
Pearl (PEARL) activated its FP8 mining upgrade on its test network on November 17, 2026, setting the fork at block 67,370 in release 1.6.0-fp8-testnet1, as part of a broader push to let artificial intelligence providers mine tokens using existing inference workloads. Following the milestone and recent major US exchange listings on November 13, PEARL surged roughly 27% over 24 hours and 62% over the trailing seven days to trade near $2.33 on November 19, lifting its circulating market capitalization to approximately $897 million.
Despite the sharp rally, independent testing and draft governance rules introduce significant operational friction ahead of mainnet deployment. Benchmarks from Kairos Research found that FP8 mining consumes far more server capacity than Pearl Research Labs estimates, while the draft PIP-3 hardware rules would exclude older Nvidia, AMD, and Intel GPUs, potentially cutting out roughly two-thirds of current network hashrate.
How Pearl (PEARL) Turns AI Work Into Mining
Pearl is a proof-of-work blockchain forked from Bitcoin. Its miners do not solve puzzles that have no other use. They run the same matrix calculations that AI models perform when they answer prompts, so a GPU serving AI customers can mine PEARL at the same time.
PEARL is listed on two major US exchanges, both of which added it on November 13, and crypto investment firm Syncracy Capital has disclosed a large position in the token.
What Pearl’s FP8 Upgrade Changes
Pearl’s main network currently counts only calculations done in INT7, a number format few AI providers use. FP8 is a format they already use, so they would not have to convert their models to mine.
Pearl Research Labs says mining would slow AI work by 1% to 5% under FP8, compared with 10% to 15% under INT7. Kairos Research tested the FP8 build on a rented Nvidia H100 and found that mining still cost 20% to 40% of the server’s output, well above Pearl’s estimate.
Why Two-Thirds of Miners Face Exclusion
Under draft improvement proposal PIP-3, only Nvidia Hopper and Blackwell architecture GPUs would remain eligible to mine following the fork. Legacy Nvidia cards alongside all AMD and Intel GPUs would be excluded from producing valid block certificates, a restriction Kairos Research estimates would eliminate roughly two-thirds of existing network hashrate.
Today, most of the Pearl mining appears disconnected from real AI workloads, as miners can multiply randomized matrices solely to earn block rewards. Blockworks estimated earlier this month that fewer than 1% of mined blocks correspond to genuine AI model inference. By adopting FP8, Pearl aims to make dual-purpose AI compute more economically attractive than synthetic “junk mining”.
What PEARL Traders Should Watch Before the Mainnet Fork
Pearl has not set a date to bring FP8 to its main network, and the testnet release keeps the fork switched off on mainnet. A software release naming a mainnet fork block is the next thing to watch.
Traders will also watch how miners with excluded GPUs respond, and whether the GPU rules in PIP-3 change before the upgrade reaches mainnet.
Where Pearl (PEARL) Stands After the Testnet Fork
PEARL holds a roughly 62% weekly gain near $2.33 as FP8 undergoes live testing on its test network. A mainnet rollout date remains unannounced, and the hardware restrictions outlined in PIP-3 remain in draft status.
Disclaimer: Crypto products and NFTs are unregulated and can be highly risky. There may be no regulatory recourse for any loss from such transactions. The information provided here is for general informational purposes only and does not constitute investment advice. Please do your own research before making any investment decisions.


