Looking into Micron Technology (MU)

Having looked into Micron it may be a company to invest in or have a watchlist. They are a memory maker who owns the fab and silicon producing process, along with the final product development and distribution.

They’ve put up some strong numbers recently and have a compelling narrative. Ultimately I decided to pass on starting a position, but I’ve put them on my watchlist. I’d be interested to hear what other board members think of Micron.

They product mix increasingly includes higher margin memory products which are slated to go into Nvidia’s AI chips. Some of these products are already sold out through 2025, and the company like many other semis is looking to ramp production adding to a facility in Idaho and creating a new fab in New York state.

There are new products like AI PCs and AI phones which Micron is going to be a top supplier for. I’m still somewhat unclear what an AI PC or AI phone even means yet, and I think maybe even Micron is too, but the hardware providers are building these, and they require significant more memory. This is looking to be a huge catalyst for Micron going forward.

The reason I’m passing on investing is that their quarterly run-rate for the last quarter is only on par with earnings from years like 2018 where they had 30.4B in revenue, and 2022 where they had 30.8B in revenue. They have explained that 2023 was a bad year for the semis and especially storage and memory. I’d really like to see this company get past their previously yearly revenue range.

They also have a 6.1B deal with the government for the CHIPS act which reimburses for the fabs build out in Idaho and New York. It sounds like the government won’t reimburse later until the factories are built. I’m not sure the impact if this regulation ends and if it only makes sense for them to build the fab because of the regulation.

DRAM which is short term memory (69% of Micron’s sales) storage has three main players,
Micron 22.8%
Samsung 38.9%
SK Hynix 34.3% (Also South Korean company)

If only looking at the last year, the trend looks impressive,

4.01B → 4.73B → 5.82B → 6.81B
-39% → 16% → 57% → 81% (yoy growth)

476M → 787M → 2115M → 2636M

Gross margin
-11% → -1% → 19% → 27%

Their last earnings guides and actuals,
Revenue 6.6B → actual 6.8B
Gross margin 25.5% → actual 26.9%
Diluted EPS 0.17 → actual 0.30

Next quarter guide is,
Revenue 7.6B
Gross margin 33.5%
Diluted EPS 0.60

Some standout items I like from their last earnings and presentations,

  • Gaining share in high margin products like HBM (High bandwidth memory)
  • Operating cash flow of 2.48B, last Q was 1.22B, and previous year was 24M
  • Robust AI demand
  • 2025 demand for AI PC and AI smartphones
  • Gaining share in data center SSDs
  • During quarter tripled bit shipments of 232-layer based 30TB SSDs (they measure volume in “bits”)
  • AI PCs have 40-80% more DRAM than typical CPUs
  • Automotive is robust and growing well
  • Storage revenue (SBU) is +50% qoq, and +116% yoy and this quarter was 1.35B
  • Revenue breakdown is 69% DRAM, 30% NAND (NAND is faster growing)
  • AI products are higher margin: HBM, high capacity DIMMs, data center SSDs
  • Seeing increased interest from many customers across market segments
  • HMB3E (latest product), has 30% lower power consumption than competitors (Keep in mind low power consumption is a top priority for Nvidia)
  • HBM shipment began to ramp in this Q, with margins accretive to DRAM and overall company margins, expect 700M in Q4 revenue and multiple billions next year
  • HBM sold old through 2024 and 2025 which prices contracted from majority of 2025 supply
  • Hyperscale demand is improving driven by AI training and inference
  • PC replacement cycle should gain momentum through calendar 2025
  • AI PCs align well with portfolio
  • AI phone releases will use 50-100% more memory
  • Mobile DRAM and NAND solutions are now widely adopted in flagship smartphones
  • Some portion of DRAM and NAND sales growth are from a 20% price increase
  • Gross margin would have grown by eight percentage points more if not for written down inventories
  • OpEx at low end of guide, discipline and operational efficiencies
  • Expect continued gross margin expansion, pricing trends
  • CHIPS act offsets CapEx costs from fab investments
  • Lots of worldwide facilities, Idaho, Virginia, Utah, Singapore, Japan, Taiwan
  • HBM3E has qualified to be used in Nvidia H200
  • Revenue is 50% domestic, 50% international. Breakdown is USA 50%, Taiwan 17.4%, China 16.2%, Japan 6.4%, other 9.8% (Shows how all the innovation is happening in the USA and Asia)

