Privacy & Confidential Execution

Why Privacy Matters for Portfolio Managers

DeFi today is transparent by default, which is powerful, but often limiting.

In particular, vault management of public, blockchain-based DeFi products is one of the most promising business models for the next generation of finance. However, its scalability and long-term sustainability are limited by a fundamental issue: the need for portfolio managers to protect their alpha.

In today's DeFi environment, vault tokenization standards expose real-time portfolio compositions and trading behaviors. Copycats and freeloaders can easily monitor, reverse-engineer, and replicate strategies without bearing the associated research and execution costs.

Without privacy, managers interested in avoiding liquidity fragmentation are forced into an impossible trade-off between:

• Sharing all their moves and risking the loss of their edge, or
• Operating off-chain, undermining the auditability and composability of DeFi.

If DeFi is to scale professional portfolio management, confidentiality must become a first-class primitive, just as composability and security already are.

Our Commitment to Onchain Privacy

Through innovations like confidential smart contracts, Orion enables:

• Protection of proprietary strategies without sacrificing onchain composability;
• Private portfolio management with verifiable, encrypted performance;
• Auditability for users and LPs without exposing sensitive manager behavior: vault strategies remain private, with holdings and trades encrypted, while performance remains auditable.
• MEV protection, safeguarding trade execution against frontrunning and sandwich attacks.

By integrating confidential computing natively into our architecture, Orion brings institutional-grade privacy to onchain finance, unlocking a new era of scalable, professionalized, and competitive portfolio management.

Confidential Smart Contracts: A Key Enabler

Orion integrates advancements in cryptography¹ to power private vault strategies and encrypted performance tracking. This integration enables the execution of confidential smart contracts on encrypted data, ensuring both data privacy and composability within blockchain environments.

Transaction inputs and onchain states are encrypted, ensuring that sensitive information remains confidential, while fully onchain, non-custodial and verifiable by anyone.

Technical Overview

Let:

• $N$ be the number of active privacy-preserving portfolios.
• $M$ be the number of distinct assets (tokens) active in the netted portfolio.
• $a_{i,j}$ denote the (encrypted) amount of asset $j$ held by portfolio $i$, where $i \in \{1, \dots, N\}$ and $j \in \{1, \dots, M\}$.

Each portfolio $i$ is defined as a vector:

$$a_i = [a_{i,1}, \dots, a_{i,M}] \quad\forall \medspace i \in \{1, \dots, N\}$$

where entries are encrypted amounts associated with a whitelisted investment universe shared by every portfolio (i.e., the union of unique symbols across all portfolios).

Let:

• $A_j$ denote the total (observable) amount of asset $j$ across all portfolios.

$$A_j = \sum_{i=1}^{N} \text{Decrypt}(a_{i,j}) \quad \forall \medspace j \in \{1, \dots, M\}$$

These totals $A_j$ are used to compute the single, batched portfolio finally executed onchain.

Note that we can decrypt the sum, not sum decrypted entries, using homomorphic encryption, to further minimize trust:

$$A_j = \text{Decrypt} \left( \sum_{i=1}^{N} a_{i,j}\right) \quad \forall \medspace j \in \{1, \dots, M\}$$

Let:

• $R_j$ be the plaintext price of asset $j$ as returned by an onchain oracle.

The Profit and Loss of each portfolio $i$ is computed as the inner product of its holdings with the public assets return:

$$P\&L_i = \sum_{j=1}^{M} \text{Decrypt}(a_{i,j}) \cdot R_j$$

As above, using FHE:

$$P\&L_i = \text{Decrypt}\left( \sum_{j=1}^{M} a_{i,j} \cdot R_j \right)$$

Proof

For an outside observer, the number of unknowns is $N \cdot M$ (portfolio states).

The number of equations is $M + N$ (one for each asset and one for each portfolio):

$$\begin{cases} \sum_{i=1}^{N} \text{Decrypt}(a_{i,j}) = A_j \quad \forall \medspace j \in \{1, \dots, M\} \\ \sum_{j=1}^{M} \text{Decrypt}(a_{i,j}) \cdot R_j = P\&L_i \quad \forall \medspace i \in \{1, \dots, N\} \end{cases}$$

Thus, the system is underdetermined and has infinitely many solutions if and only if:

$$N \cdot M > M + N$$

Which is easily satisfied for $N, M \geq 2$.

---

Footnotes

1. Zama Confidential Blockchain Protocol ↩