Q1 FY-24 Earnings Release and Analyst Call

Stocks A to Z
1

12/20/22

Press Release and Investor Presentation Commentary

Here are the points from the press release, outside the financial statements, that I found notable:

  • Business fundamentals are expected to improve throughout 2024, with record industry TAM projected for calendar 2025. This wording doesn’t project Micron will have record revenue in 2025, but it implies that it will.
  • Revenue, gross margin, and EPS were above the high end of the guidance ranges provided in the last earnings call. They did an investor even a month ago with updated guidance, so citing back to last quarter in the lead bullet is a bit of sales spin on their results, which is typical of this executive team. They are highly concerned with appearances with investors.
  • Pricing is still “far from levels associated with necessary ROI [for capital investment in new capacity.]” They intend to stay “very disciplined” with their supply. This is a signal to the Market and to other memory makers that pricing needs to go up much more from here before they add production capacity.
  • The “vast majority” of their bits are on leading edge nodes: 1-alpha and 1-beta for DRAM and 176-layer and 232-layer in NAND.
  • Fiscal 2024 front end cost reductions will track with their long-term expectations. Those are mid-to-high single digits for DRAM and low teens in NAND.
  • Volume production on 1-gamma DRAM using EUV will happen in calendar 2025.
  • Inventory of DRAM and NAND are at or near normal levels for “most customers” across PC, mobile, auto, and industrial segments.
  • Inventory at data center customers for both memory types is improving and is expected to reach normal levels sometime in the first half of calendar 2024. This is a negative revision from last quarter when the company said DC inventories would “likely normalize in early calendar 2024.”
  • They are ahead of the industry in the transition from DDR4 to DDR5 and expect to cross over to higher DDR5 volume in early calendar 2024.
  • Server unit shipments in 2024 are expected to rise mid-single digit percentage in calendar 2024, following low double-digit declines in calendar 2023.
  • The company expects to generate several hundred dollars in HBM revenue in fiscal 2024. The company is a distant third to Hynix and Samsung in this AI market. They expect their HBM market share will match their overall DRAM market share “sometime in calendar 2025.”
  • PC unit volumes will grow low to mid-single digit percentage in calendar 2024, following two years of double-digit percentage PC unit declines.
  • The smartphone market is “showing signs of recovery” and the company believes unit shipments will grow “modestly” in calendar 2024. This is a down-revision from last quarter. The company is saying smartphone demand remains tepid and unit shipments in 2024 may well be flat compared to 2023. The mobile market is recovering more slowly than anticipated.
  • The company expects 2023 DRAM bit demand to grow high-single digits percentage range, up from prior expectations for mid-single-digit growth.
  • Their view of NAND bit demand growth in calendar 2023 to be in the high-teens percentage range.
  • For calendar 2024, they forecast DRAM bit demand growth to be near the long-term CAGR of mid-teens percentage. This is a down-revision in demand from last quarter, when they believed calendar 2024 DRAM bit demand growth would be above mid-teens percentage. In one of the analyst calls, the management team clarified that the reason for this is that they have raised their forecast for calendar 2023 bit demand upward but haven’t raised 2024 bit demand by as much, hence the delta shrunk between forecasts.
  • In calendar 2024, they believe NAND bit demand growth will be “somewhat below” the long-term CAGR of low-20s percentage range. This is also a down-revision from last quarter when the company believe 2024 NAND bit demand growth would be around 20%.
  • Micron’s bit supply growth in fiscal 2024 will be “well below demand growth for both DRAM and NAND.” In their view, industry supply growth will be below demand for both DRAM and NAND, resulting in a contraction of industry inventory levels.
  • As fiscal year 2024 progresses, Micron’s fab underutilization will gradually transition to structurally lower wafer capacity at higher utilization rates, as technology node transitions progress.
  • Micron’s FY-24 capex will be between $7.5B and $8.0B, slightly higher than last year’s levels and an increase from prior plans, primarily to support the HBM3E production ramp.
  • They expect fiscal 2024 WFE capex to be down year over year. Hopefully Micron, and the other two large DRAM players, stick with this plan.
  • They have proceeded with the expansion of their assembly and test facility in Xi’an, China. In my view, this is an indication of weakness from CEO Mehrotra and his executive team. The CCP has punished Micron with serious market access restrictions. Instead of responding by reducing investment in China, perhaps by moving out of their Xi’an facility or selling it, they reward the Chinese government with more investment and employment there. This is a poor strategy from an executive team acting out of fear.
  • DRAM was 73% of total revenue in the quarter. Bit shipments rose in the low-20s percentage quarter-over-quarter. ASPs increase in the low-single digit percentage range. Last quarter, DRAM ASPs were down high-single digits percentage, so DRAM prices are down mid-to-high single digits over the last half year. For real recovery in Micron’s financial performance, DRAM prices need to rise by several tens of percent. The jump in gross margin and earnings this quarter benefitted from selling written-down inventory.
  • NAND was 26% of total revenue in the quarter. Bit shipments decrease in the mid-teens percentage range quarter-over-quarter. NAND ASPs increased approximately 20%. Last quarter, ASPs declined in the mid-teens percentage range. This means NAND ASPs are up a percent or two over the last six months.
  • Share buybacks continue to be suspended.
  • GAAP guidance for FQ2-24 is revenue of $5.30B +/-$200M, gross margin of 12.0% +/- 1.5%, operating expenses of $1.07B +/-$15M, and diluted EPS of ($0.45) +/- $0.07.

