Intel to Eventually Exit NAND Flash

First the context

Intel announced record revenue and unit shipments in Q4'07 that beat their previous record by a half billion dollars. Intel netted $2.3B from a gross of $10.7B. Gross margins, a highly watched indicator of profitability, came in at a whopping 58%, almost unheard of in any business. Further, profits grew significantly faster than revenue, another sign of a keenly improving operation. The primary drivers for the impressive gross margins were: higher unit volumes, lower unit costs, and lower 45nm costs. The 45nm process is a breakthrough technology that utilizes for the first time hafnium based high-k metal gate technology.

Yet Intel's stock price got spanked, down 15% immediately in after-hours trading. What was the underlying reason behind the seemingly dichotomous market reaction? Weaker than expected Q4 results, uncertainty in the overall economy, and less than anticipated Q1'08 guidance contributed to the stock price demise. But the primary reason was Intel's NAND flash business.

Flash History 101

Intel pioneered the non-volatile Flash memory technology in the late 1980s. The key distinguishing advantage of flash memory was that it was non-volatile, meaning that the memory chip didn't lose data when powered off unlike DRAM for example. After transitioning to the much more lucrative microprocessor business, Intel's primary motive for remaining in the NOR flash business was to keep open a second flank on AMD so that they couldn't compete as effectively in the higher-margin processor business. Recently NOR flash started to lose out to another flash technology called NAND, which was much better suited for the growing data storage market that is utilized in devices such as mp3 players, digital cameras, and just about every other device that needs to retain data after power is turned off. And NAND was the fastest growing semi-conductor segment, period. Intel was even willing to enter the NAND business late because it found a way for it to complement its main processor business by including flash memory on its computer motherboards in its vPro platform and its still to be proven Turbo Memory offering.

The Rub

This all sounds promising for NAND, so what's the rub? Despite the near ubiquitous nature of Flash memory, Intel never found a way to make a profit with it, even when it was #1 in the NOR market for years, much less now with NAND where it's #5 with 3% market share behind Samsung, Toshiba, Hynix, and Micron respectively. In 2006, the Flash business lost $555M when Intel finally had enough and spun it off with STM to get their Flash assets off the books.

So the obvious reason Intel will jettison its remaining NAND flash business is profitability, or lack thereof. In the current environment, Intel has made clear that it no longer has the patience to tolerate money losing ventures as they have demonstrated by selling off numerous unprofitable businesses over the last two years. There's also a strategic reason that makes this prediction a near certainty. Competing in NAND exacerbates a conflict with a potentially important and strategic customer - Samsung. Samsung currently purchases significant quantities of processors for PCs and is working with Intel to develop the next generation of DRAM technology to benefit Intel's primary PC business. Further, Intel is counting on Samsung to be a strategic customer in a key future growth area, the $10B mobile internet device consumer electronics market.

The only issue that is preventing Intel from exiting NAND immediately are its existing contractual obligations. So the question is when, not if Intel will sell off NAND.