Ever since data was being recorded on magnetic media, the only way to increase storage capacity within a certain standard form factor could be only increased by writing more data into a smaller space. That's why a number of technologies were invented to pack ever more bits into the same or smaller area on the physical area on disk. Single / Double Density floppy disks were replaced by High Density media on the physical side of things, and FM recording was replaced by MFM or GCR data encoding on the software side of magnetic recording. When storage shifted towards hard disk drives, technology like perpendicular recording (PMR), helium-filled drives and Shingled magnetic recording (SMR) drives followed.
All these technologies were facing a similar set of problems, what is known as the "Magnetic Recording Trilemma": Three directions are competing against each other: the requirements of readability, writeability and stability. The problem stems from the fact that when written magnetized bits on a magnetizable surface move nearer towards each other, a medium with a higher coercivity is needed so that magnetic domains withstand any undesired external magnetic influences, to remain reliable and separate in their individual locations. As the change from Single Density to High Density floppy disks drives has shown, this change in media means different write-heads need to be employed, heads which are able to produce a stronger magnetic force to overcome this higher coercivity during write. And this is where current available technology reaches a limit. Drive heads can become smaller and smaller but with ever increasing areal density the area occupied by one single bit of data becomes so small that these smaller write-heads are not able to overcome the coercivity of the magnetic medium anymore and thus are not able to flip the magnetic domain. Again, at least with current technology.
This is where HAMR technology comes in and where it is able to break this limit. The coercivity of many materials is temperature dependent. One example of this effect can be seen by heating a magnetized metallic object, such as a needle, in a flame. When the object cools down, it will have lost much of its magnetization. The rule behind this is that if the temperature of a magnetized object is temporarily raised above its Curie temperature, its coercivity will become much less, until it has cooled down again. HAMR uses this effect to overcome the limit of not being able to overcome high coercivity with current write-heads and their limited magnetic force. Different manufacturers tackled the challenge of heating a surface differently. One current solution by Seagate is to use a tiny laser diode to temporarily heat the miniscule area on the magnetizable surface, so that if briefly reaches a temperature where the disk's materials loses enough of its coercivity so that write-heads are strong enough to flip the magnetization. The magnetic domain that is heated is very small and as such it also cools down very quickly after the heating, usually in under 1 nanosecond. Once the surface area has cooled down, the high coercivity returns and stored data is reliably magnetically stored or "baked" into the surface.
Some of the challenges during the development of HAMR drives where the implementation of a technology that is able to heat very tiny areas. Laser diodes in combination with the use of diffraction and waveguide phenomena solved the problem. The magnetic medium had to change as well and is more of a glass compound in HAMR drives, a material that exhibits optical properties in combination with a heat conduction profile that enables spot heating. These new solutions in combination with multiple positioning actuators, raised platter count and refined controller logics make drives possible with 2 TB of data per square inch areal density and will soon reach 10 TBpsi.