QSDM+ Open Dashboard →
MAINNET PREVIEW · NIST FIPS 204 · NVIDIA-LOCKED

Post-quantum distributed trust,
NVIDIA GPU-attested by design.

QSDM is the Quantum-Secure Dynamic Mesh Ledger — a non-AI, hardware-agnostic electronic cash and smart-contract fabric with a native coin, Cell (CELL). It is built on ML-DSA-87 post-quantum signatures, a dynamic 2D/3D mesh topology, and Proof-of-Entanglement BFT consensus that runs on CPU-only validators. An additive GPU mining layer is planned for Cell emission and runs independently of consensus; validators can optionally publish NVIDIA NGC proof bundles as a transparency signal, not as a consensus rule.

Open Dashboard API Health Learn more ↓
Signature scheme
ML-DSA-87
NIST FIPS 204 · 256-bit PQ security
Verify speed
0.19 ms
1.76× faster than ECDSA (0.33 ms)
Sig compression
~50%
zstd on signatures; 60–70% on tx storage
Batch signing
10–100×
vs single-tx baseline
Consensus
Proof-of-Entanglement
CPU validators · no proof-of-work · mesh-native
Storage engines
SQLite → ScyllaDB
Zstandard-compressed, linear migration path
PLATFORM

One ledger, two node roles.

QSDM is the unified platform (historically shipped under the transitional name QSDM+); Cell (CELL) is its native coin. The node software ships in two roles that share the same cryptography, wire format, and mesh topology: validators run Proof-of-Entanglement BFT consensus on CPU-only VPS-class hardware and earn transaction fees, while miners run an additive GPU proof-of-work layer that emits Cell from home rigs.

Validator · CPU-only · VPS

QSDM Validator Node

The consensus tier. A lattice-based, mesh-structured electronic cash system engineered from the first block to be quantum-resistant — no retrofits, no hybrids. Validators never require a GPU; Proof-of-Entanglement BFT runs on CPU. Optional NGC attestation acts as a transparency signal, not as a consensus rule.

  • ML-DSA-87 signatures (NIST FIPS 204) via liboqs
  • libp2p networking with GossipSub + PEX
  • SQLite or ScyllaDB storage with Zstandard compression
  • Proof-of-Entanglement consensus (PoE) with multi-parent cell validation
  • 2D/3D mesh topology, submesh profiles, and WASM contract SDK
  • Dashboard, WebSocket topology, Prometheus scrape, audit reports, governance snapshots
  • Ubuntu 24.04+ and Windows 10+ supported; macOS in development
Launch Dashboard GitHub Compare vs. Blockchain/DAG
Miner · GPU · Home rig · Planned

QSDM Miner (Cell emission) — planned

The emission tier. An additive proof-of-work layer for Cell (CELL) minting that runs independently of consensus. Home-scale GPU operators submit proofs to any QSDM validator and receive Cell rewards per the published emission schedule. The reference CPU miner ships with the node; the CUDA miner is planned to ship after external audit.

  • Separate qsdmminer binary — does not participate in consensus
  • Candidate PoW: mesh3D-tied useful proof; CPU-safe O(1) verify
  • Native coin Cell (CELL), 8 decimals, smallest unit dust
  • Halving emission schedule; 0% pre-mine, treasury cap 10%
  • Reference CPU miner in-tree; CUDA miner gated on external audit
  • NGC attestation is optional and consensus-independent
API: /api/v1/health
NVIDIA GPU EDGE

Silicon as a sybil-resistance primitive.

Most ledgers treat validators as interchangeable processes — easy to spin up, easy to fake. QSDM validators can optionally attach a fresh, HMAC-signed NGC proof bundle from a specific NVIDIA GPU as a transparency signal. NGC attestation is not a consensus rule — Proof-of-Entanglement runs on CPU and admits validators without a GPU — but operators that publish bundles give the network a cryptographic statement about the silicon behind them. The bundle is produced by a lightweight sidecar (apps/qsdmplus-nvidia-ngc, folder name retained during rebrand) that hashes CUDA device properties, runs a deterministic proof, and signs the result with a per-node secret.

The result: identity that is physically grounded. You cannot clone a validator without cloning the silicon behind it.

