January 10, 2022
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 3% share of the DRAM market, a distant fifth, behind Samsung, Hynix, and Micron. The top three companies control a combined 94% 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 2022 Performance
Following the 39.9% sequential decline in revenue last quarter, sales in Q4 fell another 27.8% compared to Q3. This is a 62.8% drop year-over-year. What a brutal business. Two thirds of revenue gone in twelve months. Lower pricing and lower bit demand are behind this trend. Bit shipments declined low-single digits % sequentially in Q4, negative but much better than the low-twenties % bit shipment decline in Q3. For the last four quarter, in order, sequential bit shipments declined 1%, 9.6%, low-twenties %, and low-single digits %. The ASP declines, on the other hand, accelerated in the fourth quarter. ASPs decreased mid-twenties % sequentially in Q4, higher than the low-twenties % sequential decline in Q3. The previous two quarter ASP declines were both mid-single-digits percent. Year over year, bit shipments decreased low-thirties % and ASPs dropped low-fifties %. Low fifties percent! Gross margins were 32.6% in Q3 and 13.0% in Q4. Operating margin was 8.3% in Q3 and -19.4% in Q4. Two quarters ago, operating margin was 29.8% and gross margin was 44.1%. That is a staggering drop in six months. The next lowest quarter for gross margin that Nanya has seen in the last five years was 22%. The lowest operating margin over that time was 9%, excluding Q3-22’s 8.3%. Margins will go lower from here though it is hard to envision a typical downturn where we have four more quarters of declines. The beginning of this downturn isn’t as clear as the previous one. COVID disrupted the shape of this cycle. I think a better view is to see calendar Q4-21 as the peak, putting us four quarters into this down-cycle. As they are typically six to seven quarters in length, we have two to three more quarters to the bottom. That means the cyclical trough is Summer to early Fall of 2023.
Bit Shipments and Capital Expenditures
As we saw with Micron, Nanya further lowered their current and future capital expenditures in this quarter compared with what they said ninety days prior. 2022 actual WFE CapEx was down by nearly half. They forecasted this would be down 40% in the Q3 call. The company further lowered 2023 total CapEx, to NT$18.5B from “no more than” NT$22B previously. Within this total, less than 50% of the spending will be on fab equipment. Eyeing the bar chart provided by the company, 2023 WFE spending will be about a third lower than 2022. 2022 year-over-year bit shipments were down twenties % because of weak demand, which “may continue” into the first half of 2023. Nanya is reducing their output by “up to” 20% in 2023 but will respond to market conditions. Nanya lost market share for certain in 2021 and will likely lose more between 2022 and 2023. The company is sub-scale and can’t migrate process technologies forward at a rate that allows them to keep up with Samsung, Hynix, and Micron.
2023 Market Outlook
Nanya believes the DRAM market will remain soft in the first quarter of 2023 for the same list of reasons given for most of 2022; lower corporate and consumer spending, slowing economic growth because of high inflation and the interest rate response, the war in Ukraine, and COVID in China. That last one needs updating. The company said about the second half of calendar 2023 that “potentially, DRAM market may gradually recover from downturn.” This is the same as saying nothing. Two quarters ago, Nanya’s management characterized the decline in DRAM pricing as a “short-term market correction.” Now it is the steepest DRAM downturn in thirteen years. Nobody knows anything.
- Server Market: The slowing economy is reducing consumer cloud vendor demand though enterprise cloud demand remains strong. Weak memory demand is like a virus. Eventually it spreads and infects every sector. The datacenter and automotive sectors are the last segments to catch the downturn virus in this cycle, but they will get sick. It is only a matter of time. No mention of prolonged datacenter construction like they said last quarter.
- Mobile Market: U.S. and Korea mobile phone shipments relatively healthy but China remains slow
- PC Market: Annual PC unit shipments continue to decline though the company didn’t quantify this
- Consumer Market: Network, industrial and automotive demand relatively healthy (see my comments above about the downturn virus). Consumer electronics sales are still slow.
Analyst Call
Prepared Remarks
- Market demand was much weaker in the fourth quarter than the company expected
- This is their first quarterly loss in “many years”
Analyst Q&A
- The steepness in price declines will slow in Q1 and hopefully turn positive in Q2. Nobody knows anything.
- The CEO doesn’t believe DDR5 adoption into datacenters will be steep. He believes it will be a gradual transition from DDR4 to DDR5.
- Newly announced restrictions on technology exports to China will slow down their DRAM aspirations but won’t end them
- They are targeting 10% to 15% of their output to be on 10nm class process technology by the end of 2023
- There were long pauses between questions, something I have not seen before. They may have been having technical issues, or more analysts chose the webcast channel to ask their questions.
*Nanya’s inventory is at the highest level in more than a decade. They have already begun to lower output in response. - SG&A and R&D will be within single digit percentage level changes in 2023 compared to 2022 spending
Summary
None of the executives at memory companies know much about the future, like the rest of us. They have better market intelligence than most, but it is still a difficult prediction to make with consistent accuracy. However, it hit me in this call that Nanya is more clueless than the rest. Two quarters ago, Nanya saw this as a short-term market correction. Now it is clearly the worst downturn in more than a decade. During their Q3 call they also said that quarter would be the low point for bit shipments, which was way wrong. They seem to have accepted the reality of this downturn and shed their optimism. They reversed course on previous statements about not lowering product and about increasing bit output in 2024. Now they are cutting output up to 20% and have reduced capital expenditures in 2023, following reductions in 2022 CapEx to levels well below plan. Something that has nagged at me since early in 2021, that I didn’t listen to, is the effect of China’s zero COVID policy on memory demand. China being locked down, effectively pushing out the start of the Pandemic for that whole nation by three years, must have affected memory demand. I think it was a significant factor in how sharply demand has fallen for memory. If that is true, China opening back up in 2023 will have the reverse effect and should help the recovery. Nothing in this release changes my view that the earliest I see the memory market bottoming is June of this year. Hynix and Samsung report later this month. Those are important prints for their commentary on production and CapEx plans.
- Smooth H (cyclical long MU)