Researchers demo cheaper resistive random access memory fabrication, raising prospect of terabyte chips for smartphones
Super-Dense Computer Memory — Researchers have discovered a new way to make chips that could pack terabytes into smartphones. — A novel type of computer memory could …
Context & Ripple Effects
Resistive RAM's terabyte promise is not new: Crossbar claimed in 2013 that RRAM would explode the $60B flash memory market with a terabyte on a chip, and the idea of nanotech-enabled terabyte storage dates back at least to Wired's 2007 thumb-drive story. What has kept those claims from reaching phones is fabrication cost — RRAM worked in the lab but not at prices NAND buyers would pay.
The new research attacks exactly that bottleneck, demonstrating a cheaper fabrication route for RRAM. It lands alongside other post-NAND bets from the same era, including IBM's racetrack memory built with existing tools in 2011 and Intel and Micron's continued shrinking of NAND flash, which shows incumbents still have runway.
First-order effects
- Crossbar and other RRAM developers gain their strongest argument yet against the cost objection that has kept the technology out of volume production since the 2013 market-entry claim.
- Smartphone makers evaluating next-generation storage now have a credible path to terabyte-class chips that does not depend solely on NAND scaling.
Second-order effects
- NAND incumbents like Intel and Micron face pressure to accelerate node shrinks or develop their own resistive-memory roadmaps rather than cede the high-density end.
- If RRAM fabrication costs approach NAND levels, handset designers can plan for storage tiers well above current capacities, shifting bill-of-materials weight toward memory.
Third-order effects
- A working low-cost RRAM process would mark the first real crack in NAND's monopoly on non-volatile memory, splitting the market between scaled flash and resistive alternatives much as DRAM and NAND split decades ago.
- The pattern — lab demos in 2007, startup market claims in 2013, cheaper fabrication by 2014 — suggests emerging memories mature through fabrication-economics breakthroughs, not physics ones, keeping the lab-to-fab gap the key uncertainty.
The trend: As NAND scaling delivers diminishing returns, a wave of alternative non-volatile memories — RRAM, racetrack, memristor variants — is competing on fabrication cost to break into mainstream storage.