Agreement
Tetrahedral consensus
Compact 4-node tetrahedra reach Byzantine agreement locally, tolerating one fault per group. Results then aggregate upward into pyramids instead of one global coordination round.
Sierpinski Pyramid Consensus · testnet live
Sierpinski Chain arranges consensus as a recursive tetrahedron — four nodes agree locally, results aggregate into pyramids, and finality lands in milliseconds. Throughput scales as 4^L because the structure itself subdivides by four.
0+
Projected TPS at L=5
20–0 ms
Adaptive finality
O(log N)
Per-node communication
≤7
Steady-state peers
The architecture
SPC doesn’t run one network-wide consensus. It runs thousands of compact 4-node tetrahedra in parallel. Each group reaches Byzantine agreement locally, then its result aggregates into the pyramid above it — until a single deterministic finality emerges from the hierarchy.
BFT
One fault tolerated per tetrahedron
~⅓
Malicious participation tolerated
Agreement
Compact 4-node tetrahedra reach Byzantine agreement locally, tolerating one fault per group. Results then aggregate upward into pyramids instead of one global coordination round.
Throughput
Every level subdivides a tetrahedron into four — so capacity scales as 4^L. Tens of thousands of TPS at lower levels, 1,000,000+ around L=5.
Coordination
Communication stays O(log N) per node with a bounded peer set. Scaling the network doesn't mean scaling the chatter between every pair of validators.
Finality
Finality lands in 20–200 milliseconds, tuned to transaction risk. No probabilistic confirmations, no reorg anxiety.
What makes it different
Other chains scale with bigger machines, fewer validators, or more layering. Sierpinski scales by geometry — a fractal layout where work expands exponentially while coordination cost grows logarithmically.
Capacity doesn't come from hiding congestion behind larger blocks or delayed settlement windows.
Performance isn't bought by pushing consensus into fewer validators or heavier machines.
High-frequency activity stays on-chain rather than being forced into off-chain escape hatches.
The platform
Deterministic execution via a zwasm engine, with SRC-20 tokens — including the MYRT stablecoin and AUGR gold reserve — as first-class citizens.
Soul-Bound Tokens carry KYC and KYB credentials on-chain, issued through Gladius and synced across every node.
Chain IDs isolate networks at the P2P and hash level. Mainnet, testnet, and per-customer chains run side by side without replay risk.
Dedicated binaries for AI/ML operations, storage providers, and permissioned chain configuration — a full node surface, not just a ledger.
Identity on SPC
Gladius is SPC’s identity layer. It verifies people and businesses, then issues Soul-Bound Tokens — KYC and KYB credentials that live on-chain, follow the holder, and can’t be transferred. Liveness checks, document OCR, and provider approval, all settled to the chain.
Individuals and businesses, verified against documents.
Non-transferable SBTs that follow the holder on-chain.
Live-camera checks and document reading in the browser.
Authorized verifiers issue and revoke credentials.
Live now
Developer flow
Zig at the core, TypeScript at the edges. Spin up a local cluster, create a wallet, and query the chain — no account, no wait.
Safety
Byzantine fault tolerance with one fault tolerated per 4-node tetrahedron.
Consensus properties verified in the formal pipeline, not just asserted.
Failure probability drops exponentially as the recursive structure deepens.
The shift
If you need on-chain scale without probabilistic settlement or off-chain escape hatches, Sierpinski Chain is the architecture to evaluate.