Private by Design: The Future of Cryptography with ZK, MPC, and FHE

Elo Mukoro
Jun 16
7 min read

Background: I worked at the Ethereum Foundation’s Privacy and Scaling Explorations (pse.dev) for nearly two years, where I supported protocols like zkEmail, tlsNotary, and Semaphore. During this time, I built up a deep understanding of emerging cryptographic technologies.

These concepts are not trivial, they weren’t easy for me at first either! But over the years, I’ve learned how to break them down in a way that’s approachable. This guide is here to simplify these technologies so that anyone, whether a builder, researcher, or simply curious mind, can understand how they work and why they matter.

As we move into a new era of decentralized systems, AI, and global digital coordination, technologies that prioritize privacy, authenticity, and sovereignty will be at the core. This guide is here to help you stay ahead of the curve.

What You’ll Learn From This Guide:

What ZK, MPC, and FHE are
How these technologies work
Real-world use cases and practical applications

What is ZK, MPC and FHE?

Zero Knowledge Proofs (ZK)

Zero Knowledge Proofs (ZK) allow you to prove you know a certain piece of information without actually revealing the information itself. Sounds like magic, right?

Let’s break it down with a classic example to make it easier to grasp:

This image shows a scenario where Bob wants Alice to prove she knows the passcode to get through a locked door.

Normally, to prove this, Alice would need to reveal the code itself. But in this setup, we introduce two paths (A and B) with a door in the middle. Bob doesn’t know which path Alice takes to reach the door, but he can randomly choose one path and ask Alice to exit from it.

If Alice really knows the passcode, she can open the door and appear at whichever path Bob selects without ever revealing the code itself.

This challenge-response process can be repeated many times. If Alice consistently gets it right every time, Bob can be statistically confident (with over 99.9% certainty) that she does, in fact, know the code, even though she never shared it directly.

Makes sense? Now, let’s get a bit technical

In a Zero Knowledge Proof, there are two main roles:

The prover — the one who knows the secret (like Alice in the cave)
The verifier — the one who wants to be convinced the prover knows the secret (like Bob)

Just like the cave example created a physical setting (with doors and paths) where Alice could prove her knowledge without revealing the password, in real-world ZK systems, the setting is mathematical. Specifically, it’s built using polynomials.

Specifically, in real-world ZK systems, the prover encodes the problem they want to prove as a polynomial equation. Instead of revealing the raw inputs, they create a mathematical proof showing that the polynomial holds true under certain conditions.

For example, imagine the polynomial 3x + 2 = 12. Only the prover knows the exact equation and their proposed solution, say x = 3.333. The verifier does not see the value of x, but they can check the proof and confirm that the equation holds, meaning the prover's claim is valid, without ever learning the actual input.

Multi-Party Computation (MPC)

Multi-party Computation (MPC) is the ability for multiple parties to collaborate on privately shared inputs. This works perfectly in situations when multiple parties want to collaborate with each other without revealing sensitive information or details. In this way you can enable collaboration amongst competitors to better forward industries as a whole. Let me give you a practical example of MPC.

A classical example involves the Millionaires problem. You have 4 millionaires that want to figure out who the richest is amongst themselves without sharing their actual investments or secret sauce as you would say.

In the diagram, each participant submits encrypted data to the MPC platform. The platform computes on the encrypted data and determines who is the richest, without revealing anyone’s actual inputs. The answer is then decrypted, only if a set threshold of participants agree to reveal the result. In the image above, 3 of the 4 needed to agree to decrypt the answer.

A real-world example of MPC is multi-signature architecture. In this setup, you can require a certain number of signers to approve a transaction. A good example of this in action is Safe, which allows users to set thresholds for executing actions without revealing individual inputs. In the example above, Kale is revealed as the richest.

MPC is powerful because it unlocks collaboration without compromising privacy. Whether it’s determining the richest person, securing crypto wallets, or enabling private analytics across organizations, MPC proves that you don’t have to choose between working together and keeping your data safe.

It’s a foundational building block for the future of private, trustless systems.

Fully Homomorphic Encryption (FHE)

Fully Homomorphic Encryption (FHE) lets you perform computations directly on encrypted data. That means you can run operations without ever needing to decrypt the inputs. Let me show you an example that applies to something real today.

