Zama

Zama is a Paris-based, open-source cryptography company specializing in Fully Homomorphic Encryption (FHE) for blockchain and confidential computing.

Reviewed by 7wData
API Available

On this page

Publisher review

Zama is a Paris-based, open-source cryptography company specializing in Fully Homomorphic Encryption (FHE) for blockchain and confidential computing. Founded in 2020 by Dr. Rand Hindi and Dr.

Pascal Paillier (inventor of the Paillier cryptosystem), Zama hit unicorn status in June 2025 after raising $57 million in Series B funding from Pantera Capital and Blockchange Ventures, bringing total funding to over $150 million. The company provides a protocol enabling confidential smart contracts on any L1 or L2 blockchain without forking the underlying chain, plus open-source libraries (TFHE-rs, Concrete, Concrete ML, FHEVM) that let developers write encrypted applications in Solidity-like syntax. The architecture separates execution from encrypted computation using off-chain coprocessors and threshold decryption, preserving public verifiability while keeping data encrypted at rest and in transit.

Current throughput stands at 20+ transactions per second on CPU, with roadmap targets of 500–1,000 TPS by end of 2026 via GPU migration and 100,000+ TPS with dedicated ASICs by 2027–2028. Use cases span confidential DeFi (swaps, lending), private payments, sealed-bid auctions, identity verification, and enterprise healthcare applications. The team includes ~40% PhD holders and world-class cryptographers including Chief Academic Officer Nigel Smart.

Unlike competitor approaches (Secret Network's Trusted Execution Environments, Aztec's ZK-based privacy), Zama bets on FHE's computational guarantees at the cost of current latency—operations run roughly 100x slower than plaintext equivalents, though bootstrapping latency dropped from 53ms to under 1ms on NVIDIA H100 GPUs. The bet is that hardware acceleration (GPU, FPGA, ASIC roadmap) will close the performance gap, but execution risk remains material.

Get the AI & data signal, daily.

48k+ subscribers read this every morning. One email, both newsletters. Unsubscribe anytime.

How it works

  1. FHEVM (FHE Virtual Machine)

    EVM-compatible virtual machine letting developers write confidential smart contracts in Solidity using encrypted data types (euint), with operations executed off-chain on encrypted inputs and results stored on-chain without revealing plaintext.

  2. Open-source FHE libraries (TFHE-rs, Concrete, Concrete ML)

    Rust and Python libraries providing foundational homomorphic encryption primitives, machine learning inference on encrypted data, and compiler tooling for building applications without cryptographic expertise.

  3. Threshold decryption (MPC-based key management)

    13-node multi-party computation committee managing decryption keys with honest-majority assumption, enabling access control and preventing single points of compromise at the cost of collusion risk.

  4. Chain-agnostic coprocessor architecture

    Off-chain compute layer separating encrypted computation from blockchain consensus, allowing confidential apps to run on Ethereum, L2s, or other chains without modifying base layer protocols.

  5. Access control lists (ACL)

    Fine-grained policies defining who can encrypt, decrypt, or temporarily access encrypted values within transactions, enabling selective data visibility without breaking composability.

  6. ZAMA token and protocol fees

    Usage-based pricing ($0.005–$0.50 per zero-knowledge proof, $0.001–$0.10 per decryption operation) paid in ZAMA tokens, capturing value from on-chain encrypted computation.

  7. Hardware acceleration roadmap (GPU, FPGA, ASIC)

    Planned migration from CPU-only execution to GPU acceleration (189,000 bootstraps/sec on H100) then custom silicon, targeting 100,000+ TPS per chain to close the 100x latency gap with plaintext computation.

Strengths and trade-offs

Strengths

  • Pioneer advantage in practical FHE: dramatic performance improvements (bootstrapping latency <1ms on H100) and first-to-market for FHE as a blockchain layer, backed by IP depth (Pascal Paillier co-founder is cryptography legend).
  • Developer experience design: Solidity interface hides cryptographic complexity, familiar toolchains (Hardhat, Foundry), and free open-source libraries lower adoption friction vs. other encrypted-computing stacks.
  • Architectural positioning: acts as privacy layer for existing chains rather than competing as a blockchain, licensing technology to other projects (Fhenix, Inco) and capturing value regardless of ecosystem winner.

