GPU ambition.
Evidence first.

Boreal is developing a payment-chain research ecosystem with NVIDIA B300-class datacenter GPUs as a primary mining-performance target.

What we are building

A specified protocol, independently verifiable node, NVIDIA miner, recoverable encrypted wallet, pool/job interface, explorer, SDK, documentation and reproducible releases. Attention and adoption should follow useful working software and reproducible results.

Algorithm and hardware

The current small CPU memory experiment is separate from the running chain. The next research examines cache and recomputation shortcuts, verifier cost, and a shared-epoch dataset design. A 20 GiB dataset is an evaluation point, not a finalized network rule or proof of ASIC resistance.

B300 and H200 are datacenter targets; RTX 5090 is a practical comparison target. Runtime hardware inventory and measured results must distinguish product variants and CUDA architectures. AI inference FLOPS are not a mining benchmark.

Security is a body of evidence

Post-quantum transaction authorization is one component. Parser limits, wallet recovery, atomic persistence, reorg and difficulty simulations, hostile-network tests, independent implementations, reproducible builds and outside review are separate requirements. The project cannot currently claim to be the most advanced or most secure cryptocurrency.

Cross-chain research

Plan BOREAL representations on Ethereum and Solana with shared backing accounting, plus a separately backed BTC asset moving into and out of Boreal. No live bridge exists. Read the interoperability design →

Community work

Facebook and Instagram drafts accompany verified milestones. Account connections are pending; nothing is presented as externally published. The research site remains private. Community measures will track real testers, contributors and reproducible submissions rather than invented popularity.

Primary references

NVIDIA DGX B300 guide · NVIDIA compute capabilities · NIST ML-DSA standard