Tenzro Testnet is live. Get testnet TNZO
← Back to Whitepapers

Tenzro Ledger: L1 Settlement Layer

Date: March 2026

License: MIT OR Apache-2.0

Abstract

Tenzro Ledger is the L1 settlement layer of the Tenzro Network — a purpose-built blockchain for the AI age providing four core pillars:

All transactions and settlements occur in TNZO, the native utility token used for gas fees, staking, and provider payments. The ledger supports three execution environments (EVM, SVM, DAML/Canton) through a unified multi-VM runtime, enabling autonomous agents to execute financial transactions on behalf of humans or autonomously.

This whitepaper describes the technical architecture of Tenzro Ledger, including consensus mechanism (HotStuff-2 BFT), execution layer (multi-VM with parallel Block-STM), fee market (EIP-1559), account abstraction (ERC-4337), settlement protocols, storage, and networking.

1. HotStuff-2 BFT Consensus

Tenzro Ledger uses HotStuff-2, a two-phase Byzantine Fault Tolerant (BFT) consensus protocol optimized for low-latency finality and linear communication complexity O(n). The protocol operates in sequential views with rotating leaders and provides optimistic responsiveness — it progresses at network speed during stable periods.

1.1 Protocol Parameters

ParameterValueDescription
Block time400msTarget interval between blocks
Max block size2 MBMaximum serialized block size
Max transactions/block10,000Maximum transaction count per block
Max gas/block30,000,000Maximum gas units per block
View timeout2,000msLeader failover timeout
Min validators4Minimum network size (3f+1 for f=1)
Epoch duration10,000 blocks~67 minutes at 400ms block time
Mempool size100,000Maximum pending transactions
Transaction TTL600sTime-to-live for pending transactions

1.2 Two-Phase Commit Protocol

HotStuff-2 reduces message complexity through a two-phase design:

  1. PREPARE phase: Leader proposes block, validators vote to prepare. Requires 2f+1 votes for quorum.
  2. COMMIT phase: Upon PREPARE quorum, validators vote to commit. Requires 2f+1 votes.
  3. DECIDE: Upon COMMIT quorum, block is finalized and appended to chain.

The quorum threshold is defined as 2f+1 where f = floor((n-1)/3) for n validators, tolerating up to f Byzantine failures. All votes include cryptographic signatures verified before collection.

1.3 Leader Selection

Three leader rotation strategies are supported:

1.4 TEE-Weighted Validation

Validators running within Trusted Execution Environments (Intel TDX, AMD SEV-SNP, AWS Nitro, NVIDIA GPU) receive 2x voting weight in leader selection and consensus quorum calculations. TEE attestations are verified before weight assignment. This incentivizes hardware-backed security and provides additional protection against validator key compromise.

1.5 Epoch Management

Epochs last 10,000 blocks (~67 minutes). At epoch boundaries:

Epoch transitions are atomic and include a view change to ensure clean validator set rotation.

1.6 Finality and Fork Choice

Finality is tracked via FinalityTracker with sequential finalization guarantees. Once a block receives a COMMIT quorum, it is immediately finalized — no further confirmations required. The fork choice rule follows the longest finalized chain by block height.

2. Multi-VM Execution Layer

Tenzro Ledger supports heterogeneous smart contract execution through three independent virtual machines orchestrated by the MultiVmRuntime. Transaction routing is determined by the VmType field in the transaction header.

2.1 Supported VMs

VMImplementationLanguagesUse Case
EVMEthereum VMSolidity, VyperEthereum-compatible DeFi, NFTs, DAOs
SVMSolana VM (BPF)Rust (BPF)High-performance programs, parallel execution
DAML/CantonCanton Ledger APIDAMLEnterprise contracts, privacy, composability

2.2 EVM Executor

The EVM executor provides full Ethereum-compatible execution with the following features:

2.3 SVM Executor

The SVM executor provides BPF program execution compatible with Solana:

2.4 DAML/Canton Executor

Each validator runs a Canton participant node connected via Ledger API v2:

2.5 Execution Constants

ConstantValueDescription
Max gas limit30,000,000Maximum gas per transaction
Default gas limit10,000,000Default if unspecified
Min gas price1 GweiMinimum base fee (1e9 wei)
Max contract size24,576 bytesEIP-170 limit
Chain ID1337Tenzro Ledger chain identifier
Max call depth1,024Maximum contract call stack depth

2.6 Precompile Registry

Tenzro-specific precompiles extend EVM functionality:

2.7 State Management

A unified state interface provides consistent access for all VMs:

3. EIP-1559 Dynamic Fee Market

Tenzro Ledger implements EIP-1559 for predictable transaction fees and network congestion management. The fee market dynamically adjusts the base fee based on block gas utilization relative to the target.

