New research confirms that AI is now the leading driver of storage growth. Yet, balancing AI readiness and performance against more traditional priorities can be a challenge. How does your organization stack up?
There’s no letting up on the demand for more data center storage, with artificial intelligence (AI) and machine learning (ML) workloads emerging as the leading growth driver. It’s time to ask, “Is our data center ready for AI storage demands?”
Organizations have many options for meeting their growing storage needs: cloud, hybrid cloud, hard disk drives (HDDs), solid-state drives (SSDs), and tape archival solutions. That can make it challenging to come up with the optimal solution as companies try to match their strategic priorities with investment criteria, which may not always perfectly align.
Unsurprisingly, the demands of AI and ML workloads are increasingly driving storage growth, now consuming an average of 24% of storage infrastructure, according to a February 2025 survey conducted by Foundry on behalf of Western Digital. This research included 109 decision-makers in IT-related roles, data and business intelligence, R&D, and executive management. The respondents were almost evenly split between midsize organizations (500–2,499 employees) and large enterprises (2,500+ employees).
Participants could choose multiple reasons for their increased storage needs. Their most-cited drivers include:
- 68%: AI/ML workloads
- 54%: Expansion of private cloud and hybrid cloud environments
- 45%: High-resolution content
- 42%: Internet of things and edge computing
- 39%: Compliance and data retention policies
As they strive to balance available options with top priorities and bottom-line considerations, participants said, they are increasingly relying on HDDs for both fast and mass storage. For example, 82% of the survey respondents expect to increase HDD investments over the next two years due to AI adoption — of those, 28% are forecasting “significant” increases.
Priorities and objectives
When the survey participants were asked about the decision-making process for storage investments, all factors seemed vital:
- 90%: Longevity and durability of devices
- 86%: Total cost of ownership (TCO)
- 82%: Energy efficiency and power savings
- 72%: Compliance with sustainability goals
- 66%: Availability of trade-in or refurbishment programs
- 65%: The vendor’s commitment to environmental, social, and governance (ESG) initiatives
- 62%: Recyclability and circularity of storage components
The survey further reveals that there is no one-size-fits-all approach to the TCO conversation. Respondents cited many factors as being critical to TCO, including storage density, power and cooling requirements, reliability, performance, and acquisition costs. All of these considerations should be taken into account, from both a CapEx and an OpEx perspective.
When the respondents were asked about storage vendor selection, their priorities changed slightly. Although they still highly ranked the need for reliability, durability, and performance, other issues bubbled up — most notably AI and analytics readiness (47%) and scalability (42%).
Although energy efficiency and sustainability remain crucial, they were more commonly cited by midsize organizations as being extremely or very important (88%) to their storage provider selection, compared with 74% of enterprise-size organizations.
Growth comes with growing pains
Storage expansion is causing major pain points. On average, organizations said their storage demand grew 27% over the past year. Yet, 51% of the respondents said they had experienced increases of 25% to 50% or more. No one in the survey reported decreases in storage demands.
The pain arises predominantly from the costs associated with storage expansion — particularly for midsize companies, of whom 60% rated it as top challenge, compared with 43% of the large enterprises. Other pains include:
- Meeting AI and analytics performance demands
- Security and compliance concerns
- Data access speeds and latency issues
- Managing unstructured data growth
- Migration challenges (on-premises to cloud and vice versa)
- Vendor lock-in or lack of flexibility
With many companies still in the relatively early stages of deploying AI, growing data demands will almost certainly amplify new or evolving storage challenges. Performance for AI and analytics workloads is table stakes, as are data retention and compliance requirements. Yet, there’s also the issue of ensuring and paying for high-speed data access and retrieval.
AI applications are causing a surge in data generation and usage. Users need to access and process AI workloads at high speed and with low latency, creating new obstacles to evolving storage architectures. As organizations accelerate their adoption of AI, the demands on enterprise storage are changing rapidly, and decision-makers must consider how the scale, speed, and complexity of AI applications will impact their infrastructure.
The survey found that most companies use two or more storage solutions for AI workloads. Midsize companies rely more on hybrid cloud and HDD solutions, whereas larger companies have greater reliance on hybrid cloud and cloud object storage. Midsize organizations are also more reliant on SSDs than are enterprises. Each workload is unique, and success depends on aligning the right storage technologies — whether HDD, high-performance flash, or cloud — with the business outcomes that organizations are seeking from AI.
Critical needs and challenges
Reliability, scalability, power, cooling, and overall costs contribute to the strategies and decision-making for enterprise storage. The impact of AI overlays all these concerns and increases pressure for organizations to develop storage strategies that are high-performing but also intelligently tiered, scalable, and cost-aware.
Data in the data center is not homogeneous; different applications and data types have varying access requirements such as frequency, latency, and cost sensitivity. This necessitates a tiered storage approach that is cost-effective and flexible enough to accommodate new and evolving AI use cases.
Meeting these challenges demands an approach that combines next-gen storage technologies with deep architectural flexibility. HDDs continue to offer opportunities for data center “warm” storage opportunities where the vast majority of data lives. For some organizations, tape still has a role for “cold” archival and backup data, due to low cost and infrequent access.
Aligning storage with the specific nature of the workload is key to unlocking AI’s full potential while optimizing total cost of ownership.
Whether your storage priority is AI readiness, sustainability, TCO, or performance, Western Digital can help ensure that you meet your goals. Click here for more information.
