2026 Designing Storage Architectures for Digital Collections

The following post was authored by Vincent Coltellino from the Library of Congress. 

In 2007, the Library of Congress organized the first Designing Storage Architectures for Digital Collections (DSA) meeting. The purpose of that first meeting was to bring together storage architecture professionals to discuss the unique data storage requirements of the Library’s new National Audio-Visual Conservation Center (NAVCC). In the years since, the event has transformed into a venue for broader discussions of digital storage advancements, challenges, and solutions. On March 9-10, 2026, the Library hosted its 20th iteration of this collaborative (and now international) meeting.  

This year’s event attracted 172 of the world’s leading experts in data storage infrastructure, best practices, innovation, and future projections of the digital preservation ecosystem. The Library’s DSA organizers arranged a robust agenda to foster open discussion of current challenges in digital storage and proactive solutions to mitigate predicted future risks. Attendees represented a variety of government agencies, cultural heritage institutions, the digital storage industry, and academic and research organizations. 

The Library’s Chief Information Officer, Judith Conklin, welcomed the crowd and set the tone for the following two days of enthusiastic discussion. Library staff then provided their annual update on the Library’s current storage environment, data growth rates, and the pursuit of innovative digital storage techniques.  

Topics included the proliferation of digital data and the challenge to storage providers in keeping up with this demand; the energy and environmental considerations required to maintain the data storage ecosystem; the impacts of artificial intelligence (AI) on storage demand; issues and best practices from a variety of data stewards; and a series of technology updates from both novel and legacy data storage providers.  

While the presentations and subsequent conversations varied greatly, two core themes consistently resonated:

• How can the data storage community maintain the required level of service given the rapid data proliferation?
• How does the community address the demands of AI accessing data historically stored on high-latency media?  

While the former theme has been a consistent point of discussion through the years, this year was different. The group is now discussing the challenge as a current problem, rather than a problem for the future. It appears that an inflection point has been reached: data storage providers are unable to keep up with increased demand. This issue is observable in the marketplace through supply limitations and increasing cost per terabyte of storage. Many attendees had encouraging words for the emerging technologies now being developed to fill the gaps not met by existing media. 

Storing data in appropriate media based on frequency of access has long been discussed at this forum, and while this issue was still a topic of conversation, a new element was introduced: foundational and large language models accessing data at rest. With the use of AI ever-increasing, even data historically stored in “cold (less accessible)” media needs to be dynamic and accessible if the data is to be effectively used by AI models. This trend could influence data stewards to store their data in “hotter (more accessible)” energy intensive media. These hotter solutions are typically expensive and have higher carbon footprints.  

Although these two challenges were a big part of the conversation throughout the meeting, much of the discussion focused on solutions. Novel data storage media, such as high-density optical and molecular solutions, experienced more support than previous years, and the companies developing them are making headway toward enterprise-level commercialization. These solutions will hopefully fill some gaps in the marketplace in the coming years. Additionally, there were several discussions about assessing the energy requirements of a digital archive, using digital archives for disaster mitigation, and how the digitization of an archive can advance the historical and genealogical knowledge of a community.  

The DSA meeting consistently assembles a unique combination of experts that represent data storage providers, technologists, data users, and archivists to discuss the challenging topics surrounding the data storage ecosystem. That was no different in 2026, but this year’s forum carried a different tone. There was a greater sense of urgency from community members regarding long-predicted challenges that have now been realized, as well as challenges yet to be solved.  

How can data storage providers meet very high market demands due to uncontrolled data proliferation? Will data storage scarcity be a limiting factor in the AI boom? How can organizations address proper data stewardship and the energy requirements associated with it? While the answers to these questions are not yet solidified, the broader community collaboration fostered at this event is an essential step to finding the answers required to ensure a healthy future for the world’s data.  

For detailed information on the presentations and access to background reading, please visit the DSA 2026 repository.