3.1 Fee Market Parameters

ParameterValueDescription
Target gas15,000,000Target gas per block (50% of max)
Min base fee0.1 GweiMinimum base fee floor
Max base fee1,000 GweiMaximum base fee ceiling
Base fee adjustment±12.5%Max change per block

3.2 Base Fee Calculation

The base fee for block n+1 is calculated based on block n's gas usage:

if gas_used > target_gas:
  new_base_fee = base_fee * 1.125 (increase by 12.5%)

if gas_used < target_gas:
  new_base_fee = base_fee * 0.875 (decrease by 12.5%)

if gas_used == target_gas:
  new_base_fee = base_fee (no change)

The base fee is clamped to [min_base_fee, max_base_fee] and burned, permanently removing TNZO from circulation.

3.3 Priority Fees

Users specify a max_priority_fee_per_gas to incentivize block producers. The priority fee is paid to the validator and is not burned. Total transaction fee:

total_fee = (base_fee + priority_fee) * gas_used
burned = base_fee * gas_used
to_validator = priority_fee * gas_used

3.4 Fee Suggestions

The FeeMarket provides priority fee suggestions based on desired confirmation urgency:

4. Block-STM Parallel Execution

Block-STM enables optimistic parallel transaction execution with automatic conflict detection and reexecution. This provides significant throughput improvements for independent transactions while maintaining correctness.

4.1 MVCC Multi-Version Data

The MultiVersionData structure maintains versioned state:

4.2 Execution Flow

  1. Optimistic execution: Execute all transactions in parallel assuming no conflicts
  2. Conflict detection: Compare read/write sets to identify dependencies
  3. Reexecution: Conflicting transactions reexecuted sequentially with updated state
  4. Fallback:If conflict rate > 50% or reexecutions > 16, fall back to sequential

4.3 Performance Metrics

The BlockStmExecutor tracks detailed execution metrics:

MetricDescription
total_executionsTotal transaction executions (including reexecutions)
conflicts_detectedNumber of read/write conflicts found
reexecutionsNumber of conflict-driven reexecutions
sequential_fallbacksTimes execution fell back to sequential mode
parallel_time_msTime spent in parallel execution
conflict_resolution_time_msTime spent detecting and resolving conflicts

4.4 Configuration

ParameterDefaultDescription
max_reexecutions16Maximum reexecution attempts before sequential fallback
conflict_threshold0.5Conflict rate threshold for sequential fallback

5. Account Abstraction (ERC-4337)

Tenzro Ledger implements ERC-4337 account abstraction, enabling smart contract wallets with programmable transaction validation, gas sponsorship, and modular security features.

5.1 Core Components

EntryPoint Contract: Singleton contract that processes UserOperation bundles. Validates operations, handles gas payment, and executes calls. Max bundle size: 100 operations.

SmartAccount: Contract wallet with customizable validation logic. Supports modular extensions for different security models.

AccountFactory: Deploys smart accounts using CREATE2 for deterministic addresses. Address computed from factory address, salt, and initialization code hash.

Paymaster: Third-party contract that sponsors gas fees for user operations. Enables gasless transactions and subscription models.

5.2 Smart Account Modules

ModuleDescription
SocialRecoveryGuardian-based account recovery with threshold signatures
SessionKeyTemporary signing keys with limited permissions and expiration
SpendingLimitDaily/weekly spending caps with automatic reset
BatchingMulti-call execution in single transaction

5.3 UserOperation Structure

sender: Address // Smart account address

nonce: u64 // Anti-replay protection

init_code: Bytes // Account deployment code (if not deployed)

call_data: Bytes // Execution payload

call_gas_limit: u64 // Gas for execution

verification_gas_limit: u64 // Gas for validation

pre_verification_gas: u64 // Fixed overhead gas

max_fee_per_gas: u64 // EIP-1559 max fee

max_priority_fee_per_gas: u64 // EIP-1559 priority fee

paymaster_and_data: Bytes // Paymaster address + context

signature: Bytes // Account-specific signature

5.4 Execution Flow

  1. Validation: EntryPoint calls account's validateUserOp() with signature and gas limits
  2. Gas payment: If paymaster specified, call paymaster's validatePaymasterUserOp(), else debit account
  3. Execution: Call account's execute() with call_data
  4. Gas refund: Refund unused gas to payer (account or paymaster)

6. Settlement Layer

The settlement layer provides flexible payment protocols for AI inference, TEE services, and provider payments. All settlements occur in TNZO with a 0.5% network commission on provider payments (flows to treasury).

