Qubic Enables Parallel Dogecoin Mining and AI Training

Qubic shifts to a parallel architecture, running Dogecoin mining on ASICs while CPU/GPU AI training operates at full capacity, in a phased rollout to safeguard stability.

Rich by Coin

2026-03-27

On April 1st, 2026, Qubic flips the switch on Dogecoin mining, and the entire mining architecture of the network changes with it.
Qubic is transitioning to a purpose-built system, one where ASIC miners handle Dogecoin and CPUs and GPUs handle AI training.
The destination is a network running at 100% capacity across both workstreams simultaneously, and the path there is a carefully phased rollout.
If you’ve been following Qubic, you know the network has always been about making computation useful.

This transition takes that philosophy from promising to proven.
Under the previous model, Qubic miners split their time between two tasks.
Roughly 50% of compute time went toward mining Monero (XMR).
The other 50% went toward training Aigarth, Qubic’s own AI.

CPUs toggled back and forth, and while the system worked, neither task got the full attention of the hardware running it.
Dogecoin uses the Scrypt hashing algorithm, which runs on ASIC hardware: dedicated machines built for that specific type of work.
Qubic’s AI training runs on CPUs and GPUs.
Different hardware.