How it works ↓ Read the spec
01 · Attestation
NGC Proof Bundles
JSON envelopes with architecture, CUDA proof hash, GPU fingerprint, optional HMAC. Replayed-proof detection via nonce + freshness window.
02 · Binding
Node-ID Lock
nvidia_lock_expected_node_id binds a validator identity to exactly one signed GPU. Strict-secrets mode forbids demo/weak HMACs in production.
03 · Enforcement
P2P Gate
With nvidia_lock_gate_p2p=true, libp2p drops inbound transactions from peers without a recent valid proof — before consensus, before the mempool.
04 · Acceleration
CUDA Validation
Phase 3 uses the same GPUs for parallel hash verification and 3D mesh traversal, turning attestation hardware into productive compute.
CONSENSUS BOUNDARY

PoE is consensus. GPUs are not.

Proof-of-Entanglement and BFT finality do every piece of consensus work on QSDM — block proposal, transaction ordering, and finality — on CPU-only validator nodes. GPUs on the mesh serve two narrow, additive roles: hardware attestation (today) and coin emission (planned). Neither role touches consensus.

CONSENSUS · ALWAYS-ON
Proof-of-Entanglement + BFT
Runs on validators (CPU). Multi-parent cell validation, stake-weighted BFT voting, finality-gadget depth checks. No GPU, no proof-of-work, no waste cycles. The chain lives here.
ATTESTATION · OPT-IN
NVIDIA NGC proofs
Validators can opt into a signed GPU-attestation badge. Purely a transparency signal on the API boundary — it is not a consensus rule; blocks are never rejected for a missing proof.
MINING · ADDITIVE (PLANNED)
GPU coin emission
A separable reward track under design: home GPU miners submit hash-based proofs to earn newly-issued supply on a published schedule. Mining never proposes blocks, never orders transactions, and can halt with zero effect on chain liveness.
HARD GUARANTEE
Validators are CPU-only
Validator builds ship without CUDA code, K8s manifests pin CPU-only nodes, and a startup guard fails fast if mining is mistakenly enabled. Running a validator requires zero GPU hardware — now and after the mining layer ships.

PoE + BFT consensus is live today. The mining layer is a planned additive subsystem; introducing it will not change validator hardware requirements, consensus rules, or finality guarantees.

PARTICIPATE · MINE AT HOME · PLANNED

Mine Cell at home — planned.

The GPU mining layer is the additive emission track for Cell (CELL). It is deliberately separate from consensus: miners never propose blocks, never order transactions, and can halt with zero effect on chain liveness. Mining Cell is how new coins enter circulation — rewards depend on network difficulty and the current block reward, both of which change over time. The reference CPU miner ships with the node today; the CUDA miner is planned to ship after an external audit. A home-scale mining pool directory will arrive alongside the incentivized testnet.

REFERENCE · IN-TREE
CPU reference miner
cmd/qsdmminer builds a real proof under MINING_PROTOCOL.md and runs a --self-test. Good for protocol compliance and one-off validation. Hash rate is low by design — this is not the production miner.
PRODUCTION · PLANNED
CUDA miner
Optimized for NVIDIA (CUDA 11+, ≥ 8 GB VRAM, RTX 20/30/40/50-class). The protocol does not prohibit other vendors; we simply ship NVIDIA-first because that is where the hardware ecosystem is. Gated on external audit before release.
HOW YOU EARN
Block rewards — halving schedule
New Cell is emitted on every block according to the published emission schedule — 90 M lifetime mining cap, 4-year halvings, 0 % pre-mine, 10 % treasury cap. There are no daily-earnings promises here: difficulty and the block reward are both moving targets.

Why NVIDIA for mining? (and what you miss without one)

QSDM is not NVIDIA-only. The reference CPU miner is always supported and the protocol will accept a valid proof from any device. In practice, though, NVIDIA is the first-class path today — the mesh3D CUDA kernels ship in-tree, the production miner image is built against CUDA in CI, and the free NVIDIA NGC tier is the only attestation path currently published. This table is what that means for an operator choosing hardware.

Tier Hardware Relative hashrate Practical use NGC attestation badge?
Reference CPU Any x86-64, no GPU 1× baseline (single-digit H/s) Protocol conformance, self-test, occasional testnet block under low difficulty No GPU to attest
Entry NVIDIA
RTX 3050 / 3060		 CUDA 11+, 8 GB+ VRAM, NVIDIA Container Toolkit Thousands× CPU baseline (CUDA kernels unlock) Home miner — occasional mainnet blocks once mainnet ships Yes — free NGC tier
Mid-range NVIDIA
RTX 3080 / 4070		 CUDA 11+, ≥ 10 GB VRAM Scales with SM count and memory bandwidth Serious home miner, longer profitable runway as difficulty rises Yes — free NGC tier
High-end NVIDIA
RTX 4090 / H100-class		 CUDA 12+, ≥ 16 GB VRAM Rack-grade — matches or exceeds farm economics Aggressive difficulty ramps, best power-per-hash, longest runway Yes — free NGC tier
Non-NVIDIA GPU
AMD / Intel		 OpenCL / ROCm / Level Zero — no first-party kernels shipped today Requires community port of pkg/mesh3d Research / goodwill contributions only NGC is NVIDIA-only tooling

Qualitative tiers, not benchmark commits. Actual hashrate depends on driver version, thermal/power budget, DAG size, and the current mesh3D epoch.