6.1 Settlement Modes

Immediate Settlement: Direct one-shot payment from consumer to provider. No escrow. Executed in single transaction.

Escrow Settlement: Funds locked in escrow contract until release conditions met. Supports multiple release conditions:

Micropayment Channels: Off-chain per-token billing for high-frequency payments (e.g., streaming inference). On-chain channel open/close with cryptographic state updates.

6.2 Micropayment Channel Protocol

  1. Open: Consumer deposits TNZO into channel contract, specifies provider and expiration
  2. Off-chain updates: For each inference token, consumer signs new state: (channel_id, sequence_number, total_paid)
  3. Close: Provider submits latest signed state on-chain, receives total_paid, remainder refunded to consumer
  4. Challenge period: Consumer can dispute with newer state (higher sequence_number) within challenge window

6.3 Batch Settlement

The BatchProcessor enables atomic execution of multiple settlements in a single transaction. If any settlement fails, entire batch is rolled back. Useful for:

6.4 Fee Collection

The FeeCollector routes network fees:

Fee TypeRateDestination
Transaction gas (base fee)EIP-1559 dynamicBurned (removed from supply)
Transaction gas (priority fee)User-specifiedBlock producer (validator)
Provider payment commission0.5%Network treasury

7. Storage and State

Tenzro Ledger uses persistent key-value storage with a Merkle Patricia Trie for state commitment. This provides efficient range queries, atomic batch writes, and cryptographic state proofs.

7.1 Storage Organization

Storage is organized into dedicated column families for efficient access:

Data CategoryContents
BlocksBlock height → serialized Block
StateMerkle Patricia Trie nodes for state commitment
AccountsAddress → account data (balance, nonce, storage root, code hash)
TransactionsTransaction hash → serialized Transaction
MetadataChain metadata (latest_height, genesis_hash, chain_id)
SnapshotsCompressed state snapshots for fast sync
SettlementsSettlement ID → settlement data and status
Payment ChannelsMicropayment channel state (deposits, balances, sequence)
Dispute ChallengesChannel dispute challenges and resolution data

7.2 Merkle Patricia Trie

The state trie uses a modified Merkle Patricia Trie (same as Ethereum):

7.3 Snapshots

Periodic snapshots enable fast sync for new nodes:

7.4 Storage Configuration

ParameterValueDescription
Block cache size1 GBStorage block cache for hot data
Write buffer size256 MBMemtable size before flush
Snapshot retention100Number of snapshots to keep
Bloom filter bits10Bits per key for bloom filter (false positive rate ~1%)

7.5 Write Durability

Finalized blocks are written with fsync to ensure durability across power loss or crashes. Non-finalized blocks use async writes for performance. All batch operations are atomic — either all writes succeed or all are rolled back.

8. P2P Networking

Tenzro Ledger uses a modular peer-to-peer networking stack with multiple protocols for discovery, messaging, and content routing. The network layer supports multiple transports and security protocols.

8.1 Networking Stack

LayerProtocolPurpose
TransportTCP, QUICNetwork layer connectivity
SecurityNoise XXEncrypted, authenticated connections
MultiplexingStream multiplexingMultiple streams over single connection
DiscoveryDistributed Hash TablePeer discovery and content routing
Pub/SubGossipsubTopic-based message propagation
IdentificationPeer identificationExchange peer info and supported protocols

8.2 Gossipsub Topics

The network uses topic-based pub/sub for different message types:

TopicMessage Type
tenzro/blocks/1.0.0Block propagation (proposals and finalized blocks)
tenzro/transactions/1.0.0Transaction propagation to mempool
tenzro/consensus/1.0.0HotStuff-2 consensus messages (PREPARE, COMMIT)
tenzro/attestations/1.0.0TEE attestation announcements
tenzro/models/1.0.0Model registry updates and provider announcements
tenzro/inference/1.0.0Inference requests and results
tenzro/status/1.0.0Status updates and peer discovery
tenzro/agents/1.0.0Agent-to-agent communication

8.3 Peer Management

The PeerManager tracks connection state and metrics:

8.4 Message Deduplication

Message deduplication prevents gossip amplification attacks:

8.5 Bootstrap and Discovery

New nodes join the network through bootstrap peers specified via CLI:

tenzro-node --boot-nodes /ip4/35.188.123.45/tcp/9000/p2p/12D3KooW...

