Data Storage Technologies Driving the Global Tech Race

Originally published 16-05-2018 | Revised June 2025

The data storage technologies market is growing strongly on the back of the emergence of new storage services that deliver access to increasingly larger and more valuable data sets. This expansion in big data continues to be driven by growth in files, distributed file systems, machine-generated telemetry and—new since 2022—training datasets for generative AI models. The on-demand nature of the cloud is not only changing how enterprises consume and pay for IT resources; it is also altering how storage providers deliver features such as de-duplication, replication, compression and encryption—often via API-first architectures and “storage-as-code” tools.

Global Players and the Geopolitics of Data

The two giants when it comes to big data and data storage remain the US and China. The US still leads in commercial data volumes, with companies like Meta (formerly Facebook), Google, Amazon and Microsoft running hyperscale infrastructures. According to industry analysts, data and data storage are now viewed by successive US administrations as a key strategic asset in the ongoing technology competition with China. Interestingly, many innovations continue to originate from small, fast-moving start-ups.

Spin Transfer Technologies, a California-based start-up, developed what it described as “a breakthrough technology” that dramatically improves MRAM speed, endurance, and feature size. This technology, called Precessional Spin Current (PSC), is a simple add-on to existing MRAM manufacturing processes. MRAM (Magnetoresistive Random-Access Memory) is a non-volatile memory technology that stores data using magnetic storage elements and continues to attract investment for edge applications requiring high endurance and low latency.

Simplified structure of an MRAM cell. By Cyferz – Wikipedia, Public Domain, Link

While MRAM offers advantages over long-established flash in terms of write speed and endurance, challenges remain around data retention and cost per bit. The PSC add-on was reportedly cost-effective at $1 per wafer, and STT has since licensed the technology to multiple MRAM vendors. As of 2024, MRAM based on this approach has entered sampling for embedded use in automotive and AI edge processors. STT has raised over $130 million in R&D funding to date, and while commercial-scale production was initially expected in 2019, viable products are only now reaching niche applications.

China’s AI Strategy and Expanding Data Repositories

Meanwhile, algorithm development in China has significantly accelerated. AI-focused companies and research institutes—heavily supported by government policy—are fuelling vast state-operated data lakes. President Xi Jinping’s AI 2030 strategy remains a cornerstone of national digital infrastructure planning. China is at the forefront of facial recognition technologies. For example, Shanghai-based start-up Yitu Technology, which previously won top honours in US benchmarks, has developed facial recognition engines capable of scanning databases containing over 1.5 billion images—systems widely adopted by law enforcement agencies.

China is also prioritising the advancement of deep learning, a more sophisticated evolution of conventional machine learning. According to McKinsey Global Institute’s latest report, deep learning now has the potential to add up to 15% in value to a company’s revenues—up from the 9% projected in 2018. The Chinese military is leveraging these tools as well, deploying swarm drones whose real-time imaging feeds contribute to petabyte-scale data processing pipelines, further driving demand for efficient and resilient storage architectures.

Next-Generation Storage Innovations from China

Some of the most intriguing storage developments continue to emerge from Chinese universities. As early as 2018, scientists at Northeast Normal University claimed to have developed a high-density storage medium capable of holding more than 1,000 times the data of a DVD within a 10x10 cm film.

Thanks to nanoparticle-based materials over 80 times thinner than a human hair, researchers led by Shencheng Fu demonstrated that such films could “holographically archive” data. Holographic data storage involves the use of lasers to write and read three-dimensional interference patterns in specially prepared materials, offering massive parallelism.

Because this method can write and retrieve millions of bits simultaneously, it greatly outpaces traditional optical and magnetic media, which process one bit at a time. Metal-semiconductor nanocomposites composed of titania and silver nanoparticles have shown promise as mediums for recording nanoscale holograms with extremely high spatial resolution. This innovation could help meet the long-term demand for high-capacity storage and may one day enable compact wearable devices capable of recording and displaying 3D imagery.

From Lab Discovery to Commercial Space Applications

In a 2024 paper in Optical Materials Express, Fu noted, “In the future, these new films could be incorporated into a tiny storage chip that records 3D colour information that could later be viewed as a 3D hologram with realistic detail. Because the storage medium is environmentally stable, the device could be used outside or even brought into the harsh radiation conditions of outer space.” His team has since formed a spin-off, HoloStor, to develop prototype drives for archival use in satellite imaging and scientific space payloads.

As global competition between the United States and China intensifies on the world stage, it’s increasingly clear that the accumulation, processing and storage of accessible data will play a pivotal role—both in visible infrastructure and behind the scenes. Whether through next-gen MRAM embedded in edge-AI silicon or nanoscale holography tested for orbital deployment, data storage technologies remain central to the future of computing, national security and global information flow.