ARCHITECTURAL SPECIFICATION · TRL 8/9 VALIDATED

Eliminating Entropy in Distributed Consensus Through Mathematical Orchestration.

Neurallity SA formulates, models, and licenses deterministic orchestration architectures that treat decentralized computing clusters as a single coherent cryptographic engine. Backed by its dedicated internal engineering and scientific research division, Neurallity enforces disjoint search sub-spaces and native L1 state execution, mathematically eliminating stale work dissipation and latency penalties across heterogeneous networks.

HASHSTRIKE ARCHITECTURE BRIEF

EXECUTION RUNTIME · 00:32

Efficiency Derivation

|dη_B/dC| < |dη_T/dC|

Invariance of efficiency under complexity

Search Collision Factor

0.00% Verified

Strictly disjoint candidate domains (Si ∩ Sj = ∅)

Protocol Layer

Direct L1 Bypass

Eliminates Stratum intermediary fee/latency taxes

Legal Territorial Moat

158 PCT Nations

IT App. 102025000019186 (Grant Apr-Jul 2027)

// 01 · SYSTEM TOPOLOGY & EXECUTION PIPELINE

The Cluster is the Single Engine

Conventional consensus networks operate through uncoordinated probabilistic addition: individual nodes connect blindly to external aggregation pools, incurring stale-share penalties, duplicate work, and high latency transport overhead. The specification inverts this model: the coordinator ingests the raw blockchain state, constructs canonical candidate tuples, and governs disjoint candidate leases across heterogeneous nodes without intermediate fragmentation.

Stage 01

Job Ingestion

Direct L1 block template generation or Stratum pool job subscription.

Stage 02

Strategic AI1

Solo-mining lifetime estimation, difficulty evaluation, and domain scoping.

Stage 03

AI2 Segmentation

Dynamic disjoint lease generation across heterogeneous resources.

Stage 04

Pipeline PoW

Complete execution of consensus algorithms with midstate optimizations.

Stage 05

L1 Verification

Independent mathematical validation and microsecond block settlement.

MODULE 1.1

Canonical Tuple Construction

The orchestrator locks immutable block fields (version, prev_hash, nBits, Merkle roots) and extracts the protocol-permitted mutable dimensions. Work is assigned as verifiable full candidate spaces rather than arbitrary, unauthenticated fragments.

Search Space: D(J) Mapped
Tuple Consistency: Strict Consensus
Integrity Check: Internal Hash Matrix

MODULE 1.2

Disjoint Lease Ledger

Search intervals are leased deterministically. Each worker receives a strictly bounded subspace defined by [nonce_start, nonce_count] and algorithm parameters. Completed, active, and stale ranges are tracked in a real-time state ledger.

Sub-Domain: Si ∩ Sj = ∅
Lease Expiry: Dynamic Heartbeat
Collision Rate: λ_dup → 0

MODULE 1.3

Independent PoW Validation

When an engine returns a candidate satisfying the difficulty target, the coordinator rebuilds the entire candidate buffer and recomputes the cryptographic proof independently before external broadcast, preventing invalid block submissions.

Audit Engine: Clean-Room Logic
Target Evaluation: Meets Network / Pool
Settlement Latency: < 500 Microseconds

// 02 · RIGOROUS MATHEMATICAL FORMULATION

Decomposition of Useful Hashrate & Efficiency Retention

The central objective is not to artificially generate raw compute beyond physical limits, but to prevent the non-linear degradation of useful work (η) as systemic complexity (C) escalates.

H_useful = η · H_gross = η · Σ H_i

Where H_gross is the sum of raw physical hashes executed across N active cluster devices, and η ∈ [0, 1] represents the net utility factor. In conventional architectures, η rapidly degrades under scale due to latency, collisions, and state divergence.

