Report on Trendfocus Trip in Taiwan and Japan

This is an Executive Brief from Trendfocus, Inc.‘s Information Services published on May 25, 2024.

The Chase Is On,
but Nearline Supply Chain Wary of Recently Elevated Build Plans

Our just-concluded trip visiting storage-related companies in Taiwan and Japan both confirmed our thesis on the recovering nearline market and revealed other areas of interest.

With demand up, driven by accelerating cloud orders, HDD suppliers have thus far executed build-to-order practices well, warning customers of long lead times and little upside flexibility through the remainder of this year. As a result, pricing is not only holding, but rather upward pricing movement continues for 3CQ24 even in the hyperscale market. The dynamics have forced OEM customers to also place firm purchase orders or sign volume purchase agreements to secure limited supply, even if the actual demand in non-cloud markets remains less certain later in the year.

As suggested in our Asia trip Executive Brief published in March, the higher demand has forced the HDD supply chain to quickly ramp component capacity following a year of factory shutdowns, layoffs and cost reductions. Now, in May, component suppliers in many cases are under increasing pressure, closing supply gaps only to see them open again as demand has continued to rise. Seagate’s delay in its HAMR ramp has also resulted in some increase in non-HAMR product builds in the meantime, causing additional churn in the supply chain.

Quarterly nearline build plans through 2024 which were flat at 14 million-plus at the beginning of the year have turned solidly upward – no surprise as Trendfocus has been telling the HDD supply chain, even in late 2023, to expect growing demand this year. Latest 1CQ24 4-quarter forecast upgraded the 2024 nearline market to 939EB and 53 million units with peak quarterly shipments rising to 15 million later in 2024. However, build plans, which should not necessarily reflect actual shipment expectations, have risen over the past few months, blowing past 16 million per quarter in each of the last 2 quarters of the year to top 18 million/quarter. This has raised concerns at component suppliers who fear the builds do not reflect real demand, especially as questions swirl about the actual build capacity at HDD companies.

For perspective, nearline market shipments peaked in 3CQ21 at 19.75 million during the over-hyped pandemic demand rush. Supply chain records indicated build plans for that quarter were just over 20 million. Since the collapse of demand in 2022, HDD build capacity has been slashed with some cuts temporary and others permanent. The temporary reductions have required factories to hire back staff and re-activate assembly and test lines and then chase after component suppliers to do the same. It is no surprise that supply continues to tighten as increasing output requires time to get a long and complex component supply moving again.

Can the HDD industry build/ship 18+ million nearline units in a quarter?
Compared to the previous nearline peak, the average nearline capacity today is much higher – 17.2TB in 1CQ24, the latest reported quarter, compared to 12.9TB back in 3CQ21. So, even if all previous nearline capacity remains today and is activated, the higher product mix would strain available test capabilities thus limiting maximum output, especially since no HDD company has invested in manufacturing output since the great storage collapse of the past 18 months. Assuming lines and staff are back at full operation (almost) and component supply is unconstrained (not even close), what is the real nearline capacity going to be in 3CQ24 and 4CQ24? Trendfocus estimates that current quarter build capacity is around 16 million or slightly higher, but it seems unlikely that the actual builds to meet the 18+ million quarterly plans in 2H24 are achievable. Another question is since HDD companies have insisted they will only build what is backed by firm orders, long-term agreements or volume purchase agreements, one begins to wonder what volumes are really on the books?

Historically, Trendfocus’ opinion was that build plans were more of a tool to exercise the supply chain into action when demand was increasing or, to expect little when end markets declined or product inventories were elevated. So, the large build plan volumes currently circulating through the component supply chain could be viewed in those terms. That said, if HDD suppliers are keeping a tight lid on builds and finished inventory, the supply chain will be left holding the bag and absorbing the volatility of demand if component companies ramp up to service unrealistic build plan volumes.

Clearly, HDD companies are anticipating a new era of stability and profitability to support future technology investments critical to the long-term existence of the industry and to cost-effectively service unceasing storage demand, especially in the cloud. Recovering markets, especially driven by expected economic improvements into 2025, will go a long way in supporting the HDD industry’s newfound discipline. However, history also is littered with past growth successes that have devolved into messy declines when end-markets contract. How disciplined will HDD companies remain when obvious and unavoidable patterns such as hyperscale digestion and longer-term macroeconomic cycles occur? Is the industry able to temper volatility with its current practices and will it actively manage expectations to stabilize a critical component supply chain? It is easy for now to answer ‘yes’ to these questions as end demand rebounds over the next year or so, but the real test for the industry lies ahead.