Darkpool Deep Dive #4: Fully Homomorphic Encryption Explained

Welcome to the fourth edition of the Singularity Darkpool Deep Dive Series, where we will take a deeper look into the architecture, vision, and innovations powering Singularity’s institutional-grade on-chain Darkpool orderbook.

In each edition, we unpack a specific layer of the Darkpool - from how we prevent MEV and frontrunning to the role of cutting-edge cryptography like ZKPs, FHE, and MPC.

Built to offer compliance without compromise, Singularity delivers confidential, capital-efficient execution tailored for institutions ready to trade on-chain - privately, securely, and without leaving a trace.

‍

Last edition, we discussed How ZKPs Power Confidential Trading in Singularity and why it's essential in DeFi. 

‍

This week, we're diving into another core technological component of our Darkpool, Fully Homomorphic Encryption (FHE).

What Is Fully Homomorphic Encryption?

FHE allows you to perform computations on encrypted data - and get encrypted results - without ever decrypting the input.

In simpler terms:

• If Alice encrypts two numbers and sends them to Bob,
  
  
• Bob can add or compare them without knowing what the numbers are,
  
  
• And return an encrypted result that Alice can decrypt herself.

Why FHE Is Critical to Private Trading

For Singularity’s Darkpool to maintain complete confidentiality, we must ensure:

• The asset pair, amount, and price of any order remain hidden
  
  
• The system can still compare encrypted values to match compatible orders

FHE allows our book nodes (matching engines) to process encrypted orders as if they were plaintext, without ever exposing sensitive information on-chain or to validators.

How Singularity Uses FHE in Practice

Here’s a step-by-step of how FHE fits into our matching workflow:

1. Encryption:  A trader creates an order and encrypts the asset, price, and amount using a public key generated by distributed book nodes.
   
   
2. Order Submission: The encrypted order is sent to the Singularity Darkpool - fully private, yet ready for matching.
   
   
3. Encrypted Matching: A book node performs matching logic on the encrypted values:
   
   
   
   • Is the asset pair compatible?
   • Is the bid price ≥ ask price?
   • Does the size meet minimum thresholds?
     
     
4. No Decryption Required: Even during these comparisons, no party sees the decrypted values.
   
   
5. MPC + FHE for Result Decryption: Once a match is found, Multi-Party Computation (MPC) is used to decrypt only the outcome (match/no match), without revealing trade specifics.

Why Not Just Use ZKPs?

ZKPs are great for proving facts. But:

• ZKPs prove that a statement is true (e.g., "I have enough tokens").
• FHE actually executes logic on encrypted data (e.g., comparing prices or matching amounts).
  
  

Together, they serve different but complementary roles:

• ZKPs = proof of validity
  
  
• FHE = private computation
  
  

A New Model of Confidential Execution

FHE enables a new class of confidential trading infrastructure - where:

• Orders remain encrypted end-to-end
• Matching happens invisibly
• Execution is auditable, but not traceable

This cryptographic foundation allows Singularity to support true dark liquidity - not just hidden intent, but hidden everything.

Next Up: Multi-Party Computation

In Part III, we’ll break down how MPC enables key management and distributed decision-making - and how it ensures that no single node can ever compromise the system.

Thanks for reading our fourth edition of Singularity’s Darkpool Deepdive series.In the subsequent deep dives, we’ll take you behind the curtain to explore additional core components that power Darkpool.