April 10, 2025
About the Company
Nanya is a small Taiwanese DRAM manufacturer. They produce DRAM on trailing-edge process nodes (licensed from Micron) but are developing their own more advanced, technology. Nanya has about a 1.6% share of the DRAM market, a distant fourth, behind Samsung, Hynix, and Micron. The top three companies control a combined 96.4% of the DRAM market.
I don’t own any shares of Nanya and would not buy any, because they lack the scale and the technology to be competitive in DRAM. I keep an eye on their quarterly releases as a data point on the DRAM market.
Market Commentary and Q4 2024 Performance
Nine quarters have passed since Nanya stopped including the time trend of their gross margin with their earnings. I warned last quarters that if the weakness in the non-AI DRAM market proves not to be temporary that Nanya’s existence would be threatened. This quarter’s results are a reprieve for the company. Following ASPs down low-teens % last quarter, the decline in pricing slowed sharply. For the first quarter of 2025, Nanya’s ASPs were down just low-single digits percent. Also, the trend in bit shipments reversed, from down high-single digits % last quarter to up high-single digits % this quarter. For the last thirteen quarters, Nanya’s ASPs have been down low-single digits %, down low-teens %, up mid-single digits %, up low-teens %, up high-single digits %, up low-single digits %, down high-single digits %, down mid-single digits %, down mid-single digits %, down high-single digits %, down mid-20s%, down low-20s%, down mid-single digits %, down mid-single digits %. Still, Nanya is feeling the pain. While revenue was up 9.3% sequentially on the strength of higher bit shipments, gross margin eroded further, from (10.6)% last quarter to (15.0)% this quarter. Operating margin was (43.9)% in the quarter and net margin was (27.0)%. More than a quarter of every dollar of Nanya’s sales is lost at the bottom line. The company needs pricing to continue improving. Otherwise, this slowing in price declines will just be a stay of execution.
Bit Shipments and Capital Expenditures
I made a bet last quarter on Nanya’s 2025 capital expenditures. They forecasted NT$20B, pending board approval. I took the under. The company’s board approved NT$19.6B. This is a level of CapEx approaching what the company spent in 2022, which was a sum of NT$20.7B. What is notable about their investment and production forecast is the company plans to increase bit shipments by more than 30% in 2025. That would be the first year of bit shipment growth for Nanya since 2020. In that year, Nanya’s bit shipments were up 30% sequentially. Also worth noting on this bit shipment forecast is that the “up >30%” forecast is an increase from their prior plan of up >20%. I think this means Nanya recognizes their subscale size means they have to take the risk of growing bit shipments in order to have a prayer of any semblance of financial health in the future. If they don’t grow revenue, they will be eaten by their fixed costs. They’ll probably be eaten either way, but at least their death will be slower.
DRAM Market Outlook for Q2 of 2025
There is “opportunity” for DRAM market demand improvements in 2025 because of AI demand and inventory reductions, the same story the DRAM industry has been hoping for the past six months. Nanya most of all. Per tradition, Nanya cites a risk from the global economy in this section, specifically the tariff conflicts. They also generally speculate about a shift to “compact language models” stimulating DRAM demand as some AI workload moves to edge devices. This is wishful thinking.
- · Server Market: CapEx remains strong in the second quarter across the server segment.
- · Mobile Market: Stimulus from the Chinese government has helped local demand in the short term, as had been the company’s hope.
- · PC Market: No comment on current conditions, which means they are weak. The one comment provided speculates on AI PCs “maybe” seeing higher demand.
- · Consumer Market: As with mobile, stimulus from China has helped in the short term. I wonder how much of the improvement in Nanya’s DRAM markets that was seen this quarter was from Chinese government stimulus and how much was from inventory reductions at customers bottoming out.
Analyst Call
Prepared Remarks
- · Technology migration to the 1-b node is a main reason for their higher costs in the quarter.
- · Within the 2025 CapEx budgeted spending, 30% will go to equipment (mostly for tech node migration) and 70% to cleanroom and facilities construction.
- · The company is sampling a 16Gb DDR5 part currently.
Analyst Q&A
- · Outside of tariff effects, DRAM demand is gradually improving because of inventory reductions and shifts in wafer starts to AI products. If there were a 10% tariff on DRAM agreed to in the next three months, the effect on demand would not be very much. However, if it takes longer to reach resolution, management cannot say what the effects will be.
- · The company’s new DDR5 product is mostly used in PCs at this time, and into low density servers. AI servers need high speed and high density memory, specifically HBM cubes.
- · They hope to start to see reduction in costs and increase in bit output from the transition to the 1-b node in the end of Q2 or in Q3 of this year.
- · “The pricing situation may continue to improve” even with tariff uncertainty.
- · DDR3 and DDR4 demand is being helped by Chinese economic stimulus.
- · The CEO was non-committal on what their DDR5 output goal for the year 2025 is.
- · The company is positive on the trend in improvements in their markets that is underway. Their one concern is the uncertainty in tariffs. Otherwise, spot prices are rising and inventory levels are declining, as was the hope across the DRAM industry since last fall.
- · DDR3 and DDR4 inventory levels across the industry are declining because of lower production and improving demand.
- · The company currently has some idle costs, but the are “not very big.” Maybe a few million dollars. Those costs will be lower in Q2.
- · The percent of total company revenue from server applications today is in the single digits. They hope this will improve in the beginning of 2026.
Summary
Last quarter Nanya was hoping to see a recovery in their product ASPs in the second quarter of 2025. In a surprise to me, this improvement has happened, and early at that. The decline in ASPs slowed dramatically and spot pricing has begun to improve. Along with that, demand has improved. In the quarter, Nanya saw an average decline in ASPs of low-single digits and a high-single digit improvement in bit shipments. After a Q4 that saw a drop of almost 20% in pricing, ASPs are now increasing. The reduction in inventory at mobile, PC, and consumer customers that has been forecasted by memory companies for six months seems to have come to pass. I have been skeptical this would happen, because I believed that Chinese memory companies are bringing supply on-line and causing excess bits in the market. Nanya is still in a dire financial situation. Better pricing reduces the level of alarm they should be feeling, but they are far from out of the woods. As for the overall DRAM market, Nanya’s results here – which give a pure view into the non-AI DRAM market – are encouraging. They tell a story that the bits being drawn into HBM and other leading-edge DRAM applications have indeed brought the rest of the DRAM market into a state of mild undersupply. Tariff uncertainty hangs over everything today. If that were to be resolved, the DRAM market looks to be headed to at least a couple quarters of better demand and rising prices. My earlier observation that the DRAM upturn for Nanya was over after one year looks to be incorrect.
– S. Hughes (short MU)