Thesis: Physical AI is starving for ground truth. Satellites can’t see sidewalks. Dashcams can’t see stairs. Corporate fleets can’t scale. The only viable infrastructure is 3 billion civilian sensors, already deployed in pockets. Vangrid turns them on.Documentation Index
Fetch the complete documentation index at: https://docs.vangrid.io/llms.txt
Use this file to discover all available pages before exploring further.
Phase 1 — Ignition
Activate the swarm.- Vangrid Web App live → Node 001 ships
- VideoQuest v1: pedestrian-level spatial capture pipeline
- On-device privacy filter operational (BlazeFace + SISO-SSD)
- PTC reward layer wired
- iOS submission
Unlock: Bootstrap the perception grid. Prove the swarm captures what satellites and fleets cannot.
Phase 2 — Distribution
Turn capture into category.- POIQuest: targeted missions on transit hubs, entrances, plazas
- RouteQuest: pedestrian-corridor coverage (sidewalks, alleys, stairs)
- MarketConnect: enterprise data licensing layer live
- First defense pilot signed
- First robotics/AV pilot signed
Unlock: Move from “we have data” to “enterprise pays for it.” Category creation.
Phase 3 — Moat
Build the IP layer.- EnrichQuest: metadata, object inventory, semantic labels on every capture
- Vault One: proprietary spatial dataset becomes core defensible IP
- World-model lab partnerships (frontier AI labs)
- Enterprise API v1
Unlock: Anyone can run nodes. Only Vangrid has this refined, enriched, cryptographically verified geometry. The dataset is the moat.
Phase 4 — Inevitability
Become infrastructure.- ValidatorNet: economic layer for data quality verification
- SyntheticQuest: real-capture × synthetic-training hybrid for World Models
- Enterprise Connect: full API stack for defense primes, robotics fleets, AV labs
- Sovereign data deployments (allied governments)
Unlock: Default spatial layer for Physical AI. Not a vendor. Infrastructure.
2027 → Endgame
The Spatial Cortex for every robot, drone, and autonomous agent in the physical world.- Real-time spatial API for autonomous fleets at city scale
- Embodied AI training contracts (humanoid robotics primes)
- Sovereign perception grids for allied defense
- Vangrid Protocol: decentralized data marketplace, on-chain provenance
- $1B+ ARR run-rate
ground_truth_score variance in covered zones — denser corroboration means tighter confidence bounds, fewer contested observations, and higher data quality floors that no single-source competitor can replicate. Every new query adds signal to coverage prioritization, compounding the dataset’s value with each request.
The infrastructure layer for the multi-trillion-dollar Physical AI race.