SanDisk Flash Summit Keynote
August 9, 2011, Flash Memory Summit Conference, Santa Clara, CA—Yoran Cedar, CTO at SanDisk, look at the future prognosis for the flush industry. New devices, applications, and platforms are driving increases in the overall market.
On one end of the spectrum, social networking and cloud services are driving the use of more flash at the enterprise level. Cloud servers and some cloud storage is converting to SSDs can for both performance and power improvements.
On the user side, 220 million devices operating on iOS were sold last year and over 500,000 Android platforms were activated. These client devices use flash for storage and mobile devices overall use one third of all NAND flash bits. Depending on the analyst, tablets are projected to grow to over 300 million units by 2015 and will account for 15 percent of all flash bits sold.
These trends anticipate long-term growth as well as predictable supply and demand in the market leading to a healthy industry. SSDs are projected to grow to 25 percent of the NAND market for both consumers and enterprise storage products by 2015. One quarter of all flash will still be for cameras, flash drives, and other end markets.
Technology enables innovation, but this innovation can be very disruptive because we don’t know what is coming next. Supply for flash is increasing as more fabs come online and new designs into scaled processes increased device performance and capacity. Industry growth rates are moving towards a level of slow but steady increases of about 50 percent growth per year. Total bits will continue to grow on a year-to-year basis.
One of the challenges facing the industry is the need for application awareness for adaptive flash management. In the next five years, the industry will move from the 2x nanometer node down to the 1x nanometer and even the 1y nanometer levels. If possible it will be done using existing immersion lithography because the extreme UV equipment isn’t available for production yet.
One developmental area is the move to vertical stacks of memories. This development can use older technologies and thereby extend the life of existing processes while still delivering the desired increases in bit density per area. Some of the alternatives to flash include 3-D R-RAM which requires EUV and have devices which are difficult to manufacture. Another alternative is to develop 3-D stacks, but the crosspoint interconnect can be very resistive and the TSVs are not easy to implement.