EQUATION 2.1 EFFICIENCY DECOMPOSITION

Factorization of Useful Work

To isolate and eliminate every structural loss vector, total system efficiency is decomposed into five orthogonal, measurable coefficients:

η = η_share · η_uptime · η_network · η_alloc · η_switch

η_share: S_acc / (S_acc + S_rej + S_stale)
η_uptime: T_active / T_total
η_network: Resilience to jitter & packet drops
η_alloc: Hardware-to-algorithm mapping optimality
η_switch: Zero-loss state migration efficiency

EQUATION 2.2 COMPLEXITY INDEX

Multidimensional Complexity Vector

System stress is parameterized by the normalized state vector x = (N, A, F, G, K, L, V, D), mapped into a scalar complexity metric:

C = Σ w_k · x̃_k | Σ w_k = 1, w_k ≥ 0

N, A: Active node count & concurrent blockchains
F, G: Switching frequency & geographic spread
K, L: Hardware heterogeneity & latency distribution
V, D: Resource variability & allocation dynamics

| dη_B / dC | < | dη_T / dC | ⟺ α_B < α_T

The Fundamental Mathematical Advantage: In the exponential decay function η(C) = η_min + (η_0 - η_min)e^(-αC), the orchestrator ensures a lower decay coefficient (α_B < α_T). The cumulative advantage over operational complexity interval [0, C_max] is strictly positive:
V = ∫ [η_B(C) - η_T(C)] dC > 0.

// 03 · DETERMINISTIC TWO-TIER INTELLIGENCE

Separation of Concerns: Strategic AI1 vs Micro-Allocation AI2

General-purpose single-agent bottlenecks are avoided by decoupling long-horizon economic forecasting from microsecond hardware scheduling.

AI1 ENGINE STRATEGIC SELECTION (SOLO MINING)

Macro Economic & Work Ingestion

Operates exclusively when the orchestrator commands its own node and mempool. AI1 evaluates dynamic template lifespans, difficulty gradients, expected monetary returns, and thermodynamic power costs:

max_a [ E(R) - c_E·E - c_S·S_stale - c_R·S_rej - c_T·T_switch ]

AI1 defines the exact residual sub-space S ⊆ R(J,t) with the highest mathematical expectation of solution prior to state expiration, passing cleanly bounded jobs to the execution tier.

AI2 ENGINE MICROSECOND ALLOCATION (HYBRID/POOL)

Real-Time Deterministic Scheduling

Active across both solo e pool modes. AI2 partitions the permitted mutable search range into strictly mutually exclusive leases based on hardware capabilities and real-time physical telemetry:

Telemetry Inputs: Core temp, bus latency, error frequency
Partition Rule: S = ⊎ S_i with S_i ∩ S_j = ∅
State Reclaim: Instantaneous reallocation of expired leases
Stale Abort: Immediate invalidation upon new block arrival

*In pool mode, AI1 is intentionally bypassed: the pool imposes work via Stratum (mining.notify). AI2 strictly partitions the extranonce2/nonce space permitted by the pool.

// 04 · SCIENTIFIC RESEARCH & IN-HOUSE ENGINEERING

Dedicated Technology & Applied Research Division

Neurallity SA maintains an independent in-house multidisciplinary research and development laboratory. Operating from Zug, our team unites senior mathematicians, cryptography researchers, low-latency firmware architects, and protocol engineers to continuously advance consensus orchestration science.

LAB 01 PURE MATHEMATICS

Mathematical Modeling & Optimization

Formulates real-time optimization functions, bounds degradation factors, and continuously proves efficiency retention bounds. Dedicated to analytical complexity models, stochastic probability engines, and formal consensus proofs.

Specialization: Dynamic Resource Optimization
Deliverable: Analytical Consensus Models
Objective: |dη_B/dC| Minimization

LAB 02 SYSTEM ENGINEERING

Low-Level Cryptographic Architecture

Develops custom micro-kernels, in-memory consensus state engines, and zero-latency interconnect protocols. Our engineers eliminate intermediate transport layers and bridge L1 daemon state pipelines directly into compute memory.

Specialization: Direct L1 Ingestion & Zero-Copy Buses
Language Stack: Bare-metal C/C++, Rust, Assembly
Execution: Microsecond Validation Loops

LAB 03 IP ADVANCEMENT

Patent Development & Bitstream Synthesis

Translates mathematical breakthroughs into enforceable industrial patent claims and cloud-reconfigurable acceleration logic. Manages continuous patent disclosures, experimental testbeds, and hardware-agnostic synthesis engines.

