Conservative
$0 net per month- Reward rate
- 60%
- Reference price
- 60%
- Gross value
- $0
Operator lab
Compute opportunity lab, for GPU operators.
Model the net value of adding Pearl proofs to AI work you already sell. The point isn’t a number to trust: it’s a test plan, showing which values you must measure before any production decision.
- Runs in your browser
- Your inputs never leave it
Your infrastructure
Every field is editableA label for your scenario: Pearl’s official miner supports only H100 and H200 today
GPUs
Running the workload you already serve
hrs
Busy with real work, not idle capacity
%
Work compatible with Pearl proofs: validate in a pilot
%
Share of eligible work that yields accepted proofs
PRL
A scenario input, not a quoted rate
USD
CoinGecko on Oct 22, 2026
%
Serving throughput lost to proving
$/GPU-hr
What an hour of this GPU earns today
W/GPU
Extra draw of the proof path
$/kWh
Your all-in rate
%
If you mine through a pool (Pearl itself charges no protocol fee)
%
Only when you settle in USD
Settlement
Converting needs a venue and adds fees and timing risk
Modeled monthly result
Scenario estimate · not liveNet incremental value
$0
…
- Total GPU hours
- 0
- Eligible accepted hours
- 0
- Gross PRL
- 0
- Gross modeled value
- $0
- Pool fee
- $0
- Conversion and settlement
- $0
- Incremental energy
- $0
- Throughput given up
- $0
- Net per eligible hour
- $0
All values are scenario assumptions, not live network data.
Sensitivity, not a forecast
Reward and price can move independentlyBase assumptions
$0 net per month- Reward rate
- 100%
- Reference price
- 100%
- Gross value
- $0
Upside
$0 net per month- Reward rate
- 130%
- Reference price
- 150%
- Gross value
- $0
How it’s calculated
From GPU hours to net value.
-
Hours that can carry proofs
GPU count × monthly hours × eligible share × proof success rate.
-
Gross PRL and value
Accepted hours × PRL per eligible hour, valued at the reference price.
-
Costs
Pool fee and (if you settle in USD) conversion fee on the gross; energy = extra watts ÷ 1,000 × eligible hours × $/kWh; throughput = eligible hours × your revenue per hour × overhead.
-
Net
Gross value minus all four costs. The headline ratio compares it with the revenue those GPUs already earn.
Before a pilot
What you’d have to measure.
- The real eligible share of your workloads, model by model.
- Accepted proofs per hour on your exact GPUs, drivers and miner release.
- End-to-end overhead on latency and throughput, not only on a benchmark. Pearl reports 3.9% to 5.08% on two setups.
- Where, how fast and at what cost you could sell PRL.