Some points about the business that I see as drawbacks,

  • PC and mobile product lines are growing low single digits
  • 6.1B deal on CHIPS act which could end for their fab build out
  • Silicon wafer supply could be a blocker, HBM3E consumes 3x the wafer supply of previous products, and HBM4 will consume even more

Overall I’d prefer to have my money invested in Nvidia and Supermicro, but Micron is certainly interesting as a growth stock. I’d like to observe at least another quarter or two to see what type of numbers they can put up. Ultimately I’d rather be in the top tier, fastest growing AI players as opposed to ones which might be the best second tier like Micron and Celestica (CLS).


I took a 2% starter position in MU last week. MU indicated strong demand for its High bandwidth memory (HBM) for the next 18 months. My impression is that they are basically sold out. HBM is critical for AI. The fundamentals look good, and their valuation is reasonable. I felt it was worth a starter position with an aggressive stop loss to act as a hedge as I learned more.


I have been following MU and the memory industry for a long time (4 memory cycles). Each one has been quite a rollercoaster. The latest downturn in 2022/2023 was especially painful for the 3 major memory producers (Samsung/SK Hynix/Micron) due to simultaneous downturns in both the PC & Mobile end-markets. Samsung management foolishly continued to flood the market with DRAM bit supply well into 2023 causing all 3 to lose a lot of money. Since Samsung announced DRAM production cuts last year and the PC & mobile sectors stabilized there has been rapid increases in DRAM/NAND ASP’s.

I considered bringing MU to the board late last year, but decided against it since it is not a traditional Saul stock. Micron has been an extremely cyclical company for a long time that sells commodity memory parts. This is not a bad thing if the DRAM cartel (oligopoly) plays nicely and doesn’t try to get into market share battles like Samsung was a year ago. I started a large position around then after Samsung announced it was finally cutting Dram production and it was clear ASP’s would be recovering. Also with Micron announcing it would be ramping HBM3e production this year for Nvidia’s H200 GPU’s. It has grown to my largest position at this time (~21%) and best performer in my PF YTD.

As you can see from the numbers Wpr posted the DRAM/NAND ASP recovery has been rapid causing explosive YoY revenue/EPS compares. I remain invested in MU because of (1) high growth expectations for HBM due to AI applications & (2) HBM wafer requirements keeping DRAM/NAND from reaching an oversupply situation. The following drawback @wpr101 mentioned is actually a positive for Micron/Sammy/SKH in my opinion since it will keep the supply of the more commodity like DRAM/NAND memories in check.

HBM is basically vertically stacked DRAM parts that allow for higher bandwidth for AI applications. It is very new memory & hard to manufacture so the yields are likely low right now across all the 3 members of the oligopoly. They will all need to put a lot of resources into ramping this and this product is not a commodity right now. Micron management has stated they have booked out their HBM supply through next year. The margins are better than they are on DRAM/NAND. Samsung is currently not even able to get their HBM qualified on Nvidia’s GPU’s due to heat and power consumption issues.

[News] Reasons for Samsung’s HBM Chips Failing Nvidia Tests Revealed, Reportedly Due to Heat and Power Consumption Issues | TrendForce Insights.

The HBM market is expected to grow at a ~50% CAGR for several years so there could be a big opportunity here. Micron management stated they expect multiple billions of HBM revenue in their FY25 (starting this Sept.) vs. a couple hundred millions HBM revenue for current FY24.

If you do buy this is not one to set and forget. There is likely to be some point where the memory cycle again peaks out and revenue may fall some at that point. But due to the above reasons and the demand behind the AI buildout I think that point will be at least 1-2 years out. I think there is a good chance MU can earn $12-15 EPS in their FY25 (Sept start). Who knows after that… but if that happens I expect the stock to be above $200 with potential for higher if FY26 looks strong.



Yeh I’ve held Micron before and is on my watch list and will will certainly will be swept along in the AI tornado.

The only issue with waiting a few quarters is that Micron and memory chips are immensely cyclical.

Entering Micron now is maybe in time to ride the chip cycle but the share price typically leads the cycle in anticipation by a quarter or two so effectively we are a little late at this stage. Waiting 2 more quarters and you maybe entering approaching peak cycle ahead of the next roll over unless you believe they will experience an Nvidia AI driven supercycle.