Specialization: Patent Expansion & Bitstream Logic
IP Portfolio: PCT 158 International Scope
Testing Tier: Continuous TRL 8/9 Verification

// 05 · MULTI-ALGORITHM CRYPTOGRAPHIC OPTIMIZATION

Mathematical Consensus Adapters

True orchestration respects consensus rules: Proof of Work cannot be computed by concatenating arbitrary, partial hashes. Parallelism occurs across candidate states or within native algorithm pipelines.

Algorithm	Candidate Construction & Consensus Logic	Partition Strategy	Hardware Invariance & Optimization
SHA-256d	80-byte header, Merkle branch, double SHA-256 compression against target.	Partition of nonce, extraNonce2, nTime, and version rolling fields.	Precomputation of midstate for invariant block prefixes. Full round pipelining.
Scrypt	Header serialized with PBKDF2-HMAC-SHA256, ROMix sequential memory buffer, and final digest.	Disjoint candidate intervals. Each candidate maintains isolated ROMix RAM states.	Strict memory-bus alignment. Memory buffers are never merged across distinct nonce states.
X11	Chained sequential execution through 11 discrete cryptographic functions (Blake to Simd).	Header mutable range partitioning across multiple flying jobs.	Pipelined data-flow across execution units; preserves strict mathematical non-commutativity.
RandomX	Variable-length mining blob, dynamic seed hash, and full RandomX virtual machine execution.	Disjoint nonce ranges referenced to active template and seed state.	Zero-setup amortization: datasets and scratchpads maintained resident in high-speed L3/RAM.
Ethash / Etchash	Header hash, epoch seed, pseudorandom DAG dataset fetches, mix compression, and Keccak.	Partitioning of candidate nonces bound to exact epoch DAG pairs.	DAG preserved read-only in memory; memory access coalescing prevents bus stalls.

EXCLUSIVE COMMERCIAL LICENSING

Technology Licensing & Commercial Architecture

Neurallity SA operates exclusively as an intellectual property holding company and computational research laboratory. Neurallity SA does not sell hardware, components, or services directly to the market. All commercialization rights are granted on an exclusive worldwide basis to Aethelis LTD (London, UK), which independently designs, manufactures, and markets the operational infrastructure powered by our patented orchestration framework.

NEURALLITY SA IP HOLDING & RESEARCH

Technology & Patent Core

Maintains the global intellectual property estate, oversees algorithmic research, and executes continuous mathematical optimization across consensus topologies through its dedicated in-house scientific division.

Core Role: Pure IP Ownership & Licensing
Asset: Patent IT 102025000019186 (PCT 158)
R&D Engine: Dedicated In-House Engineering & Math Division
Territory: Global Exclusive License to Aethelis

AETHELIS LTD EXCLUSIVE OPERATOR

Commercial Infrastructure

Independently engineers, manufactures, and commercializes physical systems powered by Neurallity technology, deploying enterprise HPC solutions and custom low-power FPGA hardware.

Infrastructure: Enterprise HPC Servers & Industrial Lines
Hardware Fleet: Custom FPGA Devices (140W–160W)
Channel: Direct Institutional Sales & Global ATP Network
Commercial Inquiries: Managed exclusively by Aethelis

// 06 · INTELLECTUAL PROPERTY & LEGAL MOAT

Patented Technology & Swiss IP Holding

Neurallity SA (Zug, Switzerland) operates strictly as a research laboratory, patent holding company, and technology licensing entity.

PATENT ESTATE PCT EXTENSION

Patent Application IT 102025000019186

Priority application filed July 25, 2025. Official Technical Search Report co-examined covering claims 1–18. Definitive grant window scheduled between April and July 2027. Priority extended globally via the Patent Cooperation Treaty (PCT) across 158 member jurisdictions.

Core Claims: Distributed coordinator, multi-tier AI, PoW audit
Protection Scope: Hardware-agnostic consensus orchestration
Status: Patent Pending (Priority Registered, PCT Active)

CORPORATE DOMICILE GOVERNANCE

Neurallity SA · Zug, Switzerland

Located in the heart of Switzerland's Crypto Valley, Neurallity SA licenses protocol implementations, algorithmic cores, and orchestration frameworks to institutional computing networks, industrial data centers, and enterprise alliances.

Corporate ID: CHE-371.492.215
Legal Nature: Société Anonyme (Swiss Corporation)
Official Contact: info@neurallity.ch