FREE NGC TIER
No paid NVIDIA plan needed
Sign up at ngc.nvidia.com/setup — the free account issues an NGC CLI API key. You do not need NVIDIA AI Enterprise or any paid subscription to run attestation. The sidecar image (nvcr.io/nvidia/pytorch) and the NVIDIA Container Toolkit are free too.
TRANSPARENCY BADGE
Appear on the public trust page
Attested nodes show up on qsdm.tech/trust with a coarse region + GPU-arch signal — a cryptographic statement tying your node to real silicon. Non-attested nodes stay first-class on the network; the badge is a signal, not a gate.
SHIPPED KERNELS
CUDA is the supported fast path
pkg/mesh3d/cuda.go links libmesh3d_kernels.so, compiled with nvcc in the miner Dockerfile and built by CI on every main commit. Other vendors are welcome — they will require a community port.

Jurisdictional restrictions, tax treatment, electricity costs, and heat/noise output vary widely by location and hardware. Evaluate your own situation before running a miner. Nothing on this page is financial or legal advice.

End-to-end operator guide MINER_QUICKSTART.md Mining protocol spec Tokenomics
PARTICIPATE · RUN A VALIDATOR

Run a validator — VPS-class, CPU-only.

Validators are the consensus tier. They run Proof-of-Entanglement + BFT on CPU-only VPS-class hardware and earn transaction fees. No GPU is required — now or after the mining layer ships. Startup guards in the validator build refuse to boot if mining is mistakenly enabled, so a validator operator never has to reason about GPU drivers, CUDA, or thermal budgets. NGC attestation is available if you want a transparency badge, but it is not required to participate in consensus and does not affect block validity.

HARDWARE
Modest CPU VPS
4–8 vCPU, 8–16 GB RAM, 200 GB NVMe, 1 Gbps symmetric. No GPU. Ubuntu 24.04+ or Windows 10+ supported today; macOS in development. Dockerfile and Kubernetes manifests ship CPU-only by default.
NETWORK ROLE
PoE + BFT consensus
Multi-parent cell validation, stake-weighted BFT voting, finality-gadget depth checks. libp2p + GossipSub networking. Validators earn transaction fees — mining rewards go to the separable miner track.
OPTIONAL · TRANSPARENCY
NGC attestation opt-in
Opt in to publish a signed NVIDIA NGC proof bundle and earn the "Attested" badge on the public trust page. This is an API-boundary policy — not a consensus rule — and validators that opt out are fully first-class on the network.
Read VALIDATOR_QUICKSTART.md Operator guide (end-to-end) Node roles Attestation transparency
WHY IT MATTERS

Benefits of GPU-anchored, post-quantum mesh.

Put together, these properties let QSDM offer guarantees that traditional blockchains and DAGs cannot — without resorting to energy-hungry proof-of-work in the consensus path or stake concentrations that reintroduce custodial risk.

POST-QUANTUM

Harvest-now-decrypt-later-safe

ML-DSA-87 signatures protect ledger history from quantum adversaries that will break ECDSA and RSA. Today's signatures are still valid in 2040.

HARDWARE-BOUND

Sybil-resistant by silicon

GPU fingerprints + signed NGC proofs mean each validator is tied to a real device. Cloning a node means cloning a GPU — and we detect that.

THROUGHPUT

Mesh, not a chain

Cells can have multiple parents. Submeshes route by fee and geotag. There is no global block interval to throttle you.

ENERGY

No proof-of-work in consensus

Proof-of-Entanglement uses cryptographic linkage, not brute-force hashing. The planned GPU mining layer is an additive emission track — it never proposes or orders blocks.

NON-AI

Deterministic governance

Rule-based quarantines, snapshot voting, manual submesh templates. Auditable, reproducible, AI-free.

OPEN

libp2p + WASM + standard APIs

GossipSub transport, WASM smart contracts, REST + Prometheus endpoints. Nothing proprietary at the seams.

OPERABLE

Single-binary, dual-platform

One Go binary. Ubuntu 24.04+ and Windows 10+ supported today; macOS in development. Systemd + Caddy deploy in under an hour.

