be minimal. Tencent (OTCPK:TCEHY) can work with Inspur and Lenovo instead. Of note, Inspur is believed to receive substantial support from the Chinese government.
Huawei’s pain, Samsung’s gain in smartphones
Huawei’s smartphone market share overseas is expected to fall sharply as the major players of the smartphone ecosystem such as ARM and Google (NASDAQ:GOOG) (GOOGL) are not going to supply the necessary IP and solution to Huawei. In 2018, Huawei sold more than 100mn units of smartphones in China and another 100mn in overseas markets, which is close to what Apple sold in terms of volume. Since 2H18, its sales volume has surpassed Apple and became formidable enough to threat Samsung’s position. Unlike Samsung and Apple, Huawei’s market share in North America was a paltry 0.3% due to North American carriers’ reluctance to support Huawei smartphones. Thus Huawei’s growth outside of North America is a very impressive feat. However, since Huawei smartphones are not going to able to support Play Store, YouTube, and Gmail, its market share outside of China is expected to plummet. Within China, Google’s key apps have not been in service, so in the short term, patriotic sentiment could boost Huawei’s market share in the country considerably. Meanwhile, Huawei is expected to use its own operating system Ark (Chinese name Hongmeng) in the future, but it's unlikely to succeed outside of China as seen from the case of Nokia which met its demise by sticking to its own Symbian OS.
However, if ARM stops providing its IP to Huawei’s AP suppliers, Huawei would not be able to upgrade its smartphones’ AP and differentiate its smartphones from others, which could all but mean the demise of the Huawei brand in the Chinese smartphone market. ARM is a dominant player in the mobile AP market and its rival x86 is currently dominated by Intel and AMD. The only solution for Huawei is to design its own CPU architecture, but it will take a lot of time, not to mention many trials and tribulations that will likely happen on the way. Most likely, Huawei’s presence may all but disappear before its efforts come to any fruition. Some argue that Huawei can survive in the domestic market by illegally procuring necessary parts but it's a far cry from the key agenda of the ongoing trade dispute: Protection of intellectual property rights.
In the short term, Chinese consumers are likely to grow more attached to the Huawei brand and the demand for Apple and other Chinese brands will likely dwindle. As a result, the smartphone replacement cycle in China is expected to become longer. Against this backdrop, we cut our 2019 smartphone demand forecast in China and the world by 4.8% and 3% to 380mn and 1.4bn units, respectively.
To conclude, the Huawei sanctions by the US and ARM are expected to offer more opportunities for Samsung. If Huawei's market share climbs to 40% in China in the short term, the companies that are likely to hurt the most are Apple and Oppo/Vivo. If the sales of Huawei smartphones fall sharply in its key markets such as Europe and Latin America, it means a great opportunity for Samsung smartphones. In Southeast Asia, Oppo/Vivo and Xiaomi (OTCPK:XIACF) will likely be the major beneficiaries of Huawei's market share falls. In India, no major changes are likely since Huawei has only a 3% market share.
For Oppo, Vivo, and Xiaomi, any advantages they gain in Southeast Asia and India from the fall of Huawei will likely be offset by a market share decline in China. Apple's 8% market share in China will likely go to Huawei mostly. After losing China, Apple's total smartphone shipments are expected to decline more than 15% from 2018. In fact, Taiwanese EMS companies' iPhone production volume in 2H19 is expected to fall 21.5% YoY.
In the meantime, Samsung will likely remain mostly unaffected, as China represents only 2% its smartphone sales. Of course, there's a good chance that another Chinese company will grow to replace Huawei in the long term. However, it will take a serious amount of time before that happens because Huawei is China's top IT H/W company in terms of sales, R&D capability, profitability, and product lineup. Against this backdrop, we revise up Samsung's smartphone shipment forecast this year by 5.2% from our previous one to 300mn units. On the other hand, we revise down our shipment forecasts for Apple and Huawei by 8.4% and 18.8%, respectively, to 180mn units and 200mn units.
