July 10, 2024
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 Q1 2024 Performance
It is now six quarters since Nanya stopped including the time trend of their gross margin with their earnings. As with last quarter, I was disappointed with Nanya’s results. The company’s performance this quarter indicates that the market for legacy DRAM technology continues to be soft. While prices are rising, we are not seeing the degree of undersupply in Nanya’s part of the DRAM market that I thought we would by now. Revenue in the quarter rose 4.4% with gross margin up from (2.9%) to 2.9%, an expansion of 5800 basis points. This is a slowdown in both sales and GM. Last quarter, gross margin improved 10,700 bps, from (13.6)% to (2.9)%. ASPs were up low-teens % and bit shipments decreased mid-single digits %. For the last ten quarters, going back from Q1-24 (from newest to oldest), ASP changes have been: 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 %. We are now three quarters into rising ASPs for Nanya. The rate of price increases has been accelerating. Last quarter, I made a prediction here for the rate of ASP increases in Q2, Q3, and Q4. Those forecasts were up mid-teens %, up 20%, and up high-teens %. One quarter in, I am off by 2-3%. I said mid-teens for Q2 and the actual result was low-teens %. It is encouraging that the second derivative of ASPs is increasing. It could be that the effect of HBM drawing DRAM bits out of other segments is taking longer than I thought it would. The company did take a loss of NT$657 related to the earthquake. That is $20M in USD.
Bit Shipments and Capital Expenditures
Last quarter, the company increased their capital expenditures forecast for 2024 from NT$20.0B (same as what they said at the end of Q4-23) to NT$26.0B. They have stayed at NT$26.0B in Q3. This is heavily back-half loaded as in 1H’24 they have only spent NT$5.4B. Of the “up to” NT$26B for all of 2024, WFE CapEx will be ~50% of this. Also unchanged from last quarter is the company’s forecast for their DRAM bit shipments to grow “more than twenties % YoY,” the same forecast that goes back to January. Saying “greater than 20%” gives the company room to grow faster if pricing supports it. For the quarter, bit shipments decreased mid-single digits % sequentially. In the first quarter, bit shipments increased low-single digits %. If low single digits is 2% and mid-single digits is 5%, then Nanya’s bit shipments are down about 3% from the start of the year. The back-loaded CapEx is supporting growing bits by at least 23% in the second half of the year. This must mean that much of that equipment is installed and close to running wafers, else there isn’t enough time to spend the rest of that CapEx.
2024 Market Outlook
Nanya continued their tradition of almost useless market commentary. They expect HBM and DDR5 momentum to “assist market recovery in 2H’24.” Focus on HBM by major suppliers is helping DRAM inventory adjustment. The company also stated that major suppliers resuming full production may lead to supply increase in 2025. No kidding.
- · Server Market: Increased CapEx from cloud providers and AI infrastructure is driving demand.
- · Mobile Market: They are seeing smartphone sales in China recover, same as last quarter.
- · PC Market: Traditional PC demand remains flat.
- · Consumer Market: Flat sales in Q2, expected to see increased demand in Q3.
Analyst Call
Prepared Remarks
· The company is still losing money, NT$813M in the quarter. This would have been 75% less without the earthquake.
Analyst Q&A
- · The CEO hedges when asked about pricing by citing geopolitical risk and competition. He did say that pricing is improving.
- · In the third quarter, overall, they expect pricing to increase, and seasonality will likely lead to higher bit shipments also. DDR5 is strongest, followed by DDR4 and then DDR3.
- · For DDR4, they are seeing inventory improving at major suppliers. Bits being shifted to HBM and to DDR5 are helping this inventory situation.
- · Second quarter sales volume was weak because the DRAM market, outside HBM and DDR5, is flat. The consumer is in a seasonally weak time. Mobile is also “not so good.” PC is also not so good. Mobile is 30%, PC is 10%, consumer is 20% of the market. I think they are referring to the market overall.
- · The 20% bit growth they are forecasting this year is because bit shipments last year were so bad.
- · Their new fab won’t be ready for production until 2026 or 2027. They are relying on production improvement and technology migration to grow their bits for the next two years.
- · DDR3 pricing is still seeing a lot of pressure. Improvement is in DDR4.
- · They will have their first DDR5 shipments in the second half of 2024.
- · The DDR3 market is shrinking faster than DDR4 is.
Summary
The company is still losing money. The tone of the CEO, who is biased to pessimism is cautious. He said he believes pricing will increase again in the third quarter. He was asked three or four times to be more specific on pricing and refused. Nanya is subscale and running old technology so their gross margin will lag well behind Samsung, Micron, and Hynix. It is still less than 5%. Nanya needs large price increases from here to make their business sensible. These results tell memory investors that, outside of HBM and leading edge, the DRAM market continues to be soft. Demand is flat in PC, mobile, and consumer. Bits are being drawn out of these segments and redirected to HBM and DDR5. The collision we may witness is between the high demand growth for AI and low demand growth for most everything else. The DRAM makers are rushing away from everything else and into AI. If this AI demand holds, the supply leaving everywhere else should support higher pricing in mobile, PC, and consumer, despite tepid demand growth. If we were to see a revival in one or more of those segments, prices will likely increase more strongly, creating a healthy memory market. However, if PC, mobile, and consumer memory demand remains lukewarm, and the AI growth engine slows, the DRAM market will likely soften before it gets healthy enough for Micron, Hynix, Samsung, and Nanya to make back the money they lost in 2022 and 2023, and fund investment in future technology and capacity.
-S. Hughes (cyclical long MU)