Trade-offs

  • Performance gap remains material despite improvements: FHE operations ~100x slower than plaintext, limiting real-time applications (trading, interactive UX) and dependent entirely on executing hardware roadmap (GPU, ASIC timelines unpredictable).
  • Key management trust assumptions: 13-node MPC committee requires honest-majority consensus; collusion among threshold nodes could enable silent attacks undetectable by other participants, centralizing decryption authority.
  • Early adoption and ecosystem scale: only ~5,000 developers active in 2026, competitive pressure from established alternatives (Secret Network's TEE-based privacy, Aztec's ZK-backed by a16z), and regulatory uncertainty around privacy-preserving finance features under MiCA and other frameworks.

Pricing context

Zama operates a dual-license model: open-source libraries (TFHE-rs, Concrete, Concrete ML, FHEVM) are free and MIT/Apache 2.0 licensed, lowering entry cost for developers. The Zama Protocol itself uses usage-based, token-denominated pricing: $0.005–$0.50 per zero-knowledge proof-of-knowledge verification and $0.001–$0.10 per decryption operation, paid in ZAMA tokens. Enterprise solutions are available but not publicly priced. There is no per-seat or per-contract licensing tier advertised; pricing scales with on-chain encrypted computation volume.

Getting started with Zama

  1. Install Zama open-source libraries

    Install TFHE-rs, Concrete, or Concrete ML via Cargo or pip. For example, run `cargo add tfhe` in your Rust project or `pip install concrete-ml` in Python. These libraries provide foundational FHE primitives for building encrypted applications.

  2. Connect to a blockchain network

    Set up a connection to an Ethereum-compatible chain (e.g., Ethereum mainnet, an L2, or a local Hardhat node). Configure your wallet and RPC endpoint in your development environment to deploy and interact with confidential smart contracts.

  3. Write a confidential smart contract

    Use Solidity with Zama's FHEVM extensions to declare encrypted variables with the `euint` type. Implement functions that operate on encrypted data, such as a private token transfer, using the provided FHE operations. Compile the contract with Hardhat or Foundry.

  4. Deploy and test the contract

    Deploy your compiled contract to the connected blockchain using a deployment script. Execute test transactions that call your contract's functions, verifying that encrypted inputs are processed off-chain and results are stored on-chain without revealing plaintext.

  5. Configure access control policies

    Define access control lists (ACLs) in your contract to specify which addresses can encrypt, decrypt, or view encrypted values. Set these policies using the FHEVM's ACL functions to enable selective data visibility while maintaining composability with other contracts.

Frequently Asked Questions

What is Zama and what does it do?

Zama is a Paris-based open-source cryptography company founded in 2020. It specializes in Fully Homomorphic Encryption (FHE) for blockchain and confidential computing, enabling encrypted data to be processed without decryption. It hit unicorn status in June 2025 after a $57 million Series B.

How does Zama's FHE technology work for blockchain privacy?

Zama uses Fully Homomorphic Encryption to keep data encrypted at rest and in transit. Its chain-agnostic coprocessor architecture separates execution from encrypted computation, using off-chain coprocessors and threshold decryption. This allows confidential smart contracts on any L1 or L2 without forking the underlying chain.

What are Zama's main products and libraries for developers?

Zama offers open-source libraries like TFHE-rs, Concrete, Concrete ML, and FHEVM. These let developers write encrypted applications in Solidity-like syntax. FHEVM is an EVM-compatible virtual machine for confidential smart contracts, while Concrete ML enables machine learning inference on encrypted data.

How fast is Zama's FHE and what is its performance roadmap?

Current throughput is over 20 transactions per second on CPU, with bootstrapping latency under 1ms on NVIDIA H100 GPUs. Zama targets 500–1,000 TPS by end of 2026 via GPU migration and 100,000+ TPS with dedicated ASICs by 2027–2028. FHE operations remain roughly 100x slower than plaintext.

How does Zama's pricing and token model work?