Let’s say you're a developer and want to use ChatGPT for assistance. Maybe you need the AI to read your .env file, but because it contains private keys, you wouldn’t want OpenAI to actually see your secrets. This is where FHE comes in.

With FHE, imagine encrypting your input or prompt. The encrypted version (ciphertext) still allows the AI to compute on it. The AI doesn't see the original input, only the ciphertext, but it can still run the same logic and return the correct answer. All without ever accessing your actual .env file.

So instead of the AI seeing:

It sees something like:

And still gives you the same result without ever knowing what was inside.

FHE makes privacy preserving computation possible. You can share encrypted data, get useful results, and still keep your secrets safe.

At a more technical level, FHE enables encrypted data to be processed using special mathematical operations that preserve structure even while unreadable. The input is encrypted with a public key, computations are performed entirely on ciphertext, and only the data owner holding the private key can decrypt the final output.

This means you can offload computation to untrusted environments like cloud servers without ever revealing the original data. The processor never sees the plaintext and the data remains secure at every stage.

The image below illustrates this flow, from encryption to computation to decryption, while maintaining complete confidentiality throughout.

Use Cases for ZK, MPC & FHE

For 2 years I’ve pondered on a wide array of use cases and I will list some of the ones that I find promising for each category

Zero Knowledge Proofs (ZK)

Privacy preserving credentials using zkTLS

Users can prove claims about their online activity or identity such as income, employment, or account ownership without revealing the underlying data. This is powerful for credit scoring, onboarding, and access control.

Examples: www.cr3dentials.xyz, www.self.xyz

Decentralized on and off ramps using zkTLS

Users can prove that they’ve completed KYC or meet certain regulatory requirements without exposing their full identity. This allows seamless interaction between centralized finance and crypto without compromising user privacy. Examples: www.zkp2p.xyz

Anonymous voting and governance

ZK proofs can ensure that only eligible users vote while keeping individual choices private. This is especially important for DAOs and decentralized communities.

Example: https://maci.pse.dev/

Multi Party Computation (MPC)

Matchmaking and online dating platforms

Two people can check if they’re a match based on preferences without revealing personal details to the platform or each other unless there's a mutual match. This preserves privacy while enhancing trust.

Collaboration between competitors

Industries like insurance or finance can share insights or detect fraud across datasets without ever exposing sensitive customer data. MPC allows them to compute joint statistics or patterns while keeping their raw data confidential.

Threshold wallet recovery or access

Instead of storing private keys in a single location, key shards can be distributed across multiple parties. Only when a set threshold is met can the full key be reconstructed, boosting wallet security.

Example: https://safe.global/

Fully Homomorphic Encryption (FHE)

Federated learning for AI model training

Multiple data providers such as hospitals or banks can contribute encrypted data to train a shared AI model. Since the data remains encrypted during training, privacy is preserved and bias can be reduced through broader secure collaboration.

Private analytics in the cloud

Companies can run analytics or generate business intelligence from encrypted customer data stored in cloud environments they don’t fully trust. The data remains encrypted throughout the process and only the final result is decrypted locally.

Secure outsourcing of computation

Any company can offload expensive computations like risk modeling or genomic analysis to third parties without giving up data confidentiality. FHE makes it possible to extract value without sacrificing security.

Conclusion: The New Cryptographic Stack

ZK, MPC, and FHE aren’t just theoretical tools or buzzwords. They’re forming the foundation of a new cryptographic stack, one that prioritizes privacy, security, and collaboration without compromise. These technologies aren’t used in isolation either. In production, they often work together. ZK with MPC, for example, enables private computation with proof of correctness. MPC with FHE allows multiple parties to compute directly on encrypted data without ever revealing their inputs.

Whether it’s verifying income without exposing bank statements, securing wallets across multiple signers, or training AI models on encrypted health data, these technologies open up powerful new design spaces for the internet and society at large.

They also challenge us to think differently. In a world obsessed with transparency and data extraction, these tools remind us that privacy is not the opposite of utility. It’s part of it.

If you’ve made it this far, I hope this guide helped you better understand the power of these primitives and sparked some ideas of your own. Whether you're a developer, founder, policymaker, or curious mind, the future of cryptography is yours to help shape.

If you're building in this space or want to explore it further, feel free to reach out. Let's build what’s next together.