Reinventing Data Storage: Swiss-Greek Inventor Selected as Finalist for the European Inventor Award 2026 - EPO
12-May-2026
|
Source : The European Patent Office (EPO)
|
Visits : 364
• Evangelos Eleftheriou and his team developed key technologies that improved how digital data is stored, read and processed
• His work spans magnetic storage, flash memory and in-memory computing, enabling faster and more energy-efficient data systems
• The Swiss-Greek inventor is a finalist in the ‘Industry’ category. The winners will be announced during the Award ceremony on July 2, 2026, in Berlin
• Public voting for the Popular Prize opens today and will be running until the ceremony on July 2, 2026
MUNICH - According to the European Commission, data centers are a vital infrastructure supporting cloud computing, AI and digital services, yet they already account for around 1.5% of global electricity consumption, with demand projected to more than double by 2030. As data volumes continue to grow, improving how information is stored and processed has become increasingly critical. Evangelos Eleftheriou and his team have contributed to this challenge through a series of advances in digital storage technologies over many years, from magnetic recording to in-memory computing. Eleftheriou has been selected as a finalist in the ‘Industry’ category of the European Inventor Award 2026 by an independent jury, according to the official website of the European Patent Office (EPO).
Making dense data readable and reliable
Modern digital storage depends on packing vast amounts of data into extremely small physical spaces. However, as data signals are placed closer together, interference increases, making it harder to retrieve information accurately. Eleftheriou’s early breakthrough, noise-predictive maximum likelihood (NPML) detection, addressed this challenge by modelling and predicting signal interference rather than treating it as background noise. This approach became a key technology in advanced read channels for high-density magnetic recording.
Building on this foundation, his work at the time at IBM Research extended to magnetic tape systems, which were essential and continue to play an important role in large-scale data archiving. His team developed methods to improve positional accuracy, error correction and synchronization, allowing systems to continue operating even when data is degraded or partially damaged. Further innovations in flash memory architecture ensured that slow background processes do not interrupt active operations, improving performance consistency.
"You may not think of magnetic tape as a current technology, but it is very much alive. When large organizations need to ensure that their data is preserved reliably, they store vast quantities on tape. Tape remains a critical backbone of large-scale data archiving,” said Eleftheriou.
Processing data where it is stored
Moving information between memory and processors has become a major source of delay and energy consumption. Eleftheriou’s more recent work at his company, Axelera AI, focuses on in-memory computing (IMC), a technology that enables certain computations to be performed directly within memory structures.
This approach is particularly relevant for artificial intelligence applications, where large datasets must be processed efficiently and, in many cases, in real time. By minimizing data movement, IMC enables faster and more efficient computing systems.
"A solution that requires a fraction of the power of a competing system is, all else being equal, a strong advantage. But it is more than a commercial argument. We cannot afford to spend energy unnecessarily on computation and simply pay the price later,” said Eleftheriou.
Eleftheriou’s work spans multiple generations of storage technologies and reflects a broader shift in computing, from storing data efficiently to processing it efficiently at scale. His contributions have supported advances in modern digital infrastructure and continue to influence the future of data-intensive technologies.
Evangelos Eleftheriou is one of three finalists in the ‘Industry’ category of the European Inventor Award 2026. The other ‘Industry’ finalists are the Italian inventor Giuseppe Crippa and his team for their advanced probe card production method and the Greek-Swedish scientist Angeliki Triantafyllou for an enzymatic process that improves oat milk. The European Patent Office will announce the winners during a livestreamed ceremony from Berlin on July 2, 2026. In addition to the four award categories, the Popular Prize will be decided through a combined vote by the public and the independent jury. Public voting opens on May 12, 2026 and will be running until the ceremony on July 2, 2026.