3 Homomorphic Encryption Trends for 2025

By Jorge Myszne, CPO, Niobium Microsystems

As 2025 dawns, fully homomorphic encryption (FHE) is a privacy-enhancing technology (PET) on the cusp of going mainstream. FHE keeps data encrypted during computation so that it can't be read even if it’s intercepted. Hardware acceleration, VC investment, and new cross-industry consortia like FHETCH have gathered strong prevailing winds behind FHE, and the next two years will see the first commercial applications. In 2025, I expect to observe three trends as FHE adoption begins in earnest.

#1: Adoption Will Start on the Bleeding Edge With AI and Blockchain.

AI requires third-party computation on a massive, unprecedented scale, and that data is vulnerable during every query. Despite all the potential benefits of AI, many CSOs have put off adoption due to the security risk. AI providers need FHE to create an ironclad security solution for enterprise clients.

Blockchain is the other industry currently straining at the traces to implement FHE. Blockchain ledgers are public and immutable, which prevents certain desirable activities and makes it difficult for blockchain applications to comply fully with privacy regulations like GDPR and HIPAA. FHE would allow users to process transactions or execute smart contracts while keeping all personal identifiers private, enabling use cases like private DeFi or private auctions.

Related:Data Privacy Quick Reference Guide

Since the 2024 election swept several noted crypto enthusiasts into power in the U.S., we’re also in the middle of an undeniable blockchain resurgence. The crypto bull market will fuel a wave of new investment in blockchain technology, accelerating development efforts in an industry that’s already proven eager to embrace FHE.

#2: Law Enforcement Will Warm To Widespread Encryption.

Bodies like the FBI and the International Association of Police Chiefs have warned that encryption allows criminals to “go dark,” making it impossible for law enforcement agencies to obtain digital communications records even with a warrant. Lately, however, the FBI has also found itself promoting the use of encrypted messaging tools in response to “Salt Typhoon,” a massive telecommunications cyberattack attributed to Chinese actors.

Salt Typhoon makes clear that widespread encryption is a matter of national security. Perpetrators were reportedly able to steal communications metadata and even record exchanges involving U.S. government officials—and they were empowered by the fact that perimeter-based security cannot scale fast enough to keep pace with the billions of endpoints and attack surfaces inherent to global telecommunications. The data itself must be encrypted at every stage.

Related:The Impact and Future of AI in Financial Services

The threat extends far beyond telcos. Amazon has reported observing a 750% increase in attacks in 2024—up to nearly a billion hack attempts per day.  State and non-state actors work tirelessly to gain access to financial records, personal data, enterprise records, etc. Encrypting all this data would massively exacerbate the issue of “going dark,” but FHE offers some mitigation. Using FHE, regulatory bodies can perform compliance checks and audits without needing to view sensitive data as plaintext. Similarly, law enforcement can analyze a dataset for evidence of a suspected crime without exposing the suspect’s private life to scrutiny. FHE could prevent “digital stop and frisk” without impeding the ability to investigate serious crimes.

#3: It’s Going To Be a Group Effort.

We’re done with “security by obscurity.” Practical, scalable FHE solutions will emerge not from private skunkworks but from a foundation of collaboratively developed industry standards.

Closed development of security solutions often hides vulnerabilities until after they are widely deployed. Perhaps if Intel’s SGX had been subject to open design review, its vulnerability to Spectre- and Meltdown-style side-channel attacks would have been exposed before widespread implementation. The same could be said of the AMD Zen 2 Zenbleed exploit, or the Rowhammer vulnerabilities across multiple generations of DRAM. Open development and review encourage broader scrutiny, uncovering flaws faster.

Related:What Is a Sovereign Cloud and Who Truly Benefits From It?

What’s more, privately developed solutions can be unacceptable to many stakeholders. As a case in point, look at Google’s efforts to bootstrap new private browsing solutions. In 2019, Google launched the “Privacy Sandbox” initiative to replace third-party cookies with privacy-preserving technologies that would still allow targeted advertising. A year later, they announced plans to eliminate the use of third-party cookies in Chrome. Cookie phase-out was supposed to be complete in 2024, but UK antitrust regulators intervened to oppose the change. Why? Google’s competitors in the targeted ads business worried that the various schemes proposed under Privacy Sandbox would leave them more reliant on Google’s internal Chrome browser data. This led, in their view, to a serious risk that Google Ads would get preferential access to data. UK regulators agreed, forcing Google to reconsider. Privacy Sandbox continues to work with W3C to develop more open private browsing alternatives, but Google’s original plan to rapidly phase out third-party cookies has been consigned to the ash heap of history.

FHE hardware and software providers and users must work together to develop standards for hardware/software interoperability, implementation best practices, compute optimization guidelines, and more. This is part of the work currently being undertaken by FHETCH, among others. Open standards will help the industry build solutions that are more secure, broadly supported, and easier to implement and scale.

Despite the appetite of industries like AI and Blockchain for FHE solutions—and the urgency of the risks they face—these collaborative efforts are still in the early phases. Still, in 2025, I expect things to move fast—and as we begin to see real-world benefits in emerging use cases, mainstream compute applications will shortly follow.

About the Author

Jorge Myszne is Chief Product Officer for Niobium Microsystems, a developer of groundbreaking Zero Trust Computing hardware solutions. Jorge is a successful entrepreneur, having founded two start-ups that were later acquired and held technical and executive positions at Intel and Qualcomm. He holds a BSc and MSc degree in electrical, electronics, and communications engineering.