Of course, there's a possibility that the sanctions against Huawei smartphones will ease going forward as these are not directly related to the future defense and aerospace industries. However, even if this happens, traumatized Chinese smartphone makers led by Huawei will likely move to reduce their dependence on US parts, and Korean parts makers are likely to benefit as a result. As such, Korean parts suppliers will be able to expand their market shares in China in the long term, although Hynix is likely to get affected in the short term because of its exposure to Huawei.
Aspeed’s guidance of 10% YoY growth in sales not going to be easy
We also had a meeting with Aspeed, a major BMC supplier to most server companies except HP (HPQ), Dell (DELL), Huawei and Google. Dell gets its BMC from Nuvoton while HP, Huawei, and Google develop and make BMC themselves. Aspeed's market share of server BMC is 65%, but its market share of hyperscale BMC is 80%, as other server companies’ (HP, Dell and Huawei) presence in hyperscale server is not as big. Therefore, Aspeed is the best indicator of hyperscale server demand from data centers. BMC accounts for 92% of Aspeed’s sales, and TSMC currently provides the foundry service. Normally, a server requires one BMC. In the future, more hyperscale servers will come with two BMCs, led by Intel. While Aspeed currently provides 40nm BMC, it will likely migrate to 28nm products in the future.
BMC accounts for 92% of the company's sales, and 70% of BMC sales are for the data centers and the rest 30% for general enterprises. As such, we believe 64.4% of Aspeed’s sales are linked with hyperscale server demand. We have found that Aspeed's sales precede the demand from server ODM companies by two months, and CPU/memory chip makers’ demand by three months. Aspeed guided for low-10% YoY growth in sales for this year. However, cumulative sales from January to April were only 0.3% YoY. To be able to meet its 2019 guidance, Aspeed’s sales have to grow at least 20% YoY in 2H19. This is not an easy feat considering the uncertainties surrounding Huawei. Of course, it is notable that its 1Q19 sales increased 13.4% QoQ and tis April sales are 5% higher than March sales, which is a signal that global server shipments have bottomed out in 1Q19. Even though the Huawei issue will keep us from seeing a V-shaped recovery in global server demand, we believe a U-shaped recovery is possible.
Intel to launch 10nm Ice Lake server CPU in 3Q20
In Computex 2019, Intel unveiled its 10-nanometer notebook CPU Ice Lake, and companies like Acer, Asus, and Lenovo are expected to roll out notebooks with the CPU before the end of the year. Intel’s disclose of Cascade Lake as well as the 10th-generation 10nm CPU in early April has helped dissipate worries about the market. Intel’s launch of the 10nm Ice Lake CPU is also expected to alleviate concerns about a potential supply shortage of Intel’s notebook CPU. Intel is currently working to expand its CPU capacity at the Nevada plant by more than 20%, and plans to solve the supply shortage by migrating to a 10nm notebook CPU. Taiwan’s notebook supply chain expects the supply shortage of Intel CPUs to ease from September. Accordingly, notebook demand, which has been dwindling along with server demand, should recover from 2H19. We may even see a decoupling of PC DRAM prices from server CPU prices from 3Q19.
Meanwhile, Intel unveiled its Cascade CPU in early April, and Microsoft’s data centers are expected to use the hyperscale servers with the CPU from July. Amazon is also expected to resume its investments in data centers from September, if not sooner. Google, Tencent, and Baidu are also likely to equip their data centers with Cascade CPUs. However, Facebook and Alibaba will probably take a rain check on Cascade, the major reason being the short product cycle of Cascade CPUs. They will likely directly migrate to Intel’s Ice Lake-based 10nm server CPU, to be released in July next year. However, given Intel’s history of launch delays, the decisions by other companies to use the Cascade CPU is meaningful enough. Meanwhile, Intel’s Ice Lake is a 10nm-based CPU, but most believe that Intel’s new CPU will still outperform AMD’s (AMD) 7nm-based CPU. Intel is expected to introduce photolithography from the 7nm process. Intel also plans to introduce DDR5 memory support from 2021’s Sapphire Rapids. Meanwhile, after visiting Japan’s SUMCO (OTCPK:SUOPY), we confirmed that Intel's wafer demand is the strongest, and supply is smooth as opposed to market worries.
