PRL
Pearl · Proof-of-Useful-Work L1 (GPU matrix-multiplication mining pitched as AI compute)
Key facts
- Circulating 356.0M, 17%
- Locked or unvested 1.74B, 83%
- Max 2.10B
Market data: CoinGecko, Oct 24, 2026. Scores reviewed Oct 23, 2026.
Assessment
Pearl has no value engine: all fees go to miners, there is no burn, staking or inference marketplace, and PRL pays only transfer fees on its own chain. There was no premine, sale or VC, but difficulty-1 genesis minted 121.7M in six days, ~34% of all PRL issued, partly to fund the team. Kryptex mined ~49% of blocks over 30 days on a single-client chain under a year old, just short of a single-pool majority. Supply is the binding problem: ~1.0M PRL a day is mined, about 100% a year of the ~354M issued, on a curve that reaches ~646M by Oct 2027.
How value reaches holders
Nothing reaches holders: all fees go to miners, with no burn, staking, smart contracts or inference marketplace. Issuance follows S(h) = 2.1B*h/(h+650,226): ~342M mined by Oct 12, 2026, ~1.0M PRL/day, supply ~646M by Oct 2027 (+89%). 121.7M PRL came in the first six days at near-zero difficulty, holders unidentified. The one AI partner, Together AI, deprecated its Pearl endpoint on Aug 27, 2026
Score anatomy
Demand
6 / 45: the stronger of two routes, plus adoptionWhy would anyone hold it: does value reach the token, and is the network or product used today?
Nothing reaches holders by design
Under $0.5M/yr, or emissions-subsidized only
Pays transfer fees on Pearl; no other monetary use (panel 1).
Pays its own network's fees (gas never counts higher, whatever the size of the economy)high confidenceHow this is scored
No smart contracts or paying demand; a Jun 2026 preprint found the PoW produces no useful AI output, so mining is the only activity
No product use: the token is the product, or the network or product is not live or sits idlelow confidenceHow this is scored
Two routes, the stronger counts
Soundness
9 / 40Can a long-term holder trust the rules: supply, float, distribution and control?
~1.0M PRL/day mined (~100%/yr of the ~354M issued, Oct 2026) on S(h) = 2.1B*h/(h+650,226); supply reaches ~646M by Oct 2027.
Over 15%/yr, or a key, company, committee or multisig can mint without limit or redirect issuancehigh confidenceHow this is scored
No premine, allocation or vesting: all mined PRL is liquid; new supply is mining issuance, scored under supply.
Fully vested, no scheduled unlockshigh confidenceHow this is scored
No premine, sale or VC, but difficulty-1 genesis minted 121.7M in six days, ~34% of the ~353M PRL ever issued, partly to fund the team; holders undisclosed
Insiders received 25-50%, or one organization directs 10-30%medium confidenceHow this is scored
Kryptex mined ~49% of blocks over 30 days (52% over 24 hours on Oct 19, 2026) of a ~46 EH/s chain under a year old; one client; just short of a single-pool majority
A halt, rollback or exploited consensus bug in the last 24 months, attack cost under $10M, or a foundation-selected validator set or small multisig can halt or upgrade with no meaningful exit windowmedium confidenceHow this is scored
Durability
2 / 15Is it big, proven and reachable by institutions?
$708.6M market cap (band 3), $4.1M/day median volume (band 2), no flags.
Small: $50-250M market cap and $1M/day+How this is scored
Genesis Apr 27, 2026: 0.5 years old.
Under 1 yearhigh confidenceHow this is scored
No top-tier spot listing; trades on Ourbit, SafeTrade and BigONE (panel 0).
No top-tier exchangehigh confidenceHow this is scored
What would move it up
To reach T4 Speculative, one gate is still open:
- Supply 0needs 1(open)
Tier outlook
Stable
- Supply stays at 0 while mining adds more than 15% a year (~1.0M PRL a day now, ~0.7M by Oct 2027)
- Security: the top pool's share against Kryptex's ~49% of blocks over 30 days
- Revenue: a paid inference marketplace priced in PRL, or a new partner after Together AI's exit on Aug 27, 2026
- Access: a first spot listing on Binance, Coinbase, Kraken, OKX or Bybit, beyond Ourbit, SafeTrade and BigONE
- Usage: a shipped job-binding PoW upgrade that makes non-useful mining uneconomic
Research report
Oct 12, 2026 · source confidence 75%Pearl is a proof-of-work chain launched in April 2026 whose mining is a matrix multiplication meant to double as AI computation. There was no token sale and the code has no premine or developer fee, but issuance is front-loaded: about 342 million of the 2.1 billion cap exist today, more than a third of them mined in the first six days, and the block reward still adds roughly one million PRL a day, so supply roughly doubles over the next year. Transaction fees go to miners and there is no burn or staking. The AI link so far is one inference partner whose discount is funded by mining rewards, and an independent June 2026 study found that current mining produces no useful AI output. Hashrate is concentrated, with one pool above half. Trading is thin relative to market value and concentrated on two smaller exchanges, plus a bridged ERC-20 on Uniswap.
