TasQ Network is a decentralized infrastructure for private AI compute, launched on the Robinhood Chain (EIP-155 chain ID 4663). Its official website (tasqnetwork.io) describes a system where users rent GPU capacity with three assurance modes—Attested (enclave-based confidentiality), Redundant (multi-seat execution with hidden audits), and Proven (zk-STARK verifiable outputs)—and where hardware providers lend GPUs while remaining blind to job content. The native $TASQ token is deployed at 0xdc453b1405cfe25e2afba8e7f4272b0a00437777 on Robinhood Chain, with explorers available via Robinhood Chain Blockscout and Robin Etherscan. Public benchmarks—including enclave overhead modeling, quorum capture simulations, and floating-point nondeterminism analysis—are published on the site’s /benchmark/ subpaths. The team lists four members with academic and technical backgrounds, though no public repository, live product demo, or third-party audit is linked. Key evaluation questions remain: What is the current treasury composition and liquidity? Is the claimed 1B-token circulating supply verified on-chain? How are GPU provider payouts settled—and what economic incentives prevent collusion or underreporting in mode R?
TasQ Network presents a technically detailed design for private AI compute using attested enclaves (mode A), redundant execution with hidden audits (mode R), and zk-STARK proofs (mode P). Its website documents assurance modes, job flow, cryptography benchmarks (e.g., 0.22 ms hybrid key agreement, 60 ms Merkle root over 65k chunks), and quorum economics — all backed by CSV-linked simulations and references to arXiv preprints. However, no on-chain or off-chain usage data is supplied: no transaction volume, active users, GPU utilization metrics, or verifiable marketplace activity appear in any source. The token contract on Robinhood Chain (eip155:4663) is confirmed, but the API snapshot reports only static supply figures (1B $TASQ circulating/total) with no transfer history, holder count, or fee revenue. The team page lists founders and advisors but offers no evidence of shipped software, testnet participation, or third-party integrations. While the threat model and admission rule are rigorously simulated, real-world validation remains absent. The primary thesis — that TasQ delivers production-grade private compute with measurable adoption — lacks verified traction evidence. This gap triggers the rubric’s hard cap: without demonstrated usage, confidence is low and the score cannot exceed 6.
Overall score: 6/10 Confidence: Low