David,
You're making this way more complicated than it is ... a point of clarification ... for me 'last' quarter is the Nov. quarter and 'this' quarter is, well .. this quarter. Let's call them Q1 and Q2 so we know if we're talking about the same thing or not. Now we have revenue (unit production x price) and costs and we know quite a bit more about these numbers than you think. Onward ...
1) costs ... contrary to what you think MU does not reduce its absolute dollar costs over time (what an unbeatable business model if they could pull off that trick!). What it does reduce is its costs per unit production ... and they do an excellent job of that due to process technology skills. If you actually go and dig out the numbers, their costs (all operating costs, i.e. cost of goods, SGA, R&D) in the last 4 quarters have been $343M, $366M, $356M and $372M in the August quarter ... I'm not making these up .. if you do the work you can dig 'em out yourself. The trend is slowly up and I'm estimating $380 for Q1. Of course, unit production in this period has increased tremendously (about 100% from Q1/fiscal 97 to Q4/97) with the result that costs per unit have declined sharply. It's interesting that their absolute costs are relatively fixed (what you would expect with the huge investment in plant and equipment ... raw silicon doesn't cost much). It means there's tremendous leverage on earnings from changes in revenue.
2) revenue ... OK prices ... dateline .... Nov. 26/97 ... Boise, Idaho ... The Salt Lake Tribune .... reporter Lisa Carricaburu chats to Bill Stover after AGM ... and Bill says that the average price of a 16-megabit DRAM chip has decreased to $4.50 in the company's current first quarter from $12 in the fourth quarter of (fiscal) 1996. "Think about how you would manage your household if your annual salary decreased from $50,000 to $19,000," he said (yeah, life's a bitch Bill). So that's what I'm using for my ASP, $4.50. This is about what I would have guessed using Achilles data and nudging it up for 64 Mbit production and SDRAM. And I could be wrong, it might be closer to your $5-5 1/4 range because Bill maybe meant just the ASP for EDO DRAM and not all 16 bit DRAM ... i.e. didn't include SDRAM. BTW I disagree with you that margins on SDRAM will be higher. Prices are (temporarily) higher but so are costs ... test and assembly ... remember MU has a bottleneck in T&A because of the switch to SDRAM. Anyway, to make up for possibly being low on price I'm going to be high on unit production, 15% versus your 10% increase. A $4.50 ASP x a 15% increase in unit production equals revenue of $385M. Which is not a great number. You can see the leverage, revenue drops from $484M in the August quarter to $385M and operating income (this is just MU, excludes MUEI) drops from $111M to $5M (leverage cuts both ways, Bill).
3) Q2, this quarter ... Prices are starting at $3 and I fearlessly predict they will end the quarter at $2 for an ASP of $2 1/2. This is due to increasing supply (new fabs, e.g. Samsung in Austin .. the biggest clean room in the U.S., and the conversion to 64 Mbit ... where MU is a laggard, not a leader) and a slack PC market. OK, let's say MU increases bit production by 50% sequentially (due to ramp in 64 Mbit). With an ASP of $2.50, this gives us revenue of $321M. Well costs aren't going down ... they have to buy new test equipment, gear up to manufacture 64 Mbit chips, etc... my guess is costs bump up another $10M to $390M. This equals big loss and is why management is so vocal about the Koreans. The only thing that can change this picture is if prices go up and not down. This outlook is why I'm saying that if management could swallow their pride and do the smart thing they would kitchen sink Q1. 'Cause betting that prices will go up this quarter sure looks like a long shot!
regards, Mike |
