◢ DePIN · DECENTRALIZED PHYSICAL INFRASTRUCTURE NETWORK

Rent ethical
AI compute
Powered by idle
Apple Silicon

DePIN — a Decentralized Physical Infrastructure Network for AI inference. Developers get cheaper, sovereign compute. Mac owners earn passive EURO — without ever giving up their data, their device, or their control.

ORCHESTRATOR.logLIVE
> shard.dispatch → mac_m4_8821
  job#a21f · 412ms · llama3-70b
> node.authenticate(mac_m4_8821) ✓
  identity token verified
> result.return(dev#0x9f…a21)
  +1,284 tokens · 218ms
> payout.euro 0.0042 → 0x9f…a21
  settled · base-mainnet
> fleet 8,412 macs · 99.2% trust
FLEET NOMINAL
MAC_M4 · ONLINENODE · AUTHENTICATEDEURO · 0.0042 → 0x9f…a21SHARD · 412msMAC_M3_PRO · ONLINEPAYOUT · SETTLEDJOB · llama3-70bREGION · EU-WESTMAC_M4 · ONLINENODE · AUTHENTICATEDEURO · 0.0042 → 0x9f…a21SHARD · 412msMAC_M3_PRO · ONLINEPAYOUT · SETTLEDJOB · llama3-70bREGION · EU-WESTMAC_M4 · ONLINENODE · AUTHENTICATEDEURO · 0.0042 → 0x9f…a21SHARD · 412msMAC_M3_PRO · ONLINEPAYOUT · SETTLEDJOB · llama3-70bREGION · EU-WEST
SYSTEM ACCESS CONTROL
isolated user-space execution
SET A MAX % CPU CAP
your Mac stays yours
ZERO DATA RETENTION
shards purged post-job
DETERMINISTIC ROUTING
every job verified by the control plane
MACS ONLINE
8,412
TOKENS / SEC
1.42M
$ / 1M TOKENS
0.18USD
EURO PAID OUT
428.6K
◢ 01 / WHAT IS DePIN

We are a DePIN company.
Decentralized Physical Infrastructure Network.

DePIN — short for Decentralized Physical Infrastructure Network — is a new category of internet infrastructure where real-world hardware (compute, storage, energy, wireless) is contributed by thousands of independent operators instead of owned by a single hyperscaler. Contributors are paid directly, on-chain, for the resources they share.

AI MINING is a DePIN for AI inference. Instead of renting GPUs from a centralized cloud, developers tap into a global fleet of idle Apple Silicon Macs — each running jobs inside segregated user-space environments, with resource orchestration managed entirely by our secure control plane. The result: cheaper, sovereign, verifiable compute owned by the people who power it.

PHYSICAL

Real Apple Silicon hardware in real homes and offices — not virtualized cloud GPUs.

DECENTRALIZED

Thousands of independent Mac operators. No single point of failure, no hyperscaler lock-in.

NETWORK

An orchestrator routes inference jobs by latency, trust score, and region — paid per token, settled after reaching 100 Euros.

◢ 02/ GREEN ENGINE

The Green Engine of the New Web.

High-performance computing without the high environmental cost. Water-free data processing, on Apple Silicon.

DePIN MINING NODEAMBIENT-AIR COOLED
WATER / QUERY
0.00 L
POWER PROFILE
~30 W
HYPERSCALE DATA CENTEREVAPORATIVE COOLED
WATER / DAY
1M+ GAL
POWER PROFILE
700+ W
01

Zero Water Waste

Hyperscale AI campuses evaporate millions of liters of potable water per day to chill their GPUs. Our nodes are ambient-air cooled — the fan you already have. Net water draw per inference: zero.

02

Performance Per Watt · 5×

Apple's 3nm silicon delivers up to 5× the tokens-per-watt of a comparable data-center GPU. Same answer, a fraction of the grid load. Sustainable AI infrastructure, not a marketing label.

03

Built on Recycled Innovation

Every job runs on hardware that already exists in homes and offices. No new racks poured, no new substations built. Low-power decentralized computing that extends the life of the silicon you own.

◢ CLOUD COMPUTING, MINUS THE SMOKE & MIRRORS

One AI query on a hyperscale GPU evaporates roughly the water in a shot glass. The same query on our network evaporates none.

POWERED BY 3NM SILICON
WATER SAVED · LAST 24H
0LITERS

EST. NODES × QUERIES × 0.025 L — DISPLACED FROM EVAPORATIVE COOLING TOWERS

◢ 03 / HOW IT WORKS

Three steps. One global Mac fleet.

01LIVE →

SUBMIT

Developer posts an inference job through the async REST API with SSE streaming. Pay-per-token, no commitment, no GPU contract.

02LIVE →

SHARD

The Orchestrator splits the workload across thousands of idle Apple Silicon Macs — routed by latency, trust score, and region.

03LIVE →

EXECUTE

Each shard runs inside a segregated user-space environment on the Mac. Results returned in milliseconds. Mac owner will be paid in EUROs

◢ 04 / FLEET INTELLIGENCE

A distributed supercomputer, one Mac at a time.

No GPU farms. No data centers. Just thousands of consumer Apple Silicon devices, coordinated in real time, verified per job.

LATENCYp99 < 180ms
MACS8,400+ across 42 countries
ISOLATIONUser-space sandboxing, every job
PAYOUTSEURO on Base, settled hourly
◢ 05 / JOIN THE FLEET

DePIN Compute is the new
commons

Two ways in. Pick the side of the marketplace that pays you.