Western Digital Looks to Flash for the Future

After taking a fresh look at hard disk drive manufacturer

Western Digital is one of two leaders in the hard disk drive market and, with the purchase of SanDisk in 2016, has become a significant player in the NAND flash market. The company has invested significant resources in successfully negotiating the technological transition from HDD to flash storage. Solid-state drives have important performance benefits versus older spinning disc media, and as a result, HDDs have lost their luster for both PCs and servers. HDDs currently have superior cost metrics and higher capacities, but SSDs have been closing these gaps and quickly limiting the number of use cases for HDDs. We believe Western Digital has significant competitive advantages in the HDD market, but these are unlikely to generate excess returns on capital as that storage solution continues to decline.

We believe the secular headwinds for HDD demand in PCs will continue. With a lower total cost of ownership from SSD products in many PC applications, we anticipate there will be an ongoing decline in annual HDD sales as SSDs increase share and PC unit sales generally decline. HDDs will remain a significant component of the storage environment to accommodate the business-critical storage needs of enterprises, but SSDs have already made significant headway in the enterprise market by taking a major portion of the share in mission-critical applications. While Western Digital’s SSD capabilities give it the potential ability to curb overall HDD declines in certain enterprise applications, thus far we believe the company has been unable to translate its success in enterprise storage to share gains in the SSD enterprise market.

Western Digital maintains a joint venture (Flash Ventures) with Toshiba to develop and manufacture flash memory wafers. While flash competitor Samsung does have an economic moat, we do not believe moats are easily developed in the flash market, given the commoditized nature of that technology, the capital intensity, and the highly cyclical nature of the semiconductor market.

We Don't Foresee Excess Returns on Capital Despite its place as one of two leaders in the hard disk drive market, we do not believe Western Digital has an economic moat. Notwithstanding considerable intangible assets surrounding HDD design expertise as well as a cost advantage in HDD manufacturing, its capabilities are unlikely to allow the company to generate excess returns on capital over the next decade, given the broad technological trends away from legacy hardware. Flash technology, specifically in the form of solid-state drives, have performance improvements over HDDs that have resulted in an ongoing shift toward SSDs for mass storage needs in PCs and servers. Western Digital has attempted to offset declines in the legacy HDD market by spending considerably on new HDD technologies in addition to purchasing flash player SanDisk in 2016. While the combined company has an advantage over competitor Seagate Technology STX, we believe Western Digital will struggle to maximize its opportunities in NAND because of the considerable competition in the space.

Over the past decade, the HDD market has experienced rapid consolidation, with the number of players falling from 10 to three as others have exited because of an inability to meet the rising investments to keep pace with higher-capacity HDDs and the shift to SSDs. The result of this consolidation is that Western Digital and rival Seagate now own a combined share in excess of 80% of the HDD market, with Toshiba playing a bit part. While the oligopoly in HDDs should allow rational behavior from both companies and help prevent price wars, the future of HDDs as a technology is in serious decline. Innovations from both companies may help to stave off obsolescence in the near term, but the next decade promises to limit the number of use cases. For context, industry estimates in 2005 for average selling price per gigabyte of storage on SSD drives were over $50. That declined to less than $0.40 in 2018 amid considerable investments from NAND players, and we estimate it will continue to decrease to $0.10-$0.15 by 2022.

HDDs still offer a cost advantage over SSDs, as HDD storage is as low as $0.03 per gigabyte. However, SSDs still have significant advantages, even at a higher cost per bit. SSDs benefit from improved performance, speed, lower power consumption, increased durability (no moving parts), and a smaller form factor. Indeed, even at multiple times the cost per bit, there are several use cases where the total cost of ownership in using SSDs is far lower than using HDDs. Though the cost advantage of HDDs over SSDs is progressively narrowing, the need for cheap high-capacity storage in data centers will continue to support the demand for enterprise HDDs, which carry higher average selling prices and margins than PC HDDs. We expect the company to benefit from an acceleration in cloud deployments and subsequent need for massive amounts of high-capacity storage. At present, SSD remains a costly product for high capacities, and we anticipate that business-critical storage (that is, "nearline" storage not needed for immediate access) will continue to be in the hands of HDD technology. However, while we model for average selling price and exabyte growth from business-critical HDDs, flash technology has already made significant headway in the mission-critical piece of enterprise storage, with SSDs accounting for nearly 20% of the mission-critical market in 2018 from a base of just 8% in 2015. In our view, enterprise HDD demand will not fully offset headwinds from SSD substitution and a declining PC market.

At the moment, PCs remain the Western’s primary end market for products. We continue to see significant secular headwinds for HDD demand in PCs, including an ongoing decline in annual PC unit sales and SSDs replacing HDDs. With nearly 50% of all laptop and other mobile PCs using SSD storage as of 2017, Gartner expects that attach rate to increase to 96% by the end of our explicit forecast. Likewise, total PCs with SSDs, including both desktop and mobile, is expected to increase to 93% by 2022. In 2018, almost half of HDD units sold by Western come from desktop and notebook sales, and thus the importance of a strong SSD portfolio to capture that shift cannot be understated.