Zama uses a dual-license model: open-source libraries are free under MIT/Apache 2.0. The Zama Protocol charges usage-based fees in ZAMA tokens, ranging from $0.005 to $0.50 per zero-knowledge proof verification and $0.001 to $0.10 per decryption operation. Enterprise pricing is not publicly listed.

What are the main weaknesses and risks of using Zama?

Key weaknesses include a material performance gap—FHE operations are about 100x slower than plaintext—and reliance on a 13-node MPC committee for key management, which risks collusion. Early adoption is limited to around 5,000 developers, and competition from TEE-based and ZK-based privacy solutions remains strong.

Alternatives

How Zama compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This tool

Zama

Pricing
Zama operates a dual-license model: open-source libraries (TFHE-rs, Concrete, Concrete ML, FHEVM) are free and MIT/Apache 2.0 licensed, lowering entry cost for developers. The Zama Protocol itself uses usage-based, token-denominated pricing: $0.005–$0.50 per zero-knowledge proof-of-knowledge verification and $0.001–$0.10 per decryption operation, paid in ZAMA tokens. Enterprise solutions are available but not publicly priced. There is no per-seat or per-contract licensing tier advertised; pricing scales with on-chain encrypted computation volume.
Target
Zama is a Paris-based, open-source cryptography company specializing in Fully Homomorphic Encryption (FHE) for blockchain and confidential computing.
Strength
Pioneer advantage in practical FHE: dramatic performance improvements (bootstrapping latency <1ms on H100) and first-to-market for FHE as a blockchain layer, backed by IP depth (Pascal Paillier co-founder is cryptography legend).
Watch for
Performance gap remains material despite improvements: FHE operations ~100x slower than plaintext, limiting real-time applications (trading, interactive UX) and dependent entirely on executing hardware roadmap (GPU, ASIC timelines unpredictable).

Amazon Aurora DSQL

Pricing
Pay-per-use: DPU-based compute and per-GB storage. Free tier available.
Target
Developers needing serverless distributed SQL with high availability.
Deployment
AWS managed cloud
Strength
99.999% multi-region availability and deep AWS ecosystem integration.
Watch for
Limited PostgreSQL compatibility; no migration tools from existing databases.

Neon

Pricing
Free tier; paid plans from $19/month for 10 GB storage.
Target
Serverless PostgreSQL users seeking instant branching and autoscaling.
Deployment
Cloud (multi-cloud)
Strength
Instant database branching for dev/test workflows.
Watch for
Compute-storage separation can cause cold start latency on idle branches.

Couchbase Capella

Pricing
Starting at $0.40/AU/hour; custom enterprise plans.
Target
Teams needing multi-model (document + SQL) with low-latency caching.
Deployment
Cloud (AWS, GCP, Azure)
Strength
Built-in caching layer via memory-first architecture for sub-millisecond reads.
Watch for
SQL++ dialect differs from standard SQL, requiring query rewrites.

User reviews

No user reviews yet. Be the first to write one.

Sources

Reporting on this tool draws on these publicly available sources.

  1. tech.eu — Series B funding round ($57M), $1B unicorn valuation, Pantera Capital and Blockchange co-leads, June 2025 timing.
  2. www.zama.org — Zama Confidential Blockchain Protocol announcement, protocol architecture, use cases (DeFi, payments, sealed-bid auctions), and roadmap targets (1,000+ TPS).
  3. blockeden.xyz — Technical strengths (bootstrapping latency <1ms on H100, 189,000 bootstraps/sec), weaknesses (performance gap, collusion risk in MPC key management), competition analysis, and adoption metrics (~5,000 developers).
  4. docs.zama.org — Formal protocol architecture, coprocessor design, threshold decryption, off-chain computation, and composability with existing smart contracts.
  5. www.figment.io — Modular architecture (host chains, FHEVM, coprocessors, KMS, gateway), access control lists, on-chain auditing, and separation of concerns between blockchain and FHE execution.
  6. learn.backpack.exchange — Primary use cases (confidential smart contracts, privacy-preserving dApps, compliance-enabled access control), developer experience improvements, and current limitations addressed by Zama's design.