Has anybody tried to calculate the fair value of the stock? I tried the DCF and P/FCF approaches in a couple of ways.
1) With DCF at 4% discount rate, 1% terminal growth rate and FCF following the projected price fluctuations of DRAM and nand flash (0.4 correlation each on top of latest Qs FCF) and price extrapolated linearly against current wafer demand & capacity varied by committed planned new capacity, I got a price upwards of $1550.
2) With DCF at 4% discount rate, 1% terminal growth rate and revenue dropping to about $50B annual (remember with SCAa they've locked $20B ARR against \~25% of blended capacity so remaining 75% should get them at last to $50 especially if they are targeting locking 50% capacity), 1.3B fixed opex and gross margins varying between 65%-85%, and 0 capex, I got stock price between $950-$1250.
3) With P/FCF at 32x multiple (where Hynix is today), I got price of \~$900 adjusting their off cycle quarters. This is unrealistic I'm sure as MU and SNDK will command some US premium and Hynix will likely rise after the ADR is listed so I thin 33x-34x is not far fetched.
Either way I skin the cat, it looks like this is bottom. Technicals too seemed to be strengthening today with relative volume higher, v-shaped recovery and price closing higher than open, it looks like it's time to get in on this trade.
what am I missing, anybody else tried assessing the fair value?
Edit 3: Again on discount rate and equity risk premium. Why should it be any higher than 1-2% multiplied by beta when risks are being hedged by SCAs and revenue accounted for cash flows is almost exclusively coming from SCA and guaranteed? If you want to go with traditional 10-15% Ke/WACC, build in at least some growth/upside into base case to balance! The business growth isn't ending tomorrow. Everybody agrees there's at least 1-2 year more growth to peak left if not more with next wave. Wonderful business at fair price than fair business at wonderful price.
Edit 2: MU has SCA for 33% of NAND and 20% DRAM, that's blended 25% capacity, targeting 50%. They're locking down the down-side commodity business and non commodity HBM at worst will follow chip makers. At worst it will only be as cyclical as nvda, Intel, amd, avgo, at best less than those as hyperscalers, auto and other upcoming industries with real time needs lock in on HBM. Should MU's multiples be much worse than nvda/amd/avgo??
Edit 1: A lot of people are calling me out on 4% discount rate. Discount rate is risk free rate + risk premium. My thesis is that the moment we assume revenue close to what's guaranteed under SCA ($20b annual against 25% capacity with the management targeting 50% by next quarter), and zero debt for a company this size, risk reduces substantially. Still perhaps 6% may make people more comfortable. That won't break the thesis.
About cyclicality, again it is accounted for, the moment I bring the annual revenue closer to what the SCAs will soon guarantee for 5+ years.
I don't believe this is peak cycle. HBM is not a commodity but is designed for each oem chip, it's not standardized like DRAM so it won't follow the same economics / competitive pressures as it ramps up. Mu is 2nd largest in HBM. Going forward autonomous driving (or even without it), the deals with GM and Ford set up the next cycle as automotive sector will lap up memory in the near term; humanoids maybe, but also autonomous evtols and military drones (Trump & Putin have created an environmemt where military investment will flow).
What works better for MU than even SNDK (which sells a lot to China & MU has significantly less exposure) is the inability of Chinese manufacturers to catch up to the tech (larger dies, poorer economics) and supply to US/EMEA due to regulation. Plus they've done well to steal DRAM market share from Hynix in the last 2 years.