Financial Statements

Statements of Operations

Revenue increased sequentially by almost 18%, rising from $4.01B to $4.73B. GAAP gross margin was (0.7%) compared to (11%) in the prior quarter. These two quarters look better than reality because their COGS was reduced by prior inventory write-downs. Taking those inventory adjustments back out, gross margin in FQ4-23 was (24.9%) and it was (13.5%) in the first quarter. I have records going back to fiscal year 2011, and I there was not a quarter with a GAAP gross margin below 10% in at least twelve years. This has been an historically bad downturn. It should be noted that the market is far from healthy for memory makers. Gross margin peaked at 47.2% in the last upturn (2022) and 61.0% two cycles ago (2018.) ASPs need to more than double from here for the financials of Micron and the other memory IDMs to return to the levels of health seen in past cycles. Operating expenses rose around 20% from the prior quarter. This happened from a combination of higher employee compensation and more pre-qualification wafers. The company posted an operating loss of $1.13B in the quarter. This is not as terrible as the $1.47B operating loss in the prior quarter. Both would be almost a half a billion dollars worse were it not for the lower costs from inventory write-downs. Interest income and interest expense exactly balanced each other this quarter. There was a non-operating expense of $27M in the quarter, from marking down the value of a non-marketable investment. Even with all these loses, the company provisioned $73M for income taxes, from differences of earnings around the world. The company’s net loss for the quarter was $1.23B, or ($1.12) per share. This would be a $1.8B loss without the prior inventory write-downs. Stock price is about the future, which is why Micron’s has been rising, but the present for the company is really terrible. With share buy-backs suspended, share count rose 10 million shares, to 1,100 million.

Cash Flows

Cash flow for the first quarter was negative by $510M. Operation cash flow was positive by $1,401M but investing activities – the biggest being capital expenditures at $1,796M – consumed all of this, plus an additional $157M. Depreciation and amortization were more than CapEx by about 5%. That is a real measure showing Micron is effectively reducing capacity. At baseline, their business requires higher capital investment every period than is used up in deprecation because process technology is always migrating forward, and bit output is (almost) always increasing. Finished goods inventory declined by $334M during the quarter. That is about a 20% reduction. Cash flow for financing activities was negative $352M. A third of this cash flow was to pay dividends, and another third was “other.” Free cash flow in the first fiscal quarter was negative $395M. The cash flow statement in this quarter is about as clean a look at the current performance of Micron as you will see, with little movement of debt, no share buy-backs, and low capital expenditures. If the memory market cooperates with rising ASPs in the coming year plus and management makes good on their statements about restrained capital expenditures, free cash flow should swing positive in the first half of calendar 2024. I think the market will improve faster than is being forecasted, so my prediction is Micron will be free cash flow positive in FQ3-24.

