Chart PRL Price
-
-
Pearlonomics
Pearl Network economics: PRL price, production cost, and chain fundamentals. The status pills below show where each series comes from.
Loading live feeds
Price source pending
PRL Index pending
Chain source pending
GPU rate pending
GPU H200 pending
GPU B200 pending
GPU A100-80GB pending
GPU L40S pending
GPU RTX4090 pending
GPU RTX5090 pending
PRL Index
-
-
Median Cost
-
-
Price / Cost
-
-
Block Reward
-
-
Network Scale
-
-
Live Price and Production Cost
The price line uses hourly PRL Index fixings where they exist (see PRL Index Methodology below). The cost line is calculated from live network hashrate, observed block pace, live height, and the live H100 marketplace rate.
Waiting for live data.
Network Hashrate
-Hashrate pending.
H100 Rate Index
-GPU rate pending.
OTC Closes in Sequence
-OTC closes in sequence — not to a time scale. The pearl-otc.com chart gives prices in order but no time for each price, so this panel plots the order only. Do not read dates from it.
OTC sequence pending.
Live Source Selection
| Market | Last | 24H Volume | Status |
|---|
No source data loaded yet.
Production Cost Inputs
GPU rate model
live Vast.ai H100 spot
Marketplace rate
-
Median H100 throughput
550 TH/s
Lower band scenario
620 TH/s
Upper band scenario
500 TH/s
Rate month
-
Emission formula
live height
Block pace
observed; 194s fallback
Pool fees / stales
excluded
Block reward
reward_prl = 2.1B * 650226 / ((height + 650226) * (height - 1 + 650226))
Observed block pace
blocks_per_hour = delta_height / delta_hours; live_partial = blended(recent_completed_pace, network_avg_block_time)
Cost per PRL
marketplace_cost_per_hour = latest_h100_snapshot_at_or_before(timestamp).median_usd_per_gpu_hour
gpu_equivalent = network_hashrate_hps / scenario_hashrate_hps
cost_per_prl = (gpu_equivalent * marketplace_cost_per_hour) / (reward_prl * blocks_per_hour)
The H100 rate is the median per-GPU on-demand price across all rentable H100 SXM, NVL, and PCIe offers on Vast.ai. For each cost point, the page uses the most recent H100 snapshot at or before that time (the earliest snapshot for points before the series starts). Snapshots before live tracking began are Silicon Data monthly seeds. If the live H100 feed is unavailable, the page uses a static monthly table, then a fixed fallback rate. The lower, median, and upper bands all use the same H100 rate. The band captures H100 throughput sensitivity: 620 TH/s efficient, 550 TH/s median, and 500 TH/s conservative. Provider-class rate spreads are excluded because rational miners should use the lowest available rental market, while already-allocated compute is closer to sunk cost than marginal rental cost. Completed buckets use observed pace; the still-forming live bucket is smoothed with recent completed pace and Pearl node network average block time.
The production-cost series is calculated, not fetched as a market price. It uses Pearl chain hashrate history (from our own Pearl node, or from lordofpearls.xyz when our node feed is stale; the chain status pill shows which) plus the live H100 marketplace rental rate — the median per-GPU on-demand price across rentable H100 offers on the Vast.ai marketplace, snapshotted into D1 over time. Pre-marketplace-tracking history is seeded from Silicon Data H100 marketplace-tier medians, and a static marketplace table remains the fallback if the live feed is unavailable. The PRL price line is the PRL Index where hourly fixings exist; the chart note names the series in use. Lord of Pearls is used only as a fallback for chain history.
PRL Index Methodology
The PRL Index is the reference price for PRL in US dollars. The methodology version is prl-index/1. The Index uses only payloads that this site has stored in its raw feed archive. It makes no new external call when you read it.
At any time T, the Index first gets one price from each venue that has usable data at or before T. Data after T is never used. From pearl-trade.com, the price is the volume-weighted average price of trades in the last 60 minutes before T. If there is no trade in that hour, the price is the last trade, if that trade is not more than 6 hours old. From CoinGecko, the price is the spot price in the last stored payload, if that payload is not more than 2 hours old. From pearl-otc.com, the price is the last point of the chart payload, stamped at the time we stored that payload, if it is not more than 2 hours old. The pearl-otc.com payload has no time for each point, so no point gets an invented time.
Lord feeds are not Index inputs. Lord republishes prices from other venues. It is not an independent venue, so its price would count one venue two times.
The Index value is the median of the venue prices. At least two venues are necessary. If fewer than two venues have usable data, there is no Index value. The endpoints report the state insufficient_venues. A single venue price is never published as the Index. Each result also gives the lowest and highest venue price, the largest difference from the median in basis points, and the price, age, and basis for each venue.
A fixing is the Index value for one completed UTC hour. The fixing for the hour that starts at H is computed at the close of that hour, H plus one hour, from archive rows at or before the close. The same archive always gives the same fixings. Fixings are filled in on read, oldest first, at most 48 per request. An hour with fewer than two venues is stored with no value, so the gap stays visible.
The Index is fresh if the newest venue observation it used is not more than 2 hours old when you read it. If it is older, the value is still served, but it is marked stale. Read the live value at /api/prl-index/current. Read the hourly fixings at /api/prl-index/history. Get the same fixings as CSV at /api/prl-index/history.csv. CSV rows carry timestamp_basis=fixing and source=pearlonomics-prl-index.
Changelog: prl-index/1 — 2026-10-15 — initial methodology.