While we are generally positive on the company’s purchase of SanDisk, which in principle has the company better positioned to deal with the technological shift, Western Digital has been unable to translate its dominance in enterprise HDD storage to SSD strength thus far. The company has gained share in the PC SSD market, expanding to 14% at the end of calendar 2017, but enterprise SSD has not followed as strongly and has in fact declined to 12% from 15% in the prior year. The company has sought to buttress its enterprise capabilities through the acquisition of Tegile, a leader in enterprise flash storage systems. Competitor Seagate has been severely handicapped in its ability to benefit from the shift to SSDs due its previous lack of a captive source of NAND flash capacity to drive its SSD business. This has changed recently with Seagate completing an agreement with Toshiba Memory Corp. Western--through its acquisition of SanDisk--also uses TMC to source its NAND through a collection of joint ventures. The two companies also have considerable intangible assets related to the design and development of NAND, and this has continued, albeit with significant negotiation after Toshiba’s flash business was purchased by Bain.

Beyond SSDs, SanDisk sells NAND into mobile devices with increasing storage capacities driving growth. However, flash storage technology becomes commoditized quite quickly, with well-capitalized competitors such as Samsung pushing the technological boundaries, particularly on the component side for mobile devices. As a result, transient product differentiation hampers long-term profitability for all parties involved. Meanwhile, the delicate balance between supply and demand remains a concern, with bouts of oversupply crippling margins and undersupply leading to overexpansion. We think the consolidation that has occurred in recent years should encourage the remaining companies to act rationally in terms of future capacity additions. Nevertheless, it only takes one additional fabrication plant to tip the market scales toward oversupply. Furthermore, the SSD space has more players, including Samsung and Intel INTC. We anticipate stronger competition in SSDs will also hamper Western Digital’s returns on invested capital.

While Western Digital has intangible assets (in the form of design expertise) and a cost advantage in HDD production, we expect these advantages to deteriorate as HDDs face rapid substitution risk from SSDs and the price/performance gap between HDDs and SSDs narrows. The acquisition of SanDisk has assuaged some concern, as the combined company will focus more on the growing SSD market in lieu of HDDs. However, Western Digital is still struggling to make its SSD expertise count, thus far losing share in enterprise SSD. We don’t see significant design or manufacturing expertise in SSDs that would warrant a stable moat trend for Western Digital. We remain skeptical that the SSD business will lead to excess returns on invested capital for the combined company, as we view memory suppliers as no-moat entities due to the highly cyclical nature of the storage market and the commoditized nature of the memory industry (for example, we viewed pre-acquisition SanDisk as a no-moat company). Despite ongoing investments in technology and manufacturing, we believe that remains the case. Thus, we do not believe Western Digital has stabilized its competitive positioning.

Inherent Cyclicality of Chip Market Is Risk We believe the semiconductor market is inherently cyclical, with memory demonstrating similar supply and demand dynamics as the broader industry. Also, while Western Digital and Seagate typically act rationally as a duopoly in the declining HDD market, NAND flash memory is more commoditized. As a result of these issues, as well as the broader technological trends affecting the storage market, we assign a very high uncertainty rating to Western Digital.

The ongoing supply and demand imbalances in flash have created a challenging pricing environment for major players. With Samsung, Intel, Micron Technology MU, and SK Hynix there is also the risk that competitors will beat Western Digital to producing novel silicon products (such as Intel and Micron’s 3D XPoint) that further curb the company’s ability to benefit from flash. Most important, despite the addition of SanDisk, Western Digital still relies on HDDs for the majority of revenue. The attach rate for SSDs in PCs is increasing at a steady rate while HDDs are in secular decline. Enterprise storage provides a significant opportunity for the company in the near term, with business-critical storage demands increasing, but should energy efficiency concerns increase for data centers and hyperscale cloud players, this opportunity may quickly evaporate. Should Western Digital be unable to successfully execute on its flash and SSD portfolio, or if it is unable to eke out improvements in HDD technology, it will be unlikely to generate profitable growth.

We believe Western Digital is in decent financial health. While debt increased to facilitate the SanDisk acquisition, the company has been working to steadily pay it down. Leverage is higher than usual, but we forecast that the company’s cash flow generation will more than meet the debt obligations it currently has. We expect management to continue to deleverage, but we also expect capital expenditures to stay elevated for the duration of our forecast. The demands of flash technology require the company to invest more in equipment, especially in pursuit the newest and best in 3D NAND.

In general, we believe Western Digital has a sensible and diligent capital-allocation strategy. The company has an ongoing share-repurchase program as well as a quarterly dividend to shareholders.