The Balance Sheet

Micron’s book value continued to decline in the first quarter of fiscal 2024, falling to $42.9B at the end of the period, from $44.1B at the start. Current assets came down by a small amount, around 1%. Within this, cash declined by $500M and total inventory was down $100M. As I noted above in the cash flow section, finished goods inventory was reduced by more. This was partially offset by a rise in WIP, which is an indication the company has increased factory utilization some during the quarter. This doesn’t necessarily mean they increased the number of wafers in the fabs. While this may be part of the rise, more of it likely comes from a richer mix of more advanced process technology wafers, which are more valuable per unit. All the action in the assets section was in the current assets. Current debt rose by around $600M, mostly the effect of long-term debt sliding into the next twelve-month period. Total liabilities didn’t change much during the quarter. As we saw in the cash flow statement, we are in a quiet period for Micron’s financials. The fourth quarter of FY23 was the bottom of the cycle and the memory market is on the way up.

Conference Call

Numbers given during the conference call are non-GAAP.

Sanjay Mehrotra (President and CEO)

  • · They have raised their projections for memory ASPs compared to prior plans, though pricing is still “far from levels associated with necessary ROI [for capacity investments].”
  • · In the data center segment, total server unit shipments are expected to increase by a mid-single-digit percentage in calendar 2024, following a year of low-double-digit decline in 2023.
  • · HBM requires 2x the wafer supply as DDR5 to produce an equivalent number of bits.

Mark Murphy (CFO)

  • · Mobile customer inventories have normalized and mobile segment revenue in Q1 nearly doubled from year-ago levels (which were terrible.)
  • · Embedded Business unit revenue was up $1B sequentially, driven by strong growth across “most” end markets.
  • · Revenue in the Storage Business Unit was down 12% sequentially because of lower consumer component sales (NAND sold as components vs. in a Micron-constructed drive), partially offset by strong SSD sales. Their mix shifted to more drives and fewer loose packaged die.
  • · Their $59M tax provision is based on their improved outlook for fiscal 2024.
  • · Inventory at the end of fiscal Q1 was $8.3B or 159 days, down from 170 days in the prior quarter. I think “in the prior quarter” means at the end of the prior quarter. Current inventory of 142 days (excluding strategic stock) is 22 days above their target inventory level.
  • · Their leading-edge supply of both DRAM and NAND are oversubscribed for the full year 2024.
  • · The company is seeking to minimize customer demand pull-in that is happening in response to higher pricing. In other words, they are purposely underselling demand to support pricing. Customers response to pending shortages by stockpiling memory at lower prices. Micron is smartly responding to this proactively. For this reason, both DRAM and NAND bit shipments will decline in the second fiscal quarter.
  • · Gross margin in the second quarter will be better because of higher prices and lower charges from underutilization.
  • · The balance of previously written-down inventories are expected to clear in the second fiscal quarter. This means that Q2 gross margin will be artificially high (not representative of Micron’s true COGS,) but that Q3 will show the true cost picture.
  • · Full year operating expenses for FY-24 will be “over” $3.9B. The company believes they will have positive operating income in the third fiscal quarter of this year. They have increased their tax expense from under $200M to over $300M, based on higher taxable income.
  • · Q2 capital expenditures will be “in line” with Q1 levels ($1.796B).
  • · Free cash flow will be positive in the fourth fiscal quarter (June to August 2024, inclusive.)

Mehrotra and Murphy spent 32:50 in this call on prepared remarks. This is still way too much (write an investor letter) but is an improvement from the 40 minutes (two-thirds of the total call!) last quarter they used on prepared remarks.

