- Seagate study claims that security and storage are at the top of the agenda for AI infrastructure
- Energy is a distant last, preceded by LLM viability and regulations
- Debates about AI energy consumption will continue until the compromise is met
AI energy consumption is becoming an increasingly hot topic, with industry stakeholders and critics voicing concerns about the technology’s environmental impact.
But a recent study from Seagate points to more pressing concerns for IT leaders, arguing that energy consumption is ranked at the bottom of the agenda behind regulatory considerations, the viability of LLMs and network capacity.
In particular, security and storage were among the top focuses for business leaders looking ahead, with nearly two-thirds (61%) of respondents who predominantly use cloud storage to host AI workloads saying their cloud-based storage will increase by over 100% in the next three years.
Cost-effective storage is key
This increased focus on AI adoption is expected to lead to an increase in demand for data storage, with hard drives emerging as the “clear winner,” said Roger Entner, founder and principal analyst of Recon Analytics, which conducted the study.
“The survey results generally point to a coming increase in demand for data storage,” he said. “When you consider that the business leaders we surveyed intend to store more and more of this AI-driven data in the cloud, it appears that cloud services are well positioned to ride another wave of growth. “
A key factor in this push is the cost-effectiveness of hard drives, the study found, which offer better scalability and improve cost per unit. dollar-per-terabyte.
Another contributing factor to the appeal of hard drives is data storage, the study found. Organizations that embrace AI typically retain data for longer periods of time to train and optimize AI models.
This long-term data retention practice plays a critical role in ensuring accuracy when training the models, with 90% of respondents already using AI believing that keeping data for longer helps improve results.
“Since the vast majority of survey respondents say they need to store data for longer periods of time to improve quality results of AI, we are focused on a real density innovation needed to increase the storage capacity of each plate in our HAMR -based hard drives,” said Entner.
“We have a clear path to more than double storage capacity per plate over the next few years.”
