Technical Whitepaper · v1.0

The 4D Data Layer
Architecture & Network Design

A community-powered, DePIN-native spatial data network — designed to become the world's largest continuously-updated 4D map, governed on-chain, and priced by quality.

4B+
Smartphone nodes addressable
3
Data grade tiers
$500k+
Saved per AV vs Google
Q3 2025
Targeted TGE
Executive Summary

A new infrastructure primitive for spatial AI

4D Labs is building the coordination layer that turns distributed human effort into structured, commercially-useful spatial data — at a cost and coverage impossible to achieve with conventional methods.

Spatial AI

World models, embodied AI, and autonomous systems require physical grounding. Current training data lacks continuous, real-world spatial coverage.

4D Data

Three spatial dimensions plus time. The network captures geometry, appearance, and physical change — not static snapshots.

Decentralized Contribution

4 billion smartphones replace expensive proprietary fleets. Participants earn rewards proportional to the quality and scarcity of their data.

On-Chain Reputation

Every contribution is scored, graded, and logged. Contributor reputation is verifiable, portable, and tier-gated — not siloed.

Core Thesis

Four claims the network is built on

§ 1.1
Spatial understanding is the AI frontier.
Text and images are solved. Physical space is not.

Large language models have saturated text-based benchmarks. Vision models can identify objects. Neither can reason about the physical layout of the world — the spatial relationships, geometry, and continuous change that real-world AI systems must navigate.

§ 2.3
DePIN replaces the $500k/vehicle model.
4 billion smartphones. Continuous updates. Distributed trust.

Google's Street View methodology costs over $500,000 per autonomous vehicle equivalent for data collection. A DePIN model distributes that workload to existing hardware — smartphones already in hands globally — with cryptographic verification replacing centralized quality control.

§ 4.3
Three data grades, one marketplace.
Raw → Processed → Semantic. Graded, priced, provenance on-chain.

A tiered quality system ensures that data buyers receive what they need and contributors are rewarded proportionally. Raw frames cost less and earn less. Semantic data with full object-class labeling earns premium rewards and commands premium prices from robotics and AV buyers.

§ 5.0
Scan-to-Earn with real yield.
Quality × scarcity × uptime. Protocol revenue drives buyback & burn.

Reward calculation is deterministic: quality score × regional scarcity coefficient × contributor uptime streak. Protocol revenue from the X3D Map API feeds a buyback-and-burn mechanism, creating sustainable, fundamentals-backed token value rather than pure inflation.

§ 2 · The Problem

AI can read text.
AI can see images.
AI can't yet understand space.

World models and embodied AI need physical grounding. That data doesn't exist at scale — and the existing collection methods are too slow and too expensive to fix it.

The spatial data gap is the core infrastructure bottleneck for the next wave of AI. Robotic systems, autonomous vehicles, AR/VR platforms, and spatial computing all require continuous, high-resolution, semantically-labeled 3D data that no existing dataset provides.

$500k+
per vehicle
Google-style 3D collection cost per AV data vehicle
1–2 yrs
update cycle
Average Street View refresh rate, versus real-world change happening continuously
Locked
data silos
Big Tech controls all high-quality 3D map data. No open access layer exists.
No layer
for contributors
No open, coordinated, incentivized network for distributed spatial data collection
§ 3 · Architecture

A complete spatial data stack — from collection to marketplace

01
DePIN Collection Network
4 billion smartphones become the data collection fleet. Distributed hardware eliminates the need for proprietary vehicles or centralized infrastructure investment.
02
Dynamic Bounty Engine
A supply-demand algorithm routes tasks to contributors in regions where data is most needed. Scarcity drives reward multipliers — not inflation.
03
Three-Tier Quality System
Raw → Processed → Semantic. Every submission is evaluated. Higher quality unlocks higher rewards and higher marketplace value. Quality is non-negotiable.
04
On-Chain Contributor Reputation
Scan Score is the verifiable, portable, and tier-gated reputation layer. It cannot be bought, only earned. Tier determines access to premium bounties and governance weight.
05
X3D Map API
The B2B delivery layer that routes verified spatial data to robotics, AV platforms, and spatial computing buyers. This is the protocol revenue engine.
06
DAO Governance & Real Yield
SIP (Spatial AI Improvement Proposals) govern network parameters. Protocol revenue from API usage feeds buyback-and-burn — linking token value to actual economic activity.
§ 4 · Data Economy

Three grades.
One marketplace.
Real value.

Every scan is evaluated on resolution, coverage, depth accuracy, and semantic completeness. Data grade determines reward tier and unlocks commercial marketplace access.

Buyers on the X3D Map API receive a verified, provenance-tracked dataset. Sellers (contributors) receive deterministic, quality-weighted rewards. The marketplace price discovery is transparent and on-chain.

Raw
100–400 pts
Point clouds, depth maps, sensor frames. Minimum viable spatial capture.
Accepts most capture hardware
Processed
400–1,000 pts
Cleaned, aligned, and structured spatial data with consistent coordinate systems.
Requires stable capture + xScanner alignment
Semantic
1,000–3,000 pts
Fully labeled with object classification, instance segmentation, and physical properties.
Head Mount or Surround Rig recommended
§ 5 · Market Context

Why this problem is solvable now

Three converging trends create a narrow window for a well-designed DePIN spatial network to become the canonical data layer before Big Tech consolidates the category.

Join the network now
01
Smartphone sensor parity

Modern smartphones carry LiDAR, stereo cameras, and IMUs that match or exceed purpose-built mapping hardware from 5 years ago. The fleet already exists. It just isn't coordinated.

02
World model demand inflection

Robotics and AV companies are entering their scale phase. GPT-4V and its successors have demonstrated that vision-language alignment works — but physical grounding remains the missing layer.

03
DePIN infrastructure maturation

Account abstraction, ERC-4337, and low-cost L2 settlement have removed the UX friction that killed early Web3 consumer apps. Contributor onboarding now takes under 60 seconds — no seed phrase.

§ 6 · Roadmap

Genesis → Global

Phase 1Live
Genesis
Q1 2025
  • Platform launch
  • Community bootstrap
  • Cold start incentives
  • Contributor tier system
Phase 2
Expansion
Q2 2025
  • xScanner App release
  • Regional bounty system
  • Marketplace beta
  • B2B API preview
Phase 3
Validation
Q3 2025
  • X3D Map API GA
  • TGE
  • DAO activation
  • SIP governance live
Phase 4
Scale
2026+
  • Global coverage targets
  • Full decentralization
  • Protocol self-sufficiency
  • Ecosystem grants

Ready to be part of the network?

The whitepaper describes the architecture. The task platform is where it runs.