A Brief Analysis of Fully Homomorphic Encryption (FHE): Expanding the Possibilities for Private Transactions

After listening to the Epicenter podcast interview with the CEO of ZAMA, I felt anxious again about my limited brain capacity. To put it simply, FHE is a more advanced encryption technology than ZKP (zero-knowledge proofs), capable of enabling data privacy across broader applications such as cloud storage, genetic prediction, biometric identification, and private crypto transactions. Coincidentally, VCs have recently been paying close attention to FHE, so here’s my take:

1) FHE stands for Fully Homomorphic Encryption, which allows computations on encrypted data in specific formats without decrypting it—thus preserving data privacy. It's a long-standing technical concept at a large scale on the internet. Recently, it has gained traction in the crypto space mainly because Fhenix Network integrated FHE functionality into Ethereum by launching the fhEVM solution. Any developer can now include FHE capabilities within transaction flows via pre-compiled binaries.

2) What new possibilities does this unlock? For example, in gaming scenarios, players’ cards could remain hidden while the platform still performs computations to advance gameplay—enhancing fairness; in DAO voting, whales could participate in governance without revealing their vote amounts, yet the protocol can still correctly tally results;

In private transaction scenarios, users can send fully encrypted transactions into the mempool without exposing sensitive details like transaction amounts; in regulatory compliance, governments could monitor cross-border fund flows and flag assets linked to blacklisted addresses for deduction—all without inspecting legitimate transaction contents. The potential of FHE exceeds that of ZKP by an entire tier.

3) What’s the difference between FHE and ZKP? A simple breakdown:

• ZKP solves the problem of consistent data transmission under encryption—allowing the sender to prove data authenticity to the receiver without revealing the actual data. This is a point-to-point encryption scheme, with lightweight variants like zk-SNARKs widely adopted in Layer 2 rollups;

• FHE, on the other hand, aims to solve the problem of performing further computations on encrypted data during transmission—a broader, many-to-many encryption framework. Its limited adoption so far stems from high computational overhead and low efficiency.

4) The company ZAMA developed the fhEVM solution, enabling FHE application in blockchain. Phenix Network takes a similar SDK-like integration approach, modularizing FHE as a pre-compiled process that can be added to various blockchains. Currently, it works best with OP-rollups and can even implement pipelines to selectively apply FHE only to specific transactions.

This means users on Layer 2 can choose whether to initiate transactions using FHE for enhanced privacy—slightly more expensive than regular transactions but offering greater security. This will inevitably attract market makers (MMs) and institutional users who prioritize transaction privacy.

Demand and use cases are clear—the key now lies in balancing efficiency, performance, and cost.

I believe private transactions will be the next refined trend following scalability solutions. The core challenges around scaling are no longer purely infrastructural; rather, missing pieces lie in regulation and market readiness for mass adoption. Privacy transactions, however, refine user segmentation, transaction types, and use cases through infra-level tech improvements—potentially achieving widespread adoption similar to ZKP.

We can envision a future powered by FHE where the experience gap between DEXs and CEXs narrows, Layer 3 privacy-focused app chains offer more secure and closed-loop trading experiences, games resolve on-chain randomness security issues, and regulators engage in a compliant yet non-intrusive manner. The imagination space is vast. That said, current TPS remains very low—only about 2–5 transactions per second, barely comparable to EVM mainnet levels. Reaching the tens of thousands of transactions per second seen in rollups will require a long journey ahead.

Note: My perspective here is primarily from the crypto angle. FHE also holds significant potential in internet and AI domains—readers interested may want to explore further.