Spintronics centres on the control of the magnetic properties of matter for memory and logic applications. Efforts in the control of magnetization in ferromagnetic materials and of the spins of moving electrons have led to important achievements from both a fundamental and a technological point of view (see for example our March 2015 focus issue on spin-transfer-torque memory; http://www.nature.com/nnano/focus/stt-mram/index.html). New directions have opened in the last few years, and attention has been given, for example, to skyrmions, a particularly stable configuration of ferromagnetic materials, as well as to magnons, that is, spin waves that can transport information without any movement of electrons. In this issue, on page 231, Tomas Jungwirth and colleagues review progress in another direction in spintronics, namely the control of magnetic order in antiferromagnetic materials.

The idea that an antiferromagnet can be used to store information seems, at first, counterintuitive. Atomic magnetic moments are ordered but their net magnetization is zero, thus invisible to any magnetic probe. In reality, it is not necessary to use magnetic fields to write or read information. Modern spintronics devices rely on the interaction between the spins of moving electrons and local magnetic moments, which is, in principle, possible with antiferromagnets too. Indeed, the advantage with respect to ferromagnetic materials is that once the information is stored, it is more secure as it cannot be easily read, and it is more stable because it is unaffected by spurious magnetic fields, external or internal to the devices.

There are other reasons why antiferromagnets could be useful for spintronics devices. For example many antiferromagnetic materials are insulators, which could be useful for the propagation of spin waves. Also, the intrinsic frequency for the switching of magnetic order is in the terahertz regime, which could lead to very fast responding memory devices.

As Jungwirth and colleagues clearly state in their Review, the field is still in its infancy. It is too early to predict whether devices could be efficient and reliable enough to raise the interest of industry. The recent demonstration of the control of antiferromagnetic order by electrical currents is only the first, though significant, realization of a potentially scalable device (P. Wadley et al., Science 351, 587–590; 2016). But the prospects are numerous and varied, and we can expect rapid and interesting developments in the near future.