IBM Makes Revolutionary Racetrack Memory Using Existing Tools
Racetrack memory could someday supersede flash in terms of density and cost. — IBM has shown that a revolutionary new type of computer memory—one that combines the large capacity of traditional hard disks with the speed …
Context & Ripple Effects
IBM has been working the next-generation-memory problem all year: back in June 2011 its researchers cleared two core fabrication and scaling hurdles for the technology, and this demonstration answers the question that hurdle-clearing left open — whether racetrack memory could actually be built on the equipment chipmakers already own. That manufacturability claim is the whole ballgame, because a novel memory that needs exotic tooling never escapes the lab.
The timing matters for who IBM is in late 2011: the company overtook Microsoft in September to become the world's second-most valuable technology firm, a repositioning away from the PC era toward exactly this kind of deep R&D bet. A storage technology that pairs hard-disk capacity with faster access would give IBM leverage over the flash incumbents whose business rests on NAND's density limits.
First-order effects
- Flash manufacturers — the NAND suppliers whose density roadmap defines smartphones and SSDs — now face a credible research path to a denser rival that claims compatibility with existing fab lines, undercutting the usual 'retooling cost' defense.
- IBM converts six months of incremental physics results into a strategic patent-and-licensing position in post-flash memory, consistent with its shift up the value chain after passing Microsoft in market value.
Second-order effects
- If the existing-tools claim survives scale-up, incumbent memory makers' installed fabs flip from liability to contested asset — the same lines could produce either NAND or a successor, so the fight moves to licensing and IP rather than capital expenditure.
- Device makers planning multi-year flash supply commitments gain a second candidate architecture to benchmark against, pressuring NAND vendors to accelerate their own density roadmaps.
Third-order effects
- Racetrack is one of several architectures racing to displace flash; whichever proves manufacturable first reshapes a memory market currently structured around NAND economics, and toolchain compatibility looks like the real gate between lab demo and product.
- A successful transition would concentrate power with whoever owns the foundational IP — likely shifting value from commodity flash fabrication toward the research organizations and licensors, a pattern IBM is explicitly positioning itself for.
The trend: Non-volatile memory research is converging on displacing flash, with manufacturability on existing fab tools emerging as the decisive filter between laboratory promise and commercial reality.