Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimistic TEE-Rollups: A Hybrid Architecture for Scalable and Verifiable Generative AI Inference on Blockchain

Aaron Chan, Alex Ding, Frank Chen, Alan Wu, Bruce Zhang, Arther Tian

TL;DR

This work tackles the Verifiability Trilemma in decentralized AI on blockchains by introducing Optimistic TEE-Rollups (OTR), a hybrid architecture that leverages NVIDIA H100 Confidential Computing TEEs for sub-second provisional finality, an optimistic fraud-proof layer, and stochastic Zero-Knowledge spot-checks. It formally defines Proof of Efficient Attribution (PoEA) to cryptographically bind execution traces to hardware attestations, thereby guaranteeing model authenticity and preventing model-downgrade attacks. The authors provide a game-theoretic security analysis, showing that rational adversaries are economically disincentivized from cheating, and validate OTR through simulations and hardware benchmarks, achieving near-native throughput with a sub-second latency and per-query cost around $0.07. The approach offers a practical path to scalable, verifiable generative AI on decentralized networks, enabling interactive on-chain AI while maintaining cryptographic integrity and resilience to hardware-side-channel risks.

Abstract

The rapid integration of Large Language Models (LLMs) into decentralized physical infrastructure networks (DePIN) is currently bottlenecked by the Verifiability Trilemma, which posits that a decentralized inference system cannot simultaneously achieve high computational integrity, low latency, and low cost. Existing cryptographic solutions, such as Zero-Knowledge Machine Learning (ZKML), suffer from superlinear proving overheads (O(k NlogN)) that render them infeasible for billionparameter models. Conversely, optimistic approaches (opML) impose prohibitive dispute windows, preventing real-time interactivity, while recent "Proof of Quality" (PoQ) paradigms sacrifice cryptographic integrity for subjective semantic evaluation, leaving networks vulnerable to model downgrade attacks and reward hacking. In this paper, we introduce Optimistic TEE-Rollups (OTR), a hybrid verification protocol that harmonizes these constraints. OTR leverages NVIDIA H100 Confidential Computing Trusted Execution Environments (TEEs) to provide sub-second Provisional Finality, underpinned by an optimistic fraud-proof mechanism and stochastic Zero-Knowledge spot-checks to mitigate hardware side-channel risks. We formally define Proof of Efficient Attribution (PoEA), a consensus mechanism that cryptographically binds execution traces to hardware attestations, thereby guaranteeing model authenticity. Extensive simulations demonstrate that OTR achieves 99% of the throughput of centralized baselines with a marginal cost overhead of $0.07 per query, maintaining Byzantine fault tolerance against rational adversaries even in the presence of transient hardware vulnerabilities.

Optimistic TEE-Rollups: A Hybrid Architecture for Scalable and Verifiable Generative AI Inference on Blockchain

TL;DR

This work tackles the Verifiability Trilemma in decentralized AI on blockchains by introducing Optimistic TEE-Rollups (OTR), a hybrid architecture that leverages NVIDIA H100 Confidential Computing TEEs for sub-second provisional finality, an optimistic fraud-proof layer, and stochastic Zero-Knowledge spot-checks. It formally defines Proof of Efficient Attribution (PoEA) to cryptographically bind execution traces to hardware attestations, thereby guaranteeing model authenticity and preventing model-downgrade attacks. The authors provide a game-theoretic security analysis, showing that rational adversaries are economically disincentivized from cheating, and validate OTR through simulations and hardware benchmarks, achieving near-native throughput with a sub-second latency and per-query cost around $0.07. The approach offers a practical path to scalable, verifiable generative AI on decentralized networks, enabling interactive on-chain AI while maintaining cryptographic integrity and resilience to hardware-side-channel risks.

Abstract

The rapid integration of Large Language Models (LLMs) into decentralized physical infrastructure networks (DePIN) is currently bottlenecked by the Verifiability Trilemma, which posits that a decentralized inference system cannot simultaneously achieve high computational integrity, low latency, and low cost. Existing cryptographic solutions, such as Zero-Knowledge Machine Learning (ZKML), suffer from superlinear proving overheads (O(k NlogN)) that render them infeasible for billionparameter models. Conversely, optimistic approaches (opML) impose prohibitive dispute windows, preventing real-time interactivity, while recent "Proof of Quality" (PoQ) paradigms sacrifice cryptographic integrity for subjective semantic evaluation, leaving networks vulnerable to model downgrade attacks and reward hacking. In this paper, we introduce Optimistic TEE-Rollups (OTR), a hybrid verification protocol that harmonizes these constraints. OTR leverages NVIDIA H100 Confidential Computing Trusted Execution Environments (TEEs) to provide sub-second Provisional Finality, underpinned by an optimistic fraud-proof mechanism and stochastic Zero-Knowledge spot-checks to mitigate hardware side-channel risks. We formally define Proof of Efficient Attribution (PoEA), a consensus mechanism that cryptographically binds execution traces to hardware attestations, thereby guaranteeing model authenticity. Extensive simulations demonstrate that OTR achieves 99% of the throughput of centralized baselines with a marginal cost overhead of $0.07 per query, maintaining Byzantine fault tolerance against rational adversaries even in the presence of transient hardware vulnerabilities.
Paper Structure (42 sections, 13 equations, 3 figures, 5 tables, 1 algorithm)

This paper contains 42 sections, 13 equations, 3 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. The Limitations of Existing Paradigms
  3. The OTR Proposition
  4. Contributions
  5. Literature Review
  6. Validity Proofs: Zero-Knowledge Machine Learning (ZKML)
  7. Fraud Proofs: Optimistic Machine Learning (opML)
  8. Outcome-Based Verification: Proof of Quality (PoQ)
  9. Trusted Execution Environments (TEEs)
  10. Comparative Analysis and The OTR Proposition
  11. Methodology: The Optimistic TEE-Rollup (OTR) Protocol
  12. Formal Definitions and System Model
  13. Notations
  14. Threat Model
  15. Protocol Architecture
  16. ...and 27 more sections

Figures (3)

  • Figure 1: State transition diagram of the OTR Protocol showing the probabilistic verification path.
  • Figure 2: The Verifiability Frontier: Comparison of protocols based on Computational Integrity vs. Latency. OTR occupies the high-integrity/low-latency quadrant via hardware acceleration.
  • Figure 3: Latency vs. Model Size. ZKML scales super-linearly, becoming infeasible beyond 7B parameters. OPML has a static high latency (dispute window). OTR remains near real-time, tracking closely with PoQ but providing cryptographic integrity.