Podcast: Unlocking Credible and Confidential DeFi and AI w/ Peyman, Fairblock

This week on the FHE Onchain podcast, I sat down with Peyman from Fairblock, a project building dynamic confidential computing infrastructure for Web3. We covered a lot: threshold encryption, MPC, how FHE fits into the puzzle, and what it actually means to run a private system on public infrastructure.

The conversation wasn’t about buzzwords — it was about practical architecture decisions and what it takes to make privacy-native apps actually work in production.

What Is Fairblock Building?

At a high level, Fairblock is building a decentralized network for confidential computing — that means users can interact with onchain applications without revealing their inputs, behaviors, or intentions in plaintext.

Rather than relying on a one-size-fits-all approach, Fairblock uses a mix of cryptographic tools — mostly MPC, but also threshold FHE (CKKS), threshold IBE, and a bit of TEE/ZK when necessary. Each application uses the right tool depending on the context: Do you need low-latency? Do you need strong guarantees? Are you optimizing for frontend UX, backend decryption, or both?

They call it dynamic confidential compute — basically a flexible stack for building decentralized, privacy-respecting applications across different blockchains.

Why Confidential Computing Matters

Right now, crypto has a pretty obvious privacy problem. Wallet addresses are public. Trades are public. Salaries, DAO votes, auction bids — all visible onchain. You can’t fix the user experience or security of Web3 until you fix the data layer.

Confidential computing isn’t about hiding for the sake of hiding. It’s about making systems harder to manipulate. MEV, sandwich attacks, and opaque governance games aren’t edge cases — they’re baked into how we’ve designed transparency-first chains. Fairblock is trying to rebuild that stack so applications can preserve privacy by default, without giving up composability or decentralization.

Fairblock’s Architecture in Practice

The main engine behind Fairblock is a validator network that supports MPC operations — the core idea is that transactions can be encrypted using tools like threshold IBE or FHE, and the validators jointly decrypt or compute over that data when specific onchain conditions are met.

Think sealed-bid auctions that actually stay sealed until the right time. Think time-locked orders that no one can front-run. Think private onchain AI models that don’t leak either the model or the user’s data.

On the frontend, developers integrate simple encryption libraries — similar to how HTTPS works today. Users don’t have to install new wallets or deal with crypto math. The application just works, and the data stays private.

Fairblock also integrates with existing ecosystems. They’ve built support for Arbitrum Stylus (using Rust smart contracts), Optimism (via precompiles), and are working toward integrations with Ethereum mainnet and others. The idea is: bring privacy where users already are, rather than forcing migration to isolated chains.

What Makes It Different?

Unlike some FHE VMs or privacy chains that rely on centralized co-processors or trusted hardware to perform computations, Fairblock is designed to stay decentralized at its core. Validators run the actual MPC protocols, enabling encrypted execution without handing over trust to a single party.

Their approach isn’t to bundle everything into one massive chain, either. It’s modular. Their tech can live natively in appchains or L2s or be integrated via libraries and smart contracts — depending on where and how developers want to use it.

Use Cases That Go Beyond "Just Privacy"

Peyman made a good point — confidential computing isn't just a privacy tool. It's a prerequisite for building credible, manipulation-resistant systems. Here are some of the areas where it can make a meaningful difference:

1. Intent-Based Protocols
Whether you're submitting a trade, placing a bid, or signaling a vote, leaking your intent gives someone else the opportunity to exploit it. With confidential computing, intent stays hidden until it's settled — and that makes the system fairer for everyone involved.

2. Private Token Launches & Auctions
Fairblock’s threshold encryption lets you do sealed-bid auctions without a trusted auctioneer. Everything stays encrypted until a certain condition is met — time-based, trigger-based, or even external oracle-based.

3. Fixed Rate Lending & DeFi Primitives
One of the overlooked challenges in DeFi is that once data is public, it becomes a point of manipulation. Fairblock enables private lending and clearing markets that don’t leak sensitive info like collateral amounts or bid/ask positions.

4. Confidential AI Onchain
This isn’t about putting ChatGPT onchain. It’s about protecting data and models during training and inference — letting users interact with onchain AI without leaking sensitive inputs, and letting model developers protect IP.

5. Private, Compliant Transfers
This includes salary payments, B2B transfers, treasury management, and more. The idea is to preserve privacy while staying compliant — whether you're an individual or an institution.

‍

Closing Thoughts

There’s a growing recognition that public-by-default systems don’t scale — not socially, not economically, and definitely not when it comes to UX. Confidential computing is a key unlock, and Fairblock is one of the few projects tackling it in a modular, developer-friendly way.

As Peyman said, “privacy isn't a feature — it’s a foundation.” Without it, we're just building transparent ledgers with no trust guarantees. With it, we open the door to new kinds of applications that were previously impossible.

If you want to dive deeper into the architecture, use cases, or hear Peyman’s take on confidential compute across both Web3 and Web2 (yes, including healthcare and private data marketplaces), check out the full episode below.

🎙 Full episode ⬇️

https://www.youtube.com/watch?v=LRA_iIXmLxA