Part 1 of this series is here.
Let's start with Elpida and Rexchip. People seem to think Elpida was a weak player in the DRAM market and went bankrupt because they were a second-rate house. You might point out to people in the pro-formas (as one can see from your Feb 2013 8K), Elpida only filed bankruptcy because their creditors (including the Japanese government) appeared willing to force them to make a large debt payment on March 31, 2012. Had the creditors simply kicked the can down the road, a little, Elpida could have made it through to a very profitable 2013.
That said, from a strategic standpoint putting Elpida and Micron together is an incredibly good idea. Most would agree Elpida has better DRAM technology than Micron, or at least a number of their strengths in Mobile DRAM and Graphics DRAM certainly complement Micron's server DRAM. Elpida was able to demonstrate (maybe not build) 25nm DRAM as early as the leader Samsung (OTC:SSNLF), and clearly ahead of Micron. Clearly Micron is going to GREATLY benefit by being able to skip a 2Xnm DRAM process generation and accelerate its 20nm development.
What is not well-known is Elpida actually demonstrated 50nm ReRAM in 2011/12. Many feel this technology may be one of the winners to replace DRAM and NAND in certain applications. Maybe Elpida's announcements were more compelling than their achievements, however if they had real achievements, you might inform your stakeholders there is more.
Since many mobile applications are MCP (multichip packages) combining MDRAM and NAND in a single package, Elpida will also benefit by being able to access Micron's NAND. They will no longer be just another merchant purchaser of NAND on the open market. Working closely with Micron, their new MCP offerings should be even more tightly integrated and tailored to application needs.
One weakness of Elpida was its reliance on third party test gear. Micron's willingness to supply their internally developed technologies was an important factor in the US Chapter 15 bankruptcy proceeding and swayed a skeptical Judge Sontchi in Micron's favor. Integration of Micron's in-house test into Elpida's processes will reduce costs substantially.
In July 2012, you presented Elpida and Rexchip as having 205,000 wafer starts per month (WSPM). More recently you have presented them as only having 185,000 WSPM. Was the shift to more advanced lithography for 30nm and 25nm the reason for the reduction? How much investment will it take to regain the higher-capacity?
Before Hynix decided to hold a barbecue at their Wuxi China facility, Hynix had approximately 280,000 WSPM of DRAM, or only about 100,000 more than Elpida/Rexchip. DRAM is by far Hynix's largest product line, as their NAND capacity is less than 130,000 wafers, and their system logic capacity is relatively irrelevant.
Hynix has a market cap of $23 Billion Dollars. Let me repeat that. $23 Billion Dollars. If you ignore their relatively minor systems logic group using older fabs, Hynix only has about 2x the 300mm wafer fab capacity of Elpida and Rexchip (before setting fire to some of it).
Elpida came with about 1/2 the cash Hynix has and it came with only 20% of the debt (most of which is no interest payable from 12/2014 to 12/2019 - how sweet a deal is that - 6 years of installment payments at no interest). The negative is that Elpida's margins are not as good as Hynix's, but that cuts two ways. Now that Elpida is out of bankruptcy and merged with Micron, there is a tremendous opportunity to rationalize pricing and improve margins going forward.
Let's take a look at the apparent appraised value of Elpida with SKHynix's metrics in mind. While it is not accurate to say Elpida should be worth 40 to 50% of Hynix, it is absurd to claim Elpida/Rexchip is only worth about 10% of Hynix. Simply absurd. Unless you know Micron will screw up Elpida and Rexchip, to tell the investment community the purchase accounting gain of $1.484B appropriately values Elpida and Rexchip is crazy. Many analysts are still doing their Price Targets on price to book, and Elpida book value has been badly mangled by the purchase accounting.
I am not sure how you were able to get Price Waterhouse to sign off on the appraisals, but it sure seems as if FASB 805/820 has been pushed beyond the spirit of the rules. If your reference point was the liquidation of the Qimonda DRAM facility in Virginia in 2009, maybe you can justify the $935M value. If you based the appraisal on the liquidation of the Powerchip P3 fab, maybe. However in those two cases, they were simply selling equipment. Micron picked up operating PROFITABLE businesses with tremendous potential synergy.
Elpida/Rexchip created over $500M of operating cash flow in Q3 2013, with over $1B of revenue, while probably still recovering from the bad deals they had to accept to keep customers during their bankruptcy proceedings. So Micron writes checks for under $1B, gets control of more cash than that, has 6 years of time to pay off the debt at no interest, picks up a fair amount of good technology, and importantly gets to pick up more DRAM wafer capacity than Micron owned prior to July 31, which if we read the situation at Manassas, Va and Tech in Singapore correctly was only about 100K WSPM.
Seeking Alpha writer, Electric Phred published a prediction of how Micron might "hide" Elpida's value on Seeking Alpha back in August before the Analyst Day meeting.
As, we thought you were going to try to grossly understate the value of Elpida so as to have cushion for future earnings.
Set aside the atrocious appraisal for a moment and just look at some historical financials. Elpida just before bankruptcy had almost $9B of assets as of 12/31/2011, (which was 60% of the assets Micron reported on May 31, 2013) less than 21 months ago. After taking a $2.8B extraordinary write-off in bankruptcy, and taking the losses for the first 9 months of 2012, Micron's Feb 5, 2013 pro forma showed Elpida as having $5.5B of assets as of 9/30/2012, just over 12 months ago.
While we await your next 8K, later this week, which is supposed to report Elpida's pro forma results through June, we believe it is going to show that the assets grew from September 2012 (without any extraordinary write-offs). It appears as if Elpida probably acquired over $400M of capex since July 2012, maybe more if we believe what Kipp Bedard, Vice President, Investor relations, said on a September conference call that Elpida has ramped up spending on equipment heading toward the $800M/yr target.
So what we know now is even after buying the approximately $400M of capex, Elpida's cash balance more than doubled to $1B, its inventory and accounts receivable increased by over $100M. Seems like a pretty good bet the assets on July 31 without the $2.1B write-off would have been higher than last September.
As of September 30, 2012, Elpida's Property, plant, and Equipment (PPE) was valued at $3.5B (before Yen depreciation), and plus Elpida's capex since then, which you now value at $935M. You've done almost exactly what Electric Phred predicted on August 8, nearly to the dollar. If Elpida invested a net of $500M (after depreciation) as of August 31, 2013, that would be a $3B write-off of the PPE.
Really?! In Elpida and 89% of Rexchip, you just picked up what we believe to be about 65% of the actual wafer fab capacity which Micron directly OWNS (excluding Intel's share of Lehi's capacity, and Inotera's 120K for MU) at a cost of $935M, or about $5,000 per WSPM. As of 5/31, Micron valued your owned PPE at $6.8B at about $25,000 per WSPM (Wafer Start Per Month).
So if Micron valued the Elpida/Rexchip PPE at the same rate it valued its own, it would be over $4.75B. So there is almost $4B of PPE which is not going to be written off over the next five years, or almost $800M/yr of "excess" gross margin potential from reduced Depreciation & Amortization, or about an additional $0.70 EPS related to reduced depreciation. Analysts and shareholders need to understand this when comparing the New Micron to the metrics of its competitors.
It does make one wonder why any of the analysts would worry about the installment debt. You pay $615M, you borrow at zero interest $1.4B, payable over 6 years, and you get over $5B of assets, which are generating operating cash flow of $500M in the last quarter.
Since Micron has not recently described in a clear manner your manufacturing capacities and capabilities, we will take a stab at it in the next installment. Reading most of the analyst reports, they do not seem to have a particularly good handle on your manufacturing situation and the longer-term implications for Micron.