Structure (12 facts)
- Chain
- Pearl mainnet (native, Bitcoin-derived UTXO chain with Taproot transfers); WPRL ERC-20 on Ethereum via PearlBridge (~3.04M supply, ~0.9% of float, upgradeable proxy owned by a non-Safe contract)
- What it is
- Proof of Useful Work L1: PoW is exact integer matrix multiplication on GPUs, pitched as a side effect of AI training/inference (Komargodski, Schen, Weinstein 2025, arXiv 2504.09971); built by Pearl Research Labs (CEO Omri Weinstein)
- TGE
- mainnet 2026-04-27, no token sale, no premine or dev fee in code (genesis coinbase 0); CoinGecko history from 2026-06-22; CEX venues SafeTrade and BigONE (from 2026-09-18), no tier-1 listing
- Max supply
- 2,100,000,000
- Circulating supply (approx)
- 342,200,000
- Market cap (approx)
- $470,000,000
- Emission
- S(h) = 2.1B * h / (h + 650,226) at a 194 s target block, no halvings; block 126,600 on 2026-10-12, reward ~2,263 PRL/block (~1.0M PRL/day, ~$1.4M/day at spot); supply ~646M by Oct 2027 (+89%), then ~+33% the following year; half the cap by ~2030
- Launch concentration
- difficulty started at 1 with a slow adjustment: blocks 0-40,000 came in ~5.9 days (~15x design speed) and minted ~121.7M PRL, 35.6% of all PRL existing on 2026-10-12; the whitepaper says the low initial difficulty was partly to fund the core team, whose holdings are undisclosed
- Value accrual
- None to holders. PRL pays block rewards and transfer fees; all fees go to miners via a Bitcoin-style auction, no burn, no staking, no smart contracts, no inference marketplace (future plan only). The one AI partner, Together AI (Gemma-4-31B-it-pearl endpoint, 2026-05-15), prices inference >25% below list and funds the discount from PRL emissions, with customers paying in USD, so the AI business model depends on selling PRL
- Useful work
- Basu, arXiv 2606.04819 (2026-06): ~24 EH/s then (~320k GPU-equivalents, ~112 MW) produced zero useful AI output; the dominant miner has no inference code and the verifier accepted random matrices (44 pool-accepted shares from CPU/AMD/Apple). Single author, not peer-reviewed, no formal team rebuttal found; core devs concede there is no on-chain usefulness metric (TechFlow 2026-10-10). Real AI formats (BF16/FP8) not yet supported by the PoW
- Security
- hashrate ~45.9 EH/s across 39 pools (MiningBoard 2026-10-12); Kryptex ~54% and pearlhash.xyz ~29%, so one pool holds a majority; official pool takes a 20% fee and accepts only H100/H200; rank-penalty softfork 2026-08-06 (block 96,251); no third-party audit found
- Liquidity
- ~$3.5M/day CEX (SafeTrade ~67%, no trust score; BigONE ~33%), volume/mcap ~0.75% vs ~5% for TAO/RENDER; WPRL ~$0.5M/day on Uniswap V3/V4. Ticker collisions: Perle (PRL, Solana, the Coinbase/Upbit listings) and Parallel Governance (PRL) are unrelated
Market figures at report date (Oct 12, 2026)
- Price (approx)
- $1.37
- FDV (approx)
- $2,890,000,000
- All-time high / low
- 1.76 (2026-09-23); ATL 0.14 (2026-07-23), so +880% off the low; ~$1.65 in late May before the July crash
Sources (10)
- github.com/pearl-research-labs/pearl
- pearlresearch.ai/research/int-whitepaper
- arxiv.org/abs/2504.09971
- arxiv.org/abs/2606.04819
- explorer.pearlresearch.ai
- miningboard.com/pool-stats/pearl
- together.ai/blog/together-ai-partners-with-pearl-research-labs
- hashrateindex.com/blog/pearl-prl-ai-compute-cryptocurrency
- coingecko.com/en/coins/pearl-2
- etherscan.io/token/0x07696dcab55e62cfef953666b29fe1970518cb00
Compare with
How the ranking works and how PRL compares: see the full ranking. This entry as JSON:
GET https://tokens.10102.io/api/v1/tokenomics?symbol=PRL