Understanding Taiwanese server companies: EMS, ODM, brands
In our visits to server ODM and brands in Taiwan, our major focus was on the server supply chain. Taiwan has a number of EMS (electronics manufacturing service) providers and ODM (original design manufacturing) companies. In servers, HP, Dell and Lenovo are seeing their market shares declining in HPC (high performance computing) servers. In the meantime, public cloud companies are trying to differentiate themselves from peers by designing their own servers and using their own AI chips. For this reason, Taiwanese ODM companies’ global market share is increasing rapidly. Taiwan’s server companies are classified into three categories: the first are EMS providers like Foxconn (OTC:FXCOF). Foxxcon’s strength is competitive production cost, as it produces a full range of finished products such as PCs, TVs, smartphones, and servers. The second are ODM companies such as Inventec, Quanta, Wistron, and Mitac. The third and the last are brand companies specializing in servers, such as Wiwynn and QCT (Quanta Cloud Technology). Wiwynn, a server brand spun off from Wistron, is developing hyperscale servers for Facebook and Microsoft. Wistron is believed to be responsible for manufacturing, including SMT processes.
QCT, formerly the server division of Quantra, has recently seen growth in shipments to Chinese telcos and other companies. It is especially picking up marketing of micro data center products for edge computing.
Unlike other ODM companies in Taiwan, servers represent a large portion of ODM sales at Inventec. In 1Q19, servers made up 34% of its total sales. Inventec could be a key gauge of corporate demand as its notebooks are mostly commercial laptops rather than consumer laptops. Mitac is a major ODM company for Inspur in China. Inventec expects hyperscale server orders from Microsoft and Amazon to increase from 3Q19, and all clients will ratchet up their orders from 3Q20. This is because it expects the launch of the Ice Lake server CPU in 3Q20 to drive Facebook and Alibaba to invest in servers.
Huawei's place is likely to be replaced by China's Inspur, Lenovo or Sugon. Huawei's share of the global server market is 6.4% and its server DRAM is mostly supplied by Hynix and Micron (MU). Hyperscale servers represent about 10% of Huawei’s server sales, most of which go to Tencent. Inspur and Lenovo look to have a strong presence in the Chinese hyperscale server market. Inspur, in particular, is enjoying a rapid increase in market share in China by offering quality products at competitive prices. As Huawei is unlikely to be able to buy Intel’s server CPUs, Inspur’s market share in China should grow sharply. Hynix may want to work more closely with Inspur if it wants to minimize the impact of declining orders from Huawei.
1Q19 DRAM sales down 28.6% QoQ; Huawei ban weighs on recovery
According to DRAMeXchange, DRAM sales in 1Q19 fell 28.6% QoQ to USD16.3bn, due mainly to an over 20% QoQ decline in DRAM prices. In 2Q19, Samsung and Hynix are expected to experience higher bit growth but value- wise, a QoQ decline in DRAM sales appears inevitable, as DRAM prices are expected to fall 30% QoQ. Despite the sanctions imposed against Huawei, the DRAM market will probably remain mostly unaffected in 2Q as Huawei tries to build its server and mobile DRAM inventories. The problem is 3Q when Huawei’s inventory of US-made key parts is expected to be depleted, because that is when Huawei should stop ordering DRAM. Meanwhile, a sales decline of the new iPhone to be released in 2H19 will likely have a fairly limited impact on the DRAM sector, because the total DRAM capacity of the new iPhone will likely be limited to 4GB even if it comes with triple cameras, which is less than 50% of Android phones. A strong Chinese boycott against Apple will have the most damaging impact on NAND products such as UFS. That said, if Samsung manages to chip away at the market shares of Apple and Huawei smartphones, then its memory semiconductor business should fare relatively well.
