IBM continues to do the fundamental research that others have abandoned

It launched the first hard drive in 1956. Called RAMAC, the 1tonne device stored 4.4Mbyte of data and offered 'unprecedented performance' by permitting random access to any of the characters distributed over both sides of 50 disks, each 2ft in diameter. Then, in 1966, Bob Dennard invented the one transistor dram, inspired by IBM research into thin films; work that would later find other data storage applications. IBM has also been interested in exploring technologies at the molecular level – at the atomic level if it can be done. In 1989, researchers at IBM's Almaden Lab used a scanning tunnelling microscope to place 35 atoms precisely, creating an atomic scale rendering of the letters I, B and M. Data storage has continued to capture the company's imagination. A few years ago, it unveiled Racetrack, which uses spintronics to boost data access speeds. First discussed in 2004, the technology is getting ever closer to prototype. Now, it's taken another step forward by creating a memory bit using just 12 antiferromagnetic iron atoms – the least number needed to reliably store a bit – on a copper nitrate substrate. The team also used 96 iron atoms to create a byte; eight bunches of 12 atoms. If manufacturable, this approach could boost storage densities by a factor of 400. There was a time when leading electronics companies invested as much in R as they did in D. Times change; now, most are only interested in D. IBM – which was granted 6180 US patents in 2011, capitalising on the work of more than 8000 researchers distributed across 36 countries – remains an inspiration.