It's no secret that consolidation in the memory business, coupled with a tragic fire at one of Hynix's (a leading vendor of DRAM) plants that dramatically reduced supply, has led to substantial gains for the remaining DRAM players. While domestic investors would probably have a hard time buying shares of Hynix or Samsung (NASDAQOTH:SSNLF) (and let's be honest, when you buy Samsung you're investing in much more than DRAM), a popular memory play has been Micron (NASDAQ:MU) up a whopping 255% year-to-date. A question that's worth exploring, despite the stock's already massive run, is whether Micron is a potential takeover target.
Why is DRAM so important?
Despite being largely a commodity, DRAM is an essential component of any device that does meaningful computation. Think of DRAM as a nice and fast buffer from which the central processing unit of a system can store and fetch the most important data that it needs. To illustrate this, understand that a typical PC usually ships with 4-8 gigabytes of DRAM, but usually has at least 128 gigabytes of primary storage (much more is common with systems that use hard disk drives). The more DRAM a system has, and the faster that RAM is, the snappier the system will typically be.
Two secular trends that cannot be ignored
Each year, mobile devices tend to have more memory packed in. These devices are becoming much faster and are quickly becoming the computing devices of choice. Continued growth in shipments of devices using a non-trivial amount of memory drives a very real and powerful secular trend.
But, there's another underlying trend that cannot be ignored -- more deeply integrated "system-on-chip" products for all flavors of computing devices from servers to smartphones. Eventually, the computing power found on a single system-on-chip will become so great that traditional DRAM interfaces will simply be too slow or too expensive to service the bandwidth needs of those engines. This has always been the issue and it is why there has been such an intense focus on cache performance in CPUs (caches are very fast and very small memories built directly onto the CPU).
While caches have worked at the individual core level, and for client processors with 2-4 processor cores (in servers with at least 8 cores, the caches become very large and expensive, but those applications can afford it), in a full system-on-chip (one that doesn't sell for hundreds or thousands of dollars) with very powerful graphics, multiple CPU engines, and other IP blocks dedicated to specific functionality, a better long-term solution is needed.
A long term solution: bringing DRAM closer
A potential solution is to more tightly couple memory and the computing units together. This is already done to some degree in smartphone chips with what is known as a "package-on-package" interface, where the memory and the applications processor are "stacked," but it needs to go much further. Intel (NASDAQ:INTC), for example, designed a specialized eDRAM, short for "embedded DRAM," for its highest-end notebook processors to feed their very high performance graphics engines. However, this is but the first step and the custom eDRAM is only suitable for rather expensive parts.
Intel acquiring Micron would make sense. DRAM business is highly profitable and would be immediately accretive to earnings. It would also allow Intel to have more leverage in the mobile processor market as it would control more of the bill of materials, particularly once DRAM and apps processors are integrated as a single solution.
What about Apple?
The biggest losers from the consolidation of the memory business are the device vendors. With fewer suppliers, and those fewer suppliers modulating production to keep memory prices up, it's got to be tough for Apple (NASDAQ: AAPL), which uses plenty of memory in all of its computing devices. If Apple were to peel off a cool $35 billion to buy Micron, it would be immediately accretive. Further, Apple would be able to source DRAM and NAND at cost for its devices, allowing the company more flexibility to profitably play the "specs" game against lower margin handset players.
Further, while the acquisition would be pricey, it probably wouldn't be too difficult to integrate. Apple could just run things as usual, but with a cost structure advantage for its own product lineup. Additionally, as Apple's chip designs demand it, Apple could more successfully integrate its memory products (and perhaps even its NAND flash) into its system-on-chip products. This is a situation where Samsung would have a longer term advantage over Apple since Samsung owns its own memory factories.
Foolish bottom line
This isn't to say that Apple or Intel will buy Micron. It's also just as likely that these two players will begin to work more closely with Micron for future designs. However, given Micron's newfound profitability in a now highly consolidated environment, it would be accretive to both companies' earnings. Intel probably wouldn't be able to afford the acquisition (unless it took out significant new debt and/or did the deal mostly in stock), but Apple sure could.
Ashraf Eassa owns shares of Intel. The Motley Fool recommends Apple and Intel. The Motley Fool owns shares of Apple and Intel. Try any of our Foolish newsletter services free for 30 days. We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy.