Device Block Reward Multiplier

A detailed comparison of 2025 block reward multipliers by device type, with upcoming staking-related adjustments.

The io.net block reward system uses performance-based multipliers to fairly distribute rewards across a wide variety of hardware types. These multipliers are designed to reflect the relative utility, compute performance, and cost-efficiency of each device, ensuring the most effective hardware for AI workloads is appropriately incentivized.

Network Expansion and Updated Reward Structure

On June 4, 2025 at 3 PM UTC, io.net is significantly expanding its network with the addition of over 800 NVIDIA H200 GPUs. This is the largest single onboarding of enterprise-grade hardware in our history and is part of our commitment to scaling infrastructure for enterprise AI workloads like training and inference.

To ensure fair distribution of rewards as the network grows, the block reward system has been updated to reflect the following changes:

  • Updated device multipliers
  • New reward pool categories
  • Temporary staking grace period

Reward Pool Categories

Pool TypeDescription
🔴 No longer supportedLow-tier or outdated devices. Multiplier set to 0. No longer earn rewards.
🟡 Community PoolConsumer and mid-range devices. Earn 5% of block rewards.
🟢 Enterprise PoolHigh-performance processors. Earn 95% of block rewards.

Device Earning Multiplier Comparison

🔴 No longer supported

ProcessorCurrent MultiplierNew MultiplierNotes
M2 Max10Already dropped
M2 Pro0.750Already dropped
M2 Ultra1.250Already dropped
A100.750Dropped due to lack of demand and supply
A10G0.750Dropped due to lack of demand and supply
A1610Dropped due to lack of demand and supply
A4020Dropped due to lack of demand and supply
GeForce GTX 1080 Ti0.250Dropped due to lack of demand
GeForce GTX 2080 Ti0.250Dropped due to lack of demand
GeForce GTX 30500.250Dropped due to lack of demand
GeForce GTX 3050 Ti0.250Dropped due to lack of demand

🟡 Community Pool

ProcessorCurrent MultiplierNew MultiplierNotes
M30.51TFLOP-based adjustment with bonus for unified memory
M3 Pro0.751.25TFLOP-based adjustment with bonus for unified memory
M3 Max11.5TFLOP-based adjustment with bonus for unified memory
M40.51TFLOP-based adjustment with bonus for unified memory
M4 Pro0.751.25TFLOP-based adjustment with bonus for unified memory
M4 Max11.5TFLOP-based adjustment with bonus for unified memory
GeForce RTX 30600.250.75TFLOP-based adjustment
GeForce RTX 3060 Ti0.250.75TFLOP-based adjustment
GeForce RTX 30700.251TFLOP-based adjustment
GeForce RTX 3070 Ti0.251TFLOP-based adjustment
GeForce RTX 30800.251.5TFLOP-based adjustment
GeForce RTX 3080 Ti0.251.75TFLOP-based adjustment
GeForce RTX 30900.51.75TFLOP-based adjustment
GeForce RTX 3090 Ti0.52TFLOP-based adjustment
GeForce RTX 40600.251TFLOP-based adjustment
GeForce RTX 4060 Ti0.251TFLOP-based adjustment
GeForce RTX 40700.251.5TFLOP-based adjustment
GeForce RTX 4070 Ti0.252.5TFLOP-based adjustment
GeForce RTX 40800.52.75TFLOP-based adjustment
L40.750.25TFLOP-based adjustment
RTX 40000.40.75TFLOP-based adjustment
RTX 4000 SFF Ada Generation0.51TFLOP-based adjustment
RTX 50000.251.5TFLOP-based adjustment
RTX 5000 Ada Generation1.53TFLOP-based adjustment
RTX 6000 Ada Generation2.54.5TFLOP-based adjustment
RTX A40000.41.5TFLOP-based adjustment
RTX A50000.751.5TFLOP-based adjustment
RTX A60001.52TFLOP-based adjustment
Tesla T40.40.5TFLOP-based adjustment
Tesla V100-PCIE-16GB0.50.75TFLOP-based adjustment
Tesla V100-PCIE-32GB0.750.75TFLOP-based adjustment
Tesla V100-SXM2-16GB0.50.75TFLOP-based adjustment
Tesla V100-SXM2-32GB0.750.75TFLOP-based adjustment
Tesla V100S-PCIE-32GB0.750.75TFLOP-based adjustment

🟢 Enterprise Pool

ProcessorCurrent MultiplierNew MultiplierNotes
L40S2.253High-demand based adjustment
A100 80GB PCIe55No change
A100-PCIE-40GB22No change
A100-SXM4-80GB55No change
GeForce RTX 40900.751.25High-demand based adjustment
GeForce RTX 4090 D0.751.25High-demand based adjustment
H100 80GB HBM31010No change
H100 PCIe1010No change
H100 80G PCIe1010No change
B2003030No change
H2001512High-demand based adjustment

Overview of Staking Requirements

🚧

Important: Changes to Staking Requirements Coming July 1, 2025

Until July 1, 2025, staking requirements will continue to be calculated using the current earning multiplier.

To qualify for block rewards, each device must meet a minimum staking threshold calculated as follows:

  • Base Stake per processor: 200 $IO
  • Earning Multiplier: Varies by device performance
  • Number of processors: Total processors in the device
Minimum Stake = Base Stake × max(1, Earning Multiplier) × Number of processors

Starting July 1, 2025, staking requirements may increase for some devices, as a new calculation logic will take effect. For example:

  • A device with 8 H100 GPUs, each having an earning multiplier of 10:
    Minimum Stake calculation: 200 × 10 × 8 = 16,000 $IO
  • A device with 4 RTX 4070 GPUs, each having an earning multiplier of 0.25:
    Minimum Stake calculation: 200 × max(1, 0.5) × 4 = 200 × 1 × 4 = 800 $IO

📘

Even if the earning multiplier is less than 1, the multiplier used in the calculation defaults to 1 to ensure a minimum stake of 200 $IO per processor.

Upcoming Adjustments

With the integration of new H200 GPUs and the introduction of a Community Hardware Pool, the following changes are anticipated:

  • Updated Multipliers: Devices will have their earning multipliers recalibrated based on performance metrics.
  • Adjusted Staking Requirements: As multipliers change, the corresponding staking requirements will also adjust.
  • Grace Period: A one-month grace period will be provided before new staking requirements are enforced.

🚧

Important Note

Higher multiplier ≠ always higher rewards.
Actual rewards depend on several factors - including the number of eligible devices in a block, pool allocation (e.g., Community vs. Enterprise), and your device’s multiplier.

For more detailed information on staking requirements and calculations, please refer to the IO Staking Documentation.