Meanwhile, the Huawei ban will likely result in a 30% QoQ fall in server DRAM price in 2Q19, which will be a 56.5% drop from the 2018 peak. If server DRAM prices fall another 15% QoQ in 3Q and 8% in 4Q, it will be even lower than 2016’s bottom. However, since hyperscale servers’ DIMM capacity is expected to grow more than twofold, DRAM makers should be able to buffer the shock of falling prices through bit growth. The good news is that Huawei's presence in the hyperscale server market is miniscule, and other Chinese companies that are expected to fill the void such as Inspur and Lenovo should help to alleviate uncertainties in the mid to long term. In the short term, however, Huawei’s plight will negatively affect Chinese server DRAM demand considering the supplier approval and verification process. With Micron not selling its serer DRAM to Huawei, other Chinese firms could grow averse to using US-made memory chips, which puts Korean chipmakers in a very advantageous position. Having said that, if Micron decides to sell its chips at a discount to Chinese server companies to clear its inventory, it could negatively impact server DRAM prices in 2H19.
In the NAND space, as Western Digital (WDC) is expected to stop supplying Huawei, Korean companies' market shares are bound to increase. After YMTC unveiled SSDs based on 64-layer 3D NAND flash memory, there have been concerns over Chinese companies’ entry into the market. We have confirmed that the controller IC of YMTC’s SSD is supplied by Silicon Motion (SIMO). However, although Silicon Motion has been a total solution provider for YMTC, YMTC’s SSDs still need further verification.
Korean IT H/W sector investment strategy
Top picks Samsung and SEMCO; wait to bargain-hunt Hynix
Contraction of demand in the IT sector triggered by the US-China trade war is worsening as the US moved to place sanctions on Huawei. The problem is not just contracting demand, but increasingly uncertain demand. Huawei sells finished products so even if it lacks just one part, it could delay the purchase of other parts, and we do not know how severe this problem will be. Furthermore, the potential Chinese boycott against Apple products could increase the burden of the Apple supply chain. It should be also noted that the Huawei ban came about when Huawei, as the world’s biggest telecom equipment provider, was escalating its promotion of 5G equipment. 5G was expected to enrich the user experience of AI and become a catalyst for IT demand, ultimately ushering in the second big memory semiconductor cycle. In conclusion, the US government's decision to cut off Huawei from American technology has fueled uncertainties over the following three issues: 1) order cliff from Huawei; 2) Chinese boycott against Apple; and 3) delay of 5G infrastructure investment.
The good news is that Huawei is not the exclusive player in the telecom equipment market. It has a 31% market share but there are other companies that can make up for Huawei’s absence such as Ericsson and Nokia, although their prices are on the higher end. 5G investments will likely lose some steam this year, but the pace is expected to pick up globally from 2020 onward. As for servers, it may take time to switch to different server companies, but given that China is a major server market, it is likely that it will be replaced with other Chinese server vendors sooner than expected. For smartphones, the Huawei ban could lead to Chinese consumers’ boycott against Apple, which could result in a sharp increase in Samsung's market share. Of course, lower iPhone sales volume is negative for Samsung Display’s OLED sales. However, the expected hike in Samsung smartphone sales will buoy mobile DRAM demand as Samsung smartphones consume high-density mobile DRAM. As for Samsung Display’ OLED, Samsung’s IM business could offset a decline in orders from Apple.
One way or another, the Huawei ban will help Samsung and SEMCO to enlarge their presence in their relevant markets in the mid to long term. Hynix's mobile business will have to feel some impact as it is highly exposed to Huawei and Apple but in the mid to long term, Chinese companies’ reluctance to working with Micron will make up for the loss. LG Innotek has low exposure to Huawei but if the Apple boycott spreads, uncertainties over 2H19 earnings could increase. The Huawei ban has helped draw a clear line between the stocks we like and do not like. As top picks, we present Samsung and SEMCO for they are set to benefit the most from Huawei’s plight. As for Hynix, we advise investors to look for a bargain-hunting opportunity as the company, despite short-term earnings uncertainty, could benefit from a recovery of data center demand and increase its market share in the mid to long term.
To sum up, among Korean IT H/W names, we favor Samsung the most, followed by SEMCO, Hynix, LG Electronics, and LG Innotek.
