A new paper is available at the LTO.org website, Powering the AI Era: How LTO Tape Drives Unprecedented Economic and Strategic Business Value, by Solutions North Consulting LLC. The paper discusses how to manage AI-driven analytics thirst for data in the modern era. The paper also covers a number of AI topics, from AI data resources and usage to the economics of AI data reserves. This BlogBytes article will review key data storage costs associated with AI analytic projects, as gleaned from the paper. Here we go!
AI LIKES LOTS OF DATA
The paper explains that about two-thirds of I&O leaders are very concerned about the costs associated with running artificial intelligence. Typically, a lot of data is needed to achieve acceptable results in many AI modeling projects. Of course, that can depend on the project. A nexocode.com article states that, “The more complex the problem is, the more training data you should have. The number of data samples should be proportional to the number of parameters. According to the so-called rule of 10, often used in dataset size estimation, you should have around 10 times more data samples than parameters.” Can you have too much data? That is possible, but as the article points out, “It’s much more common to struggle with the lack of data than its excess – but it can happen. In general, it’s literally impossible to have too much quality data.”
THE HIDDEN COST OF GROWTH
As noted, many AI projects require lots of high-quality data. Storing that data on disk or in the cloud can be costly. The Solutions North paper notes that, “Artificial intelligence models are massive consumers of data, and attempting to house all foundational data, ingest and historical telemetry, and historical checkpoints on high-performance NVMe, flash, or HDD storage is simply financially unsustainable.”
Solutions North explains that if data is egressed from the cloud, there can be a cloud egress tax: “Cloud storage may look cheap initially, but just a 2% monthly data egress can trigger a massive 164% spike in total storage costs. Modern tape solutions never have egress charges.”
STORAGE COST COMPARISONS
Solutions North emphasizes, “AI-driven analytics and real-time processing propel businesses into the Exabyte era,” and that “Modern tape solutions have evolved into integrable, AI-driven architectures. LTO tape technology has emerged as the silent, highly efficient engine of the world’s most modern and capable [data storage] architectures.” The paper states that “LTO has proven to outpace the industry in scaling capabilities while keeping the cost-per-Gigabyte at an absolute minimum.” The following set of graphics from the paper compares TCO for LTO-10 tape with disk and cloud to store three different petabyte capacities over a 7-year period:
IS IT TIME TO REPATRIATE DATA?
Due in part to cost, many firms are repatriating data to on-site premises. What is data repatriation? Puppet.com explains that “Cloud repatriation is the process of moving apps, services, and data off of public clouds (like AWS, Azure, or GCP) and back to data centers, on-premises, private cloud, or a hybrid setup. Companies choose cloud repatriation to reduce cost, improve privacy, and meet changing business needs.” The changing needs include cold data necessary for long-term data protection and to feed AI analytics.
The Solutions North study notes that the business case for repatriating the data to tape storage is undeniable:
- ROI: Most organizations see a full return on investment from tape in less than 24 months.
- Cyber Resiliency: In an era of rampant ransomware, tape provides the only true physical air gap.
- Data Control: Demanding governance and regulatory requirements are pushing data back on-premises.
The paper explains that “Tape is not merely a storage medium; it is an organization’s ultimate line of defense—immutable, encrypted, and offline.” See the complete Solutions North AI paper here.
1LTO.org 2026 TCO tool, Tape data on LTO-10, 7-year retention, AWS Glacier, 1% per month egress, Bulk Retrieval, 15% contractual discount level.
2Data retention of 7-years, LTO-10 Tape technology, 30TB HDD at May 2026 average price, managed service provider Deep Archive storage tier with zero retrieval or egress charges. Tape and HDD cost includes the cost of repatriation retrieval or egress charges.
