Qubic

L1 with Useful Proof-of-Work (UPoW). 676 Computors (BFT validators, 451/676 quorum) selected weekly by UPoW solution ranking. Bare-metal hardware. Feeless transactions. CertiK-marketed 15.5M TPS. Single trusted "Arbitrator" role (override-able by supermajority).

One-line verdict: "Tokenomically clever, AI-substantively empty, brand-contaminated by the Monero campaign, founder-risk dominant. An anti-Venice case study."

Team — Founder Risk Dominant

Sergey Ivancheglo ("Come-from-Beyond" / CfB):

• Co-founder of IOTA (departed June 2019, acrimonious public dispute with David Sønstebø over tokens)
• Original NXT dev (2013, also departed)
• Qubic is project #3 with the same departure pattern brewing
• Material governance risk

AI Substance — Skeptical

• Aigarth: ternary {-1, 0, +1} representation, Helix logic gates, variable-delay synapses, "Intelligent Tissue" evolutionary training (not gradient descent)
• ANNA (public instantiation) released to X
• Famously failed 1+1 (claimed -114), 1+2 = 28
• Zero standard benchmarks (MMLU/GSM8K/etc.)
• One ResearchGate paper
• Architecture is idiosyncratic neuromorphic work with no external validation in 12+ years of CfB pursuing ternary computing

Token

• $0.066, ~$85M MCap, rank #298
• Cap 200T (down from 1000T via vote)
• Circulating ~137.5T
• Weekly epochs emit 1T
• First halving Aug 2025 (Epoch 175): 850B → 450B
• Burn-on-use plus external-mining buyback loop
• Not on Binance or Coinbase; primary venue Gate, thin liquidity

The Monero Attack (Material Reputation Datapoint)

• May-July 2025: hashrate share went <2% → 25%+
• Aug 11, 2025: Qubic claimed 51% control / reorg
• arxiv:2512.01437 (Lee & Kim) shows sustained share was actually 23-34%, never majority; pattern was selfish mining with strategic release; revenues were below honest mining — campaign was subsidized by QUBIC dilution
• Triggered Monero "Detective Mining" / "Publish or Perish" defenses
• Permanently damaged Qubic reputation in PoW circles

Traction

• 676 Computors fixed
• No meaningful external AI usage
• Real signal: Doge Connect launched April 1, 2026 — Scrypt ASICs mine DOGE, proceeds buy-and-burn QUBIC. Functional revenue loop, but entirely separable from AI work

Critiques (Sharpest)

1. Founder departure pattern (NXT → IOTA → ??)
2. ANNA can't add; no benchmarks after 6+ years
3. Monero attack was unprofitable, unethical, and overstated by Qubic
4. DOGE buy-burn loop is any-DAO-could-do-this; doesn't require Qubic tech
5. AI angle is increasingly marketing veneer atop yield-farming mechanics

Competitive Slot

• vs. Bittensor: Bittensor 32+ functional models, 118+ subnets, multi-B MCap. Qubic 1 broken model, $85M. Bittensor wins on AI.
• vs. Venice: Venice has product + customers. Qubic has neither.
• vs. Akash Network/Render Network: those have actual GPU marketplaces. Qubic's UPoW is captive to Aigarth.

Verdict

Track as a tokenomics case study (the buy-and-burn loop is genuinely interesting). Not investable on AI substance. Founder risk dominates everything else.