AFHE Technology

Product Suite

AFHE is not a research project — it is infrastructure deployed through a stack of products addressing specific market needs, from retail traders to institutional desks.

Aura Shield: Encrypted Swap Execution

Aura Shield protects swap intents from submission through execution in two phases. Phase 1 (current): Users submit swap intents encrypted client-side using AFHE. The encrypted intent is transmitted to the Aura Relayer operating inside a Trusted Execution Environment (AWS Nitro Enclave), where it is decrypted, built into a transaction using the Jupiter SDK, and submitted as a Jito bundle bypassing the public mempool. The confirmed transaction on Solana is a standard transaction — validators process it normally. Phase 1 protects the intent in transit and during execution, not the on-chain state at rest. Phase 2 (planned): Full homomorphic swap execution where AMM invariants are verified on encrypted state and settlement occurs with encrypted state transitions. Phase 2 requires the FHE Coprocessor network and represents the transition from hardware-secured (TEE) to cryptographically-secured privacy.

Why it matters: MEV protection is the single most requested feature from institutional Solana traders. Aura Shield transforms Solana from a transparent execution venue into a dark pool equivalent with on-chain settlement guarantees. Revenue model: fee per encrypted swap (basis points on notional value).

Confidential Token Standard

The Confidential Token Standard extends Solana's SPL Token standard to support encrypted balances and confidential transfers. Any SPL token can be wrapped into a confidential version where balances and transfer amounts are encrypted. Token balances are stored as AFHE ciphertexts on-chain. Transfers are verified homomorphically (sufficient balance check, overflow protection) without decrypting. The standard is designed for compatibility with existing Solana wallet infrastructure (Phantom, Backpack) through a thin client-side encryption layer, though wallet integration will require coordination with wallet development teams.

Why it matters: This is the on-chain equivalent of omnibus accounts in traditional finance — institutions can hold and transfer tokens without revealing positions to the market. Revenue model: protocol fee on confidential token operations.

FHE Coprocessor

The FHE Coprocessor is a generalized encrypted computation layer that any Solana program can call. Developers can add privacy to existing protocols without rebuilding from scratch. Solana programs submit encrypted computation requests to the AFHE coprocessor network via account-based request submission. Coprocessor nodes — GPU operators running AURA Compute Node software — pick up requests, perform AFHE computation off-chain, and return encrypted results with proofs of correct computation. Lookup tables are stored off-chain by coprocessor operators with on-chain commitments for verification, avoiding the prohibitive cost of on-chain storage for large LUT datasets.

Use cases include private voting for DAOs, sealed-bid auctions for NFTs and tokens, confidential lending with hidden collateral ratios, and private orderbooks with invisible limit orders.

Why it matters: The coprocessor model creates a platform effect — every Solana protocol that integrates AFHE privacy becomes part of the network's value proposition, driving demand for AURA tokens. Revenue model: compute fees per coprocessor call (paid in AURA tokens).

FHE Mining: Distributed Table Generation

FHE Mining decentralizes the precomputation of AFHE lookup tables across a network of GPU-equipped miners compensated in AURA tokens. Table generation tasks are posted by the protocol when new operations or parameter sets are needed. Miners compute table segments using GPUs (table generation is embarrassingly parallel). Computed tables are verified through sampling-based probabilistic checks. Miners receive AURA tokens proportional to tables contributed.

Mining rewards decrease over time as the core table library matures. Ongoing demand comes from new operations, parameter updates, and protocol upgrades. GPU miners from other networks can redirect hardware to FHE table mining.

Why it matters: FHE Mining creates a hardware-backed economic moat. The table library is a compounding asset — each new table makes the network more capable, attracting more developers, which drives demand for more tables.