In the last BlogBytes, see it here, we discussed the costs associated with storing AI data as described in the paper, Powering the AI Era: How LTO Tape Drives Unprecedented Economic and Strategic Business Value, by Solutions North Consulting LLC. In this BlogBytes, we will continue reviewing the paper, and look at other topics, like how AI data can be protected for greater cyber resiliency, and the latest news on hard disk drive (HDD) markets which may have implications for LTO technology customers. Let’s go!
EGRESS BUILDS QUIETLY.
To set the stage, let’s review a key section of the new Solutions North paper. This explains how artificial intelligence (AI) models are based upon huge amounts of information which some users may be storing in the cloud for mid-to-longer term retention purposes. However, to retrieve information from cloud storage for AI processing, organizations may incur large egress fees. The report explains that a nominal “2% monthly data egress can trigger a massive 164% spike in total storage costs.” Egressers beware!
A cloud4c.com article describes what can happen to an organization with egressed data: “Egress, though, builds quietly in the background across backup syncs, API responses, disaster recovery tests, and cross-region replication, until a budget review or an unexpected invoice forces someone to actually look at it closely.” It explains that “They grow because the business grows. New regions, more integrations, more users, larger feature sets, all of it generates outbound data. Ingress is free. Egress is not.”
CAN THE FEES BE UNBEARABLE?
In an analyst paper from earlier in the year, The Hidden Economics of Cloud Data Repatriation, Jon Brown of Omdia elaborates on cloud storage fees: “While cloud providers do offer a lower cost of entry for data storage, the economics of storing massive amounts of data in the cloud can be punitive.” Brown explains that, “At the beginning, your monthly cloud storage fees might be $3,500 for 500 TB of data. That seems affordable compared to infrastructure investments. But this is the bill every month. And as your data grows, so do the fees … until they become unbearable.”
THE HDD DILEMMA.
Organizations need vast amounts of AI data to inform critical business decision-making. They need to balance egress fees, data storage access times, data storage costs, and cyber resiliency. Huge AI data lakes, in part, have caused a scarcity of hard disk drives. In its Q3 earnings, Seagate said its nearline capacity was “largely allocated through CY27,” with strategic discussions already reaching into 2028 and build-to-order contracts in place through fiscal 2027. CEO Dave Mosley said “the demand keeps going up” and presented higher pricing as part of a continuing strategy supported by constrained supply, higher-capacity products and multiyear customer commitments.
For its part, Western Digital’s clearest availability comment came from CFO Kris Sennesael who confirmed that “Manufacturing lead times are now about a year.” WD now has LTAs extending into calendar 2029 and is giving strong indications of a structurally tight market: customers are reserving capacity years ahead, while incremental demand beyond those commitments may carry a pricing premium.
The two companies are giving consistent feedback to the market. Supply is heavily committed. Neither company is responding by aggressively adding HDD unit capacity. Both are relying on more terabytes per disk and per drive. Pricing power is strengthening: Seagate recorded mid-single-digit growth in data center revenue per terabyte, while WD reported a 9% increase. The most significant takeaway is that HDD capacity is increasingly being sold through forward commitments rather than treated as a readily expandable commodity supply. This suggests that customers needing significant nearline capacity should expect longer planning horizons, firmer volume forecasting and continued upward pressure on price per terabyte. All of these trends strongly support reevaluating the role of tape to help organizations manage spiraling AI data growth.
This is underscored by the Solutions North paper, which explains that “More than 70% of enterprise data will not have any enduring need for the performance of SSDs and HDDs, but will have greatly expanding needs for low-cost sustainability, immutability and security.” The paper further notes that the bottom line is “Data protection and durability at the lowest cost of ownership.” The analyst report illustrates that LTO tape storage has been shown to be the lowest cost form of data storage by up to 5 times vs disk or cloud environments and can provide offline data protection.
DO YOU HAVE OFFLINE DATA PROTECTION?
Offline tape data storage and protection can be the key to providing big data for AI analysis. Tape just might be the ideal alternative to HDDs because the media is widely available, easy to scale, and offers high storage density at reduced costs. If you don’t have the data, and lots of it, AI can’t give the optimal results.
Meanwhile, the Solutions North paper illustrates that more than simply providing the raw storage capacity for massive data lakes, LTO technology can also deliver the critical data security that today’s AI intellectual property requires. It emphasizes the essential reality that “there is no ‘modern data center’ without tape. It is the lifeline for cyber resiliency: the ultimate air-gapped defense.”
See the complete Solutions North paper here.
