Fujifilm plans to substantially expand production capacity for magnetic tape used in data storage. According to a Nikkei report on October 28, 2026, the company will invest ¥10 billion in its Odawara plant and aim to raise capacity-based shipments to five times the fiscal 2025 level by fiscal 2028. The goal is to meet growing demand for data storage, including AI-related demand. However, the "fivefold" figure refers to the total amount of data that the shipped tapes can store, so it does not necessarily mean the number of cartridges shipped will increase fivefold.
Fujifilm is increasing the storage capacity of each tape through improved recording materials while also strengthening its supply capability through plant investment. But even if storage capacity grows, data read speeds do not necessarily improve at the same rate. To consider how this expansion might affect data center operations and similar environments, the amount of data that can be stored needs to be evaluated separately from the time required to restore the data you need.
¥10 billion investment to expand shipment capacity fivefold
According to Nikkei Asia's report, Fujifilm will invest ¥10 billion in its Odawara plant and increase magnetic tape shipment capacity to five times the fiscal 2025 level by fiscal 2028. It would be the largest production buildup in the tape business in about 20 years. The background is growing demand for data storage accompanying the spread of AI and other trends.
The key point here is the difference between shipment capacity and the number of units shipped. Shipment capacity is determined by the number of cartridges shipped and the recording capacity of each one. For example, if the capacity per cartridge rises, total shipment capacity increases even if the number of units shipped stays the same. The fivefold target therefore cannot be interpreted as a plan to increase the number of cartridges manufactured or the amount of production equipment fivefold. Revenue is also affected by product pricing, so it will not necessarily grow at the same rate as shipment capacity.
At this point, the absolute shipment capacity for fiscal 2025, which serves as the baseline, has not been disclosed, nor has a breakdown of future supply by capacity tier. It is also unclear how much of the fivefold increase will come from larger capacity per tape and how much from more units shipped. Even so, the planned large-scale investment in the plants that support supply, in addition to higher-capacity products, is an important factor in considering the business's direction.
It is also worth checking market-wide trends. According to data released by the LTO Program on September 1, 2026, LTO tape shipment capacity in the first quarter of 2026 grew 57% year over year. Meanwhile, full-year 2025 shipment capacity was 160.3 EB (exabytes), down about 9% from 176.5 EB in 2024. Continued LTO-9 shipments and the ramp-up of LTO-10 contributed to the recent quarterly recovery, but how much of it is attributable to AI demand has not been disclosed. The 57% quarterly increase for the overall market and Fujifilm's single-company goal of a fivefold increase by fiscal 2028 differ in both scope and comparison period, so the two cannot be directly linked.
Magnetic material and film technology behind 40TB capacity
Fujifilm has already launched a 40TB (uncompressed) product compatible with the 10th generation of LTO (Linear Tape-Open), the data storage tape standard. Announced on December 22, 2025 and released in January 2026, it exceeds the capacity of the 30TB version released in June 2025. This means tape capacity itself has been increasing separately from the production expansion plan reported now.
The move to 40TB relies on combining higher magnetic recording density with longer tape. Fujifilm raised areal recording density by adopting a fine-particle hybrid magnetic material and used its proprietary thin-layer technology to extend the tape length by about 30% compared with the 30TB version. The 40TB tape is 4.0 μm thick and 1,337 m long. In other words, the capacity increase comes not only from being able to record more information in the same area, but also from fitting a greater length of tape into one cartridge.
In the specifications published by the LTO Program on November 12, 2025, aramid is cited as a base material that enables thin, smooth tape and longer lengths. The magnetic material increases data recording density, while improvements to the base film allow longer tape to fit in the cartridge. Combining the two raises recording capacity without changing the cartridge size. Plant investment, on the other hand, is aimed at supplying more of this media. Both factors push up capacity-based shipments, but without separating their respective contributions, it is impossible to judge how much production capacity will actually expand.