Analyst Q&A

  • · Their HBM DRAM supply in calendar 2024 is sold out.
  • · Mehrotra said again, while answering the second question, that memory pricing is still “far below” the level needed for the company to expand production capacity.
  • · From Q4-23 to Q1-24, about half of the gross margin increase of 1000 bps was better pricing with most of the rest being mix. The rest was lower inventory cost from write downs and lower underutilization charges.
  • · From Q1-24 to Q2-24, the ~1300 bps rise in gross margin will be mostly from better pricing. The balance is lower cost from written-down inventory and reduced idle capacity charges.
  • · They expect margin expansion in each quarter throughout fiscal year 2024.
  • · Asked directly of the through-cycle model they gave at their investor day in 2022, if they are sticking by their 30% operating margin and 10% gross margin, Mehrotra said it applies after this downturn. He said their through-cycle model is the standard they are holding themselves to, except this cycle doesn’t count, because it was especially bad.
  • · When asked how the company plans to limit strategic buying without limiting market share loss, Mehrotra made some comments, none of which resembled an answer.
  • · When asked if HBM gross margin will be pulled down to the rest of the DRAM industry average margin, Mehrotra said, indirectly and in a long-winded way, that demand for HBM will be much higher (3x higher) than the rest of the industry. In other words, growth will be hard for supply to catch up with (they hope.)
  • · Mehrotra thanked the first analyst for his question, but not after that. It’s as if he got feedback to thank analysts for their questions, then went back to his true nature after the first question.
  • · I usually feel better after listening to Micron’s analyst calls. I felt worse after this one. Combination of lack of information and the contradiction that the memory market has “healthy dynamics” and their leading edge supply is oversubscribed, and the statement that their current pricing is far from the levels necessary for capital investment in capacity.

Post Earnings Q&A call with Mark Murphy (CFO), Sumit Sadana (CBO) and Manish Bhatia (VP of Global Operations)

  • · Their push-out of data center inventory normalization is because they misjudged the amount of inventory at their large data center customers. They didn’t say that directly, but that’s what they said.
  • · Bhatia thanked the analyst for a question directed to him, a nice contrast to his boss.
  • · Their view of calendar year 2024 DRAM bit demand didn’t change much from a quarter ago, but their view of 2023 demand went up some. This is the reason for the change in year-over-year percent DRAM bit growth in the prepared remarks.
  • · Their DRAM bit output bottomed in FQ1-23 and revenue bottomed in FQ2-23.
  • · DRAM and NAND bit shipments will both be down in FQ2. DRAM will likely also be down in bit shipments in FQ3. NAND is not clear at this point for FQ3. Pricing will drive all the changes in revenue and margin performance for at least the second and third quarters.
  • · DRAM gross margin is “definitely higher” than NAND. This continues a trend that has been evident for years. NAND has declined more sharply than DRAM has in this downturn, thus it will need to see larger percent increases in pricing to make capacity investments make sense.
  • · At one point, their total wafer output was down nearly 30% from peak. Bhatia declined to be more specific. The analyst asked for wafer output numbers. This is typically not something any memory companies disclosed, so I’m not surprised they didn’t give them here.
  • · “Sometime in fiscal year 2025” is as detailed as they would get when asked when their HBM market share will match their overall DRAM bit market share.
  • · The last underutilization charges will come from Micron’s oldest nodes.
  • · Bit growth will be flat in Q2 vs. Q1, so all the sequential revenue increase between these two quarters is in pricing.
  • · Vivek Arya asked why the executive are confident that prices will rise for 2+ quarters, given the industry has a poor record of predicting pricing more than one quarter out. Sadana gave an important answer to understanding the cycle. The first part is that, from the point that a memory company decides to add capacity, it is 14 months before output is impacted, for example. That came from a five-month total cycle time for manufacturing (fab and assembly-test) plus a nine-month tool lead time, including installation and qualification. On the demand side, customer segments are heterogeneous and content density is steadily rising. Inventory is not fully healthy yet, and as that happens, demand will rise. Also, their internal models for unit growth in memory-containing devices are conservative. Finally, the downturn has been historically deep, thus memory makers have a high threshold before they increase capital investment.

Summary

The executives made no mention of the China market access restrictions, nor did any analyst ask about them. To me, that means the company has mostly adjusted their shipments to customers away from China. The overall market will have more bits from Samsung and Hynix going to China and fewer from Micron. It took a couple of quarters, but the global memory market is fluid in a similar way to the oil industry. Their guidance is for gross margin to be positive 13 in FQ2, a sharp rise in sentiment from last quarter. As I said previously, the upturn typically comes on faster than is widely believed at the time. There was no mention of the rate of pricing increase they are forecasting. I don’t think this means anything; it is just a quarter-to-quarter change I noticed. Micron’s stock price has risen sharply in the last few weeks. This is in part from the emerging evidence that the memory upturn is underway. It is also from the overall market rising to levels not seen in more than twenty years, as measured by how low the equity risk premium is. I expect to see memory pricing rise for at least another fifteen months from here. As always, the question is how quickly.

-Smooth Hughes (cyclical long MU)

4 Likes