Disk drives are smaller and use less material, but we're making more than ever. What are we to do with them when we're done with them?
I mentioned HP’s long-running circular system for printer cartridges. Implementing circularity for electronics is a near Sisyphean task today due to lack of information, expertise and management support. But it can be done — and, indeed, it must be done.
Check out Google’s approach to circularity in their data centers and self-designed servers as described on their sustainability page, and while you do that, note the first photo on this page. This shows a stack of intentionally damaged disk drives; that’s how you make sure the data on an otherwise end-of-life drive is unavailable for theft. While these appear to be a variety of brands, a couple are clearly 2 terabyte drives.
Twenty-five years ago, when I worked at Tandem Computers, we had a customer with 2 TB of disk space. That was the largest amount of disk space of any customer we had at the time. Of course this was not achieved with a single 3.5” drive, though! It took A THOUSAND 2 GB drives! A thousand drives. For two terabytes. Along with all those SCSI controllers (eight drives per controller). Supporting two terabytes of storage in 1995 required racks and racks of equipment — perhaps an entire datacenter. Just let that (and the cost of it) sink in…
So yes, storage density has been dramatically improved over the past couple decades, and, as a result, the absolute amount of material and energy use necessary per unit of data storage has been drastically reduced. On the other hand, demand for data storage has skyrocketed since then and, according to Statista, the number of hard disk drives being sold annually for storage on an absolute basis has more than tripled since then.
Despite that, this is an extraordinary achievement, yet we haven’t done much about reusability in this space. Material wearout limits the absolute life of a mechanical hard disk drive (and semiconductor solid-state drives, for that matter), so the potential for reuse at the product level (once you can get past the risk of data retrievability) is limited.
Disk drives contain a lot of aluminum, especially in the body of the product, and their magnets all have some amount of neodymium and dysprosium, two rare earth elements that (along with all other REEs) are considered “critical minerals” in the USA and EU. While recovery of the aluminum in recycling processes is well-understood, it took a major multi-stakeholder concept demonstration project at iNEMI to assess what and how to recover the REEs from used disk drives.
This clearly isn’t easy; this project only addressed magnets in disk drives (as well as securely wiping and reselling functional disk drives post-use). Are projects like this necessary for many, or most, or perhaps every other component of your products?
Extending the life of a system, component or subassembly by refurbishing and either re-using them or selling them on the aftermarket, as Google does, is a very good — and relatively low-cost and easy — first step. But that just delays the inevitable, which can often be landfill. A good alternative could be further reuse or recycling. That should be today’s goal. But what about tomorrow?
I see two ways to approach this:
Start to think about what information is needed to enable circularity for your products. See what it takes to collect that information; it’s a significant project but could lead to significant returns once basic information is provided to management for support and funding.