OPEN_SOURCE ↗
REDDIT · REDDIT// 14d agoINFRASTRUCTURE
Qubic merges AI training, Dogecoin mining
Qubic says its Useful Proof of Work turns mining into Aigarth AI training, and on April 1, 2026 it plans to run Dogecoin mining in parallel on ASICs while CPUs and GPUs keep training. The project also touts a CertiK-verified 15.52 million TPS mainnet test, but the real question is whether this is meaningful AI infrastructure or a clever crypto experiment with an AI label.
// ANALYSIS
Qubic is genuinely unusual because it tries to make consensus work useful, not wasted. The catch is that the AI side is still more architectural promise than proven research platform.
- –Official docs frame UPoW as AI training on bare metal, so the AI angle is baked into the protocol rather than bolted on later.
- –The 15.52M TPS number is a throughput headline, not evidence of model quality, training efficiency, or research impact.
- –The April 1, 2026 Dogecoin rollout is technically neat because ASICs and CPU/GPU workloads are separate, so the network can run both in parallel.
- –Compared with Bittensor's subnet marketplace, Qubic is the more literal "mining funds training" experiment, which is novel but harder to validate.
- –If Qubic wants to break into serious AI discussions, it will need reproducible training results and public evals, not just faster chains and new token flows.
// TAGS
qubicopen-sourceself-hostedresearchgpu
DISCOVERED
14d ago
2026-03-28
PUBLISHED
15d ago
2026-03-28
RELEVANCE
8/ 10
AUTHOR
srodland01