After initial connection, the distributed hash table provides decentralized peer discovery. Peers periodically refresh their routing tables and discover new peers through structured queries.

9. Identity (TDIP)

The Tenzro Decentralized Identity Protocol (TDIP) provides unified identity for both humans and autonomous machines. All identities are anchored on-chain with W3C DID compatibility and support for verifiable credentials.

9.1 DID Format

Identity TypeDID FormatExample
Humandid:tenzro:human:{uuid}did:tenzro:human:a1b2c3d4-...
Machine (controlled)did:tenzro:machine:{controller}:{uuid}did:tenzro:machine:a1b2...:e5f6...
Machine (autonomous)did:tenzro:machine:{uuid}did:tenzro:machine:e5f6g7h8-...

9.2 KYC Tiers

Human identities have KYC tiers determining permissions and limits:

TierLevelDescription
Unverified0No identity verification, limited functionality
Basic1Email + phone verification
Enhanced2Government ID verification
Full3In-person verification or biometric attestation

9.3 Delegation Scopes

When humans delegate control to machines, DelegationScope defines fine-grained permissions:

9.4 Verifiable Credentials

TDIP supports W3C Verifiable Credentials for attestations:

10. Security (TEE + ZK)

Tenzro Ledger uses a dual security model combining Trusted Execution Environments (TEE) for hardware-backed confidentiality and Zero-Knowledge (ZK) proofs for cryptographic verification.

10.1 Supported TEE Providers

ProviderTechnologyUse Case
Intel TDXTrust Domain ExtensionsValidators, key management, confidential compute
AMD SEV-SNPSecure Encrypted VirtualizationValidators, encrypted VM execution
AWS NitroNitro EnclavesCloud-based validators, custody
NVIDIA GPUHopper/Blackwell/Ada Lovelace CCGPU-accelerated inference with attestation

10.2 ZK Proof System

Groth16 SNARKs on BN254 curve (~128-bit security) with three pre-built circuits:

10.3 GPU-Accelerated Proving

The GpuProvingEngine provides batch proof generation with multi-level compression:

FeatureValue
Max batch size100
Max constraints1,000,000
Compression levelsNone, Light (2x), Medium (4x), Heavy (8x)
Merkle aggregationTree-based proof batching

10.4 Hybrid ZK-in-TEE

Combining ZK proofs with TEE attestations provides layered security:

11. Token Economics

TNZO is the native utility token of Tenzro Ledger with 18-decimal precision. It serves three primary functions: gas fees, staking, and settlement.

11.1 Token Utility

Use CaseDescription
Gas feesTransaction fees (base fee burned via EIP-1559, priority fee to validators)
AI inference paymentPay model providers for inference (0.5% network commission)
TEE service paymentPay for key management, custody, confidential compute (0.5% network commission)
Validator stakingStake to become validator, earn block rewards and fees
Provider stakingStake to register as model or TEE provider, subject to slashing
Governance votingStake-weighted voting on protocol upgrades and parameters

11.2 Liquid Staking (stTNZO)

Users can stake TNZO in the liquid staking pool and receive stTNZO representing their staked position:

11.3 Treasury and Governance

The NetworkTreasury collects protocol fees and distributes via governance:

11.4 Reward Distribution

Validators and providers earn rewards distributed at epoch boundaries:

12. Implementation Status

12.1 Core Components

12.2 Development Phases

  1. Core infrastructure (cryptography, consensus, networking)
  2. Execution layer (multi-VM integration, state persistence)
  3. Network and consensus (bootstrap, genesis, validator management)
  4. Identity and payments (TDIP, MPP, x402, Tempo integration)
  5. Agent and protocol integration (MCP, A2A protocols)
  6. Testnet deployment
  7. Economics and settlement (liquid staking, micropayment channels)
  8. Bridge and interoperability
  9. AI infrastructure
  10. Client applications
  11. Production hardening and security audits

13. Conclusion

Tenzro Ledger provides a purpose-built L1 settlement layer for the AI age, combining:

The design establishes a foundation for building decentralized infrastructure enabling autonomous AI agents to access intelligence and security services with verifiable settlement. The implementation follows a phased development approach, with core infrastructure, execution layer, and settlement protocols in active development.

Live Testnet: https://rpc.tenzro.network

Repository: https://github.com/tenzro/tenzro-network

License: MIT OR Apache-2.0

← Back to Whitepapers