Summary
When using more recent data to compare the embodied CO2e of HDDs to SSDs, the results show that the manufacturing emissions associated with SSDs are less than 2x that of HDDs, instead of 8x as found in previous studies.
There are academic papers that have been published in the last couple of years comparing the embodied carbon dioxide equivalent (CO2e) of hard disk drives (HDDs) and solid-state disks (SSDs). One example is the University of Wisconsin paper “The Dirty Secret of SSDs: Embodied Carbon,” published in October 2023. In that paper, researchers compared the embodied CO2e of consumer-level HDDs and SSDs manufactured in 2017 on a single device to single device basis. If you take its comparisons at face value, you’d walk away believing that SSDs are up to 8x worse than HDDs when it comes to manufacturing (aka non-use phase) emissions and that building storage systems from flash is much worse for the environment than building them with HDDs.
While that may have been true a few years ago, today, published data from HDD and SSD manufacturers shows that the embodied CO2e of SSDs has decreased much faster than it has for HDDs, and that a current comparison has the manufacturing emissions associated with SSDs at less than 2x that of HDDs.
Getting to an Updated Embodied CO2e Baseline
In comparing HDDs to SSDs, the University of Wisconsin study showed 20kg CO2e/TB for the HDD and 160kg CO2e/TB for the SSD, an 8x difference. Using publicly released data from SSD drive vendors for enterprise drives manufactured in 2021, I compared an 20 TB Seagate EXOS X20 HDD with an embodied carbon intensity of approximately 1.3 kg CO2e per TB to a 15TB Seagate Nytro 3332 SSD. According to Seagate’s 2021 analysis report for the Nytro 3332 SSD the results were based on ISO 14040 and 14044 compliant tools and methods:
Seagate’s sustainability assessment tools used to generate the product sustainability analysis have been verified by UL in accordance with ISO 14040, ISO 14044, and the World Resources Institute and World Business Council for Sustainable Development’s GHG Protocol Product Life Cycle Accounting and Reporting Standard.
Given the ISO compliant methodology, there is no reason to doubt the GHG Emissions by Life Stage values the analysis Seagate provided:
Figure 1: A GHG Emission by Life Stage for a 15 TB Seagate Nytro 3332 SSD3
Using the 5 year life cycle defined within the analysis for the Nytro 3332 SSD, I calculated that it is responsible for a total of 189.75 kg CO2e (2.53 kg/TB-yr x 15TB x 5 years). Per the GHG Emissions by Life Stage values shown in Figure 1 (above), approximately 80.4% of the GHG emissions are attributed to product Use Phase. That means a maximum of 37.2 kg CO2e or approximately 2.48 kg CO2e per TB is embodied CO2e. Instead of an 8x difference, the Nytro 3332 15TB SSD had only 2x the embodied carbon of the comparable HDD!
It is also important to note that enterprise SSD densities are increasing faster than enterprise HDD densities, a trend that is very likely to continue well into the future. Given that 61.44 TB SSDs are already generally available today, 122 TB SSDs are likely to become generally available within the next one to two years. The largest HDDs in general availability are in the 24TB-28TB range, with devices up to 36TB likely to become more widely available within the next one to two years. As device density increases, the CO2e/TB of SSDs is expected to decrease faster than that of HDDs.
If the decreasing GHG emissions intensity ratio trend for SSD continues alongside the accelerating SSD density advantage over HDDs, the data suggests that the SSD GHG embodied emissions per TB will soon achieve parity with HDDs.
Conclusion
Recent public statistics on the embodied CO2e per TB of HDDs and SSDs indicate the two drive technologies are much closer than what has been provided in the past. While the public analysis and data points are quite limited they do provide us with updated insight into the narrowing gap between the embodied carbon of HDD and SSD. In a device-to-device comparison, there is no disputing HDDs enjoy a slim but ever-narrowing advantage over SSDs in embodied emissions. Does this mean that SSD-based enterprise storage systems are worse for the environment than HDD-based ones? In particular, how does the stagnation in HDD throughput and IOPS impact the use of larger capacity HDD in enterprise data storage systems? We’ll look into that in Part 2 of this blog.
The Innovation Race
Reducing Risk and Navigating the AI Frontier for Future Success
Less, but Better
Learn how you can improve data center sustainability with Pure Storage.