AUDITABLE

Metrics and reports by default

Prometheus /api/metrics/prometheus, audit markdown/JSON reports with severity gates, topology WebSocket, and signed release artifacts.

TECHNOLOGY COMPARISON

QSDM vs. traditional blockchains and DAGs.

A like-for-like comparison against representative systems. Values reflect the default, in-the-box configuration of each technology — not maximal lab setups.

Dimension QSDM Bitcoin · PoW Ethereum · PoS IOTA / Hedera · DAG
Data structure Dynamic 2D/3D mesh, multi-parent cells Linear block chain Linear block chain Directed acyclic graph
Consensus Proof-of-Entanglement (PoE) + BFT · CPU-only validators Proof-of-Work Proof-of-Stake (Gasper) Tangle / Hashgraph
Mining / emission Additive GPU track (planned) · never decides block order or finality Mining is consensus No mining; staking = consensus No mining; DAG voting = consensus
Quantum-safe signatures ✓ ML-DSA-87 (FIPS 204, 256-bit) ✗ ECDSA (secp256k1) ✗ BLS12-381 / ECDSA ◐ varies; mostly classical
GPU-attested validators ✓ NVIDIA NGC proofs + P2P gate ✗ ✗ ✗
Energy profile Low; CUDA for validation, not PoW Very high (hash mining) Low (staking) Low
Throughput scaling Submesh parallelism + fee routing ~7 tps mainnet ~15–30 tps (L1) High (DAG parallelism)
Finality PoE + finality gadget, reorg-bounded Probabilistic (6 blocks) Epoch-based (~12–15 min) Varies; often probabilistic
AI/ML dependencies None — rule-based governance None None (core); apps vary Varies
Smart contracts WASM SDK (wallet, validator) Script (non-Turing) EVM EVM (Hedera), limited (IOTA)
Storage SQLite + Zstd → ScyllaDB path Flat files / LevelDB Merkle-Patricia trie Custom / DAG pruning
Deployment footprint 1 Go binary · Linux/Windows Full node required Execution + consensus client Protocol-specific tooling
Operator tooling Dashboard, Prometheus, audit reports, mTLS RPC + 3rd party RPC + 3rd party RPC + 3rd party

Figures shown are representative of default mainnet configurations. QSDM quantum-safe primitives are based on NIST FIPS 204 (ML-DSA) standardized in 2024.

ROADMAP

Three phases. One mesh.

QSDM ships in phases, with each phase deliverable and production-usable on its own. Current deployments run Phase 1 infrastructure with Phase 2 submesh routing enabled.

2D Mesh Launch

Stable 2D mesh, manual bootstrapping, Proof-of-Entanglement, ML-DSA-87, SQLite with Zstd.

  • libp2p + GossipSub networking
  • Optimized signing (memory pool, batch 10–100×)
  • Operator dashboard, HTTPS via Caddy, UFW hardening
Shipped

Scalability & Optimization

Dynamic submeshes, priority routing, WASM contract SDK, ScyllaDB migration path.

  • Fee + geotag submesh rules (loadable .toml profiles)
  • WASM wallet and validator integration
  • SQLite → ScyllaDB migrate with dry-run & stats
In progress

3D Mesh & Self-Healing

3D mesh validation, quarantine auto-recovery, reputation, CUDA acceleration.

  • NVIDIA-locked validators enforced end-to-end
  • Topology WebSocket + live 3D viewer
  • Reputation + reorg-safe partition heal
Next
ARCHITECTURE

How the pieces fit.

A single QSDM node speaks libp2p to its peers, terminates TLS at Caddy, exposes a REST API and a Prometheus endpoint for operators, and optionally ingests NGC proof bundles from an NVIDIA sidecar.

Browser / SDKs Go · JS · WASM clients Caddy · TLS edge qsdm.tech / api / dashboard QSDM binary · API /api/v1/auth, /wallet, /health JWT · mTLS admin · audit log :8443 (HTTP behind Caddy) Operator Dashboard WebSocket topology · Prometheus Login → /api/v1/auth proxy :8081 Proof-of-Entanglement CPU-only · no proof-of-work Mempool · BFT · Finality gadget Storage SQLite + Zstd → ScyllaDB (opt) Row-level compression, MVs libp2p · :4001 GossipSub · PEX Tx · Blocks · Evidence NVIDIA NGC sidecar CUDA fingerprint · signed proof bundle POST /api/v1/monitoring/ngc-proof · HMAC-signed P2P gate Drop tx from peers w/o recent valid GPU proof

Ready to explore?

Log in to the operator dashboard, hit the public API, or read the source.

Open Dashboard /api/v1/health GitHub