The flagship economic primitive

Useful Proof of Staking

Stake QUBIC to reserve intelligent compute. Use what you need. Earn from what you don't. A new economic primitive where capital and infrastructure grow together.

0.00B QUBIC
Total staked
0.00%
Avg. annual yield
0
Compute regions
0.00%
Settlement uptime

The mechanism

How capital becomes infrastructure.

Useful Proof of Staking aligns three groups: stakers, operators, and users. The result is a self-reinforcing loop where value flows back to those who contribute capital and capacity.

1. Stakers

Lock QUBIC to reserve compute capacity. Larger stakes unlock more throughput, dedicated clusters, and governance weight.

  • Reserved capacity
  • Burn discounts
  • Governance voting

2. Operators

Run verified hardware. Validate computation. Earn fees from network usage and staking rewards.

  • Block validation
  • Compute rewards
  • Quality bonuses

3. Users

Consume AI services. Pay protocol fees. Drive demand for compute that flows back to stakers and operators.

  • OpenAI-compatible APIs
  • Pay per token
  • Volume discounts

The flow

Where every QUBIC goes.

Aigarth's economic loop is transparent. Every transaction, every fee, every reward ” on a verifiable path.

Stake
Lock QUBIC
Reserve
Capacity assigned
Use
Run inference
Fees
Pay per call
Burn
Tokens removed
Treasury
Funds growth
Note: Burn percentages and treasury allocations shown here are conceptual examples. Actual values are governed by on-chain votes and may vary.

Economics

Numbers, transparent and verifiable.

These figures illustrate how the model is designed to work. They will be replaced with on-chain data once the protocol reaches mainnet parity.

Reserve ratio
1M QUBIC ≈ 8 GPU-hr / day
Indicative allocation ratio
Staker yield
6% “ 12% APY
Range across plan tiers
Illustrative burn
15% of protocol fees
Conceptual example
Slashing
Up to 5% on downtime
Operator-only

Timeline

A stake is a relationship, not a transaction.

Day 0
Stake

QUBIC locked in the reserve contract.

Day 1
Allocation

Capacity assigned to your account.

Ongoing
Use

Run inference, embeddings, training, agents.

Daily
Settlement

Protocol fees paid, rewards distributed.

Quarterly
Governance

Vote on parameters, burn rate, treasury.

Anytime
Unstake

Cool-down period applies. Capacity returned.

Comparison

How Aigarth compares.

vs AWS

Decentralized cloud
Ownership
Staked capacity
Recurring rent
Exit
Unstake anytime
Migration required
Idle cost
Earn from spare
Pay regardless
Pricing
Token-efficient
USD-billed
Vendor lock
Portable APIs
Proprietary stack

vs OpenAI

Open inference layer
Source
Distributed network
Centralized
Pricing
Staking discount
Pay per token
Monetize
Sell idle capacity
Not available
Governance
On-chain voting
Closed
Models
Multiple providers
Proprietary only

vs Traditional Staking

Productive capital
Yield source
Real compute usage
Inflation only
Backing
Reserved infrastructure
No backing
Utility
Use compute directly
Passive
Risk
Operator slashing
Inflation dilution
Returns
Fees + burn
Inflationary rewards

Interactive simulator

Estimate your yield.

Stakers earn from network usage and protocol fees. Higher tiers get burn discounts and priority routing. All numbers are illustrative.

10M500M
Base yield (8%)4,000,000 QUBIC
Burn discount (25%)-3,000,000 QUBIC
Annual est.5,000,000 QUBIC
Illustrative only. Actual returns depend on network usage, slashing, and governance parameters.

Live network

A living ecosystem.

Thousands of stakers, operators, and users connected. Every node is verified. Every flow is settled.

Active stakers
12,847
Operators online
1,204
Active models
47
Settlement latency
1.2s

Governance

Stakers steer the network.

The longer you stake, the more your voice weighs. Vote on burn rate, supported models, treasury grants, and protocol upgrades.

Active proposals

AIP-0422d left
Lower image generation fees by 12%
For78%
AIP-0415d left
Add Grok-2 to the model registry
For92%
AIP-0408d left
Increase enterprise SLA compensation
For61%

FAQ

Common questions.

Reserve compute. Earn while you build.

Pick a plan, stake QUBIC, and start running AI workloads in minutes.