Reward efficient, reproducible model outputs by scoring quality per compute, verifying dataset provenance, auditing compute claims, and maintaining rolling reputations for stable trusted AI
MeritNet is a Bittensor subnet that rewards efficient, reproducible intelligence rather than raw compute or unverifiable accuracy. Miners are scored on quality per unit of compute, with validator-audited compute claims, dataset provenance verification, and a rolling on-chain reputation system.
Current incentive schemes favor brute-force scaling and short-term benchmark gaming. MeritNet corrects this by penalizing excessive compute and rewarding sustained, cost-effective performance, improving decentralization and signal quality.
Miners submit outputs plus signed compute telemetry, dataset fingerprints, and environment hashes.
Validators re-run sampled inferences, verify compute claims, and attest to provenance.
Scoring combines quality, cost, latency, provenance, and miner reputation:
efficiency = (quality × provenance × latency) / (compute^β)
score = normalized_efficiency × (1 + γ × reputation)
Epoch rewards are distributed using an exponent α to avoid winner-take-all dynamics.
Randomized validator sampling
Stake-backed slashing for false attestations
Reputation decay to prevent permanent dominance
Task-class normalization for fair cross-task comparison
MeritNet strengthens Bittensor’s core mission by turning TAO emissions into a credible market signal for efficient, trustworthy AI, discouraging wasteful centralization while preserving permissionless competition.
<h1>Fundraising Status — Quality and Efficiency Subnet</h1><p><strong>Project:</strong> Quality and Efficiency Subnet (QE-Subnet)<br><strong>Status:</strong> Pre-Seed / Seed Fundraising<br><strong>Date:</strong> February 4, 2026</p><h2>Executive Summary</h2><p>The Quality and Efficiency Subnet (QE-Subnet) is a Bittensor subnet designed to reward <strong>efficient, reproducible, and verifiable model performance</strong> relative to compute cost. By combining cost-aware scoring, validator-audited compute claims, dataset provenance verification, and an on-chain reputation system, QE-Subnet produces a credible market signal for high-value neural compute.</p><p>We are seeking <strong>pre-seed and seed funding</strong> to complete core protocol development, bootstrap validator infrastructure, run pilot deployments, and progress toward mainnet launch.</p><h2>Funding Target</h2><ul><li><p><strong>Total Raise:</strong> $500K–$1.5M USD (or equivalent in crypto / TAO)</p></li></ul><h3>Use of Funds</h3><ul><li><p><strong>40% — Core Team Expansion:</strong> Protocol engineering, DevOps, validator operations</p></li><li><p><strong>30% — Validator Infrastructure:</strong> Hardware, deployment tooling, validator grants</p></li><li><p><strong>20% — Pilot Programs:</strong> Cloud, research, and enterprise integrations</p></li><li><p><strong>10% — Operations:</strong> Legal, compliance, marketing, and general administration</p></li></ul><h2>Investment Rationale</h2><ul><li><p><strong>Large and Growing Market:</strong> >$30B annual ML inference spend; decentralized compute exceeding $5B and expanding.</p></li><li><p><strong>Clear Differentiation:</strong> First subnet to combine efficiency-based rewards, dataset provenance, and rolling on-chain reputation.</p></li><li><p><strong>Strong Token Alignment:</strong> QE-Subnet increases the informational value of TAO emissions by rewarding sustainable intelligence rather than raw compute.</p></li><li><p><strong>Execution Path:</strong> Identified pilot partners including GPU infrastructure providers and research organizations.</p></li><li><p><strong>Technical Credibility:</strong> Team experience across Web3 infrastructure, decentralized systems, and ML operations.</p></li></ul><h2>Fundraising Phases</h2><ul><li><p><strong>Pre-Seed (Current):</strong> $100K–$300K<br><em>Objective:</em> MVP completion, testnet launch, initial validator set.</p></li><li><p><strong>Seed (Q2 2026):</strong> $500K–$1M<br><em>Objective:</em> Validator network scaling, pilot deployments, reputation system hardening.</p></li><li><p><strong>Series A (Q4 2026):</strong> $3M–$5M<br><em>Objective:</em> Mainnet launch, enterprise onboarding, ecosystem expansion.</p></li></ul><h2>Key Milestones</h2><ul><li><p><strong>Q1 2026:</strong> Testnet live with initial validator infrastructure and pilot partners.</p></li><li><p><strong>Q2 2026:</strong> Pilot inference tasks running; reputation system in beta.</p></li><li><p><strong>Q3 2026:</strong> Mainnet proposal submitted; governance activation.</p></li><li><p><strong>Q4 2026:</strong> Public mainnet launch with 50+ validators and initial enterprise adoption.</p></li></ul>