Overseas, we present Aspeed and Wiwynn as our stocks of interest for their earnings are expected to pick up in 2H19 on the back of increased investments in data centers.
Huawei’s pain, Samsung’s gain
We reiterate BUY on Samsung Electronics with a six-month-forward target price of KRW52,000 (1.4x 2019F BPS). Despite a weaker KRW, server DRAM prices plunged on the US ban on Huawei. As such, we revise down our 2Q19 operating profit estimate by 6.6% to KRW5.7tn. While it takes time for Samsung to benefit from Huawei’s faltering smartphone sales, inventory declines in memory semiconductors have a more immediate impact. In light of the change of our DRAM and NAND price projections for 3Q and 4Q, we trim our 2019-2020 operating profit forecasts by 7.8% and 13%, respectively. Having said that, we note that Samsung’s semiconductor sales exposure to Huawei is relatively small, while Samsung Display’s rigid OLED business is likely to feel some impact, its flexible OLED will help offset the possible decline when Samsung’s flagship smartphone shipments increase. While it is a certainty that Samsung’s leadership in the smartphone and semiconductor space will strengthen in the mid to long term, falling memory chip demand and inventory cuts are likely to affect this year’s earnings negatively. As for displays, it is possible that the volume supplied to Apple for its new product to be launched in 3Q may decline, but it is not a foregone conclusion as Apple may attempt to defend its top line by increasing the share of Max models which have bigger price premiums, and Samsung’s IM division could help offset the potential volume decline.
Major issues and earnings outlook
Considering that Huawei used to pose a serious threat to Samsung in the smartphone space and that Huawei has not been the biggest buyer of Samsung’s semiconductors and displays, we believe Samsung’s leadership in the components industry could strengthen further. Furthermore, Chinese finished product makers’ growing aversion to US-made components could also provide a long-term opportunity for Samsung. We advise investors to buy and hold, focusing on the prospect of stronger market leadership in the long term rather than short-term earnings.
Look at server DRAM, not Huawei
We reduce our six-month-forward target price from KRW92,000 to KRW82,000 (1.25x 2019F BPS) but keep our rating at buy given the likely recovery in demand for memory chips for data centers from 3Q19. We revise down our 2Q19 sales and operating profit forecasts by 2.1% and 47% to KRW6.87tn and KRW534.6bn, respectively, as we now expect DRAM price declines, especially server DRAM, to be more severe than previously anticipated. Hynix has substantial exposure to Huawei in server and mobile DRAM. As such, its 2Q19 shipments will likely be strong as Huawei will probably want to fill up its inventory in advance. However, there remains the possibility that the orders from Huawei will plummet starting in end 3Q19. Moreover, Apple may not be able to sell as many of its new iPhones in 2H19 as it plans. Taken together, we expect mobile semiconductor earnings to be sluggish. In particular, price competition may intensify if Micron decides to slash prices to get rid of the volume that would otherwise have been supplied to Huawei, which could fuel uncertainties over mobile semiconductor earnings. That said, the key driver of orders in 2H19 will be server DRAM, and data centers’ server orders are likely to pick up in earnest from 3Q with the launch of Intel’s Cascade CPU, which leads us to believe that earnings will rebound after bottoming out in 2Q.
Major issues and earnings outlook
Since 2017, the key driver of the memory semiconductor big cycle has been server DRAM for hyperscale data centers. With 5G services kicking off in earnest with Intel’s new CPU, it would be wise to approach the Huawei issue as nothing more than a hiccup. There's a possibility that Chinese partners will become more averse to using US-made semiconductor parts. In that case, Hynix stands to benefit in the mid to long term. The major reason that Hynix has substantial sales exposure to Huawei is because Huawei prefers Hynix’s memory semiconductors in order to achieve rapid growth. In all, we believe Hynix’s position as the world’s second-largest DRAM maker will hold. We recommend investors to take advantage of bargain-hunting opportunities from 3Q19 when server DRAM demand is expected to recover.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: Hyundai Motor Company is a passive shareholder in our bank.