Bottom in the memory industry?

• Memory prices are in the dumps.
• The creation of newer markets due to price erosion.
• Increased demand due to higher memory capacities and new markets, more than offsetting any price per bit declines.
• Memory vendors currently trade at rock bottom valuations.
• Most existing players are looking to divest or merge. The survivors would have better pricing power, and the memory industry should do well.
• The trend towards mobility and computing convergence will drive growth towards newer markets and higher capacities.

As the average selling price (ASP) continues to plummet, elasticity dictates that newer markets and applications would be discovered, which would simulate sales growth. Andy Kessler discovered the same phenomena with EPROMs in the 1980s. Really, nothing has changed in terms of demand supply curves. Elasticity is the guiding theme in this sector.

Right now, as memory prices collapse and vendors bleed red, memory sector companies like Micron Technology Inc. and SanDisk Corp. sound like scary places to be. A slump in demand, coupled with an oversupply, has forced vendors to sell memory chips below cost. Many have predicted an impending shakeout with the smaller players either exiting or being bought out. Most chip companies like Intel Corp., Advanced Micro Devices Inc., Infineon Technologies AG and STMicroelectronics N.V. have spun off or merged their 'undesirable' memory divisions, or are in the process of doing so. The memory sector has been left for dead, which is exactly the best time to get in from a contrarian perspective. A similar shakeout took place in the mid 1980s, when Intel exited the DRAM sector. That turned out to be a long term bottom in the industry.

The number one use of flash right now is in Apple's iPods. Just contemplating the possible markets where memories could be widely used gets me excited.

• Think of the growth in feature rich cell phone markets. Multiply that growth with the increasing memory capacities for each of those cell phones. You get the picture.
• The introduction of the high memory capacity iPhones by Apple should drive other phone providers to beef up the memory capacity in their phones, thus accelerating the trend towards higher capacity phones.
• While everyone is anticipating a faster 3G rollout because of the iPhone, fewer analysts are anticipating the growth in memory usage. Broadband and wireless just gets wider coverage than the commoditized memory sector.
• In 2007, 271 million PCs (laptops+desktops) were sold. In contrast, a mind boggling 1.15 billion cell phones were sold last year (and growing fast). As these become more feature rich and pack in higher capacity chips, the memory vendors who survive the current shakeout are in for a bonanza.
• Solid state drives in laptops are on the cusp of mass adoption. Five years from now, most of us would be toting around flash enabled SSD drives. Flash memories would be faster, less noisy, more power efficient and increasingly a preferred usage choice for even high capacity drives.

Check out the graph below from a Micron presentation. While I agree with the growth projections, I'd just like to add that it probably underestimates the cannibalization of MP3 player growth rates due to integrated phones with music capabilities (like iPhones), and underestimates the erosion in USB Drives due to online storage and cloud computing applications.

Most flash memory vendors with scale and a strong patent portfolio should do well in the upturn. The only caveat I would add is that there is a potential threat from newer memory technologies (FeRAM, MRAM, phase change RAM), but their commercial deployment and proven success is still many years away.

In short:

• We are truly on the cusp of a phenomenal growth cycle in memory chip usage.
• The best part is that people absolutely hate the sector right now.
• This makes it a classic contrarian pick.
• The current carnage would scare away incumbent semiconductor players toying with the idea of an entry.
• Demand-supply elasticity tells us that the memory sector should scale beautifully over the next few years.

To conclude, the discounted memory chip vendors are at the bottom of the economic and industry cycle and their future prospects look very promising.