The specifications for the 40TB version also list a capacity of "up to 100TB," but that figure assumes the data can be compressed at a 2.5:1 ratio. The physical recording capacity is 40TB uncompressed, and 100TB depends on how well the stored data compresses. Depending on the type of data, such as already-compressed images and video, the assumed ratio may not be achieved, so 100TB cannot always be stored. When estimating storage capacity, uncompressed capacity must be clearly distinguished from capacity that assumes compression.
The 40TB version can be used in the same LTO-10 drives as the 30TB version. However, higher media capacity does not improve the drive's own read/write performance. Moreover, even though the capacity increase from 30TB to 40TB is known, it is not clear how much of Fujifilm's future shipments the 40TB version will account for. As a result, the number of units to be manufactured or the rate of equipment expansion cannot be simply back-calculated from the fivefold shipment capacity target.
Reading all 40TB takes about 28 hours in theory
When evaluating larger-capacity magnetic tape, the time needed to read back stored data also deserves attention. Fujifilm's 40TB LTO-10 tape has a maximum uncompressed transfer rate of 400MB/s. If 40TB of data were read continuously at that speed, transferring all of it would take about 27.8 hours in theory.
The calculation uses the uncompressed capacity of 40TB and the maximum uncompressed transfer rate of 400MB/s from the 40TB product specifications published by Fujifilm. Treating both TB and MB as decimal units and assuming a constant transfer rate, the time required is calculated as follows.
\[ \frac{40\times10^{12}\ \mathrm{byte}}{400\times10^6\ \mathrm{byte/s}} =100{,}000\ \mathrm{s} \approx27.8\ \mathrm{hours} \]
40TB is 40 trillion bytes and 400MB/s is 400 million bytes per second, so the transfer time comes to 100,000 seconds, or about 27.8 hours. However, this calculation assumes ideal conditions in which the maximum transfer rate is sustained without interruption. In practice, cartridge loading, seeking to the read position, fluctuations in transfer speed, and data verification or retries would occur, so the actual restoration work could take longer. The figure of about 27.8 hours is a theoretical minimum transfer time derived from official specifications, not a value measured in an actual recovery operation.
Fujifilm lists the maximum transfer rate for this product with compression at 1,000MB/s. However, when calculating the transfer time for the 40TB uncompressed capacity, the maximum uncompressed speed of 400MB/s must be used. Combining the uncompressed capacity with the compressed transfer rate would mean the comparison conditions do not match. Estimating actual transfer times also requires considering the data's compression ratio and the drive's operating conditions.
Of course, restoring data does not always require reading the full 40TB. If only some files are needed, the amount of data to read will be smaller. However, magnetic tape is a recording medium based on sequential access, so the time to move to the position where the needed data is recorded must also be taken into account. If capacity per tape grows while transfer speed stays the same, the time to read the entire contents grows longer. Being able to store more data is a separate matter from being able to restore more data within a limited time.
Rising demand for long-term storage, and the remaining restoration challenge
Fujifilm cites as an advantage of magnetic tape the ability to store data in an "air-gapped" environment that is physically disconnected from the network. Making data inaccessible directly over a network can reduce the risk of data loss from cyberattacks such as ransomware. For uses that involve storing large volumes of data over long periods, this characteristic is one reason to choose magnetic tape. However, simply adopting tape does not guarantee data safety; operational design, including storage methods, access management, and backup verification, is also necessary.
For companies and data centers that actually deploy magnetic tape, what matters is not only total storable capacity. They must also consider which data needs to be restored, and how quickly, when a failure or data loss occurs. Even with high-capacity tapes, if the time needed for restoration exceeds the acceptable range, operational requirements may not be met. It is important to evaluate real-world restoration performance, including the number of drives, parallel read configurations, data placement, and cartridge management methods.
If the ¥10 billion investment proceeds as planned, Fujifilm will be able to expand its supply capacity to meet future data storage demand. However, the fivefold shipment capacity target alone does not reveal which capacity products will be supplied, when, and in what quantities. Going forward, the focus will be on whether product-by-product production plans and supply start dates become concrete, and on whether users of higher-capacity media can build systems that restore needed data within required deadlines. Whether expanded tape supply actually strengthens data storage infrastructure must be judged not only on media capacity and production volume but also on read performance and operational design.