Up nearly 10x in 4 months, Pearl's AI mining narrative gains traction, but this "pearl" still faces multiple challenges
Mining while running AI models? Pearl, an L1 project that has recently gained traction, is trying to turn this idea into reality. Pearl attempts to reuse the GPU matrix multiplication required for AI inference in blockchain mining, achieving "one computation, dual rewards."
As PEARL's price keeps climbing, the "AI version of Bitcoin" narrative continues to heat up, with several well-known crypto institutions voicing optimism and pushing Pearl's market heat ever higher.
Token surges nearly 10x in four months, but concentrated holdings and insufficient liquidity remain concerns
Over the past period, PEARL has kept refreshing its price and become a focal point of the crypto market.
CoinGecko data shows that over the past four months, PEARL rose from a low of $0.23 to $2.2, a cumulative gain of nearly 9.5x. Its market cap has now risen to $760 million, ranking among the top 100 crypto assets globally by market cap. In terms of both price gains and market cap performance, PEARL has been one of the more eye-catching assets in the recent market.
Alongside the price rise, PEARL's on-chain activity has also shown some growth momentum. Pearlhub.ai data shows that as of October 21, the number of active PEARL holding addresses has exceeded 68,000, and the recent daily number of new addresses has kept climbing, most recently reaching 4,982. However, the token distribution is relatively concentrated: the Top 100 addresses hold 42.12% of the tokens, and the Top 10 addresses hold 14.15%. In other words, although the number of holding addresses has grown, the tokens remain concentrated in the hands of a small number of addresses.
Looking at on-chain transfers, recent daily transfer volume and transaction counts have both hit new highs. Judging by the holding period of transferred tokens, tokens held for more than 90 days account for about 65.8%, those held for 30 to 90 days account for 17.5%, and tokens held for 7 to 30 days, 1 to 7 days, and less than 1 day account for 9.5%, 4.4%, and 3.1%, respectively. This indicates that the tokens transferred recently mainly come from longer-held positions, while the share of short-term held tokens is relatively low.
At the same time, the concentration of transaction amounts is also worth noting. Over the past 30 days, PEARL's cumulative transaction count exceeded 180,000, but the top 1% of transactions (291 transactions) contributed 96.3% of the total transaction amount.
However, PEARL's liquidity in the secondary market remains low. CoinGecko data shows that over the past 24 hours, PEARL's trading volume was about $9.86 million, of which 98.8% was concentrated on three relatively small trading platforms — SafeTrade, BigONE, and Ourbit — so liquidity and trading depth still need improvement.
Recently, Prep DEXLighter and Entropy, a HIP-3 market deployer, listed PEARL contracts, with daily trading volume in the millions of dollars; Entropy also spent 500 HYPE to successfully win the trading ticker PEARL. Coinbase and Kraken have successively added PEARL to their listing roadmaps, and digital asset custodian BitGo has also partnered with Pearl, bringing new expectations for expanding trading channels and improving liquidity in the future. In addition, Pearl renamed its token ticker from PRL — which was the same as another project — to PEARL, which the community also sees as a positive signal for paving the way for future listings.
PEARL's current insufficient liquidity is related both to its underlying architecture and to its token distribution mechanism.
At the technical level, Pearl adopts an independent UTXO architecture, similar to Bitcoin, rather than being an ERC-20 token issued on Ethereum. This means that if exchanges want to support PEARL, they need to adapt to its underlying chain; if DEXs want to establish a corresponding trading market, they also need additional technical support and time to build liquidity.
In terms of token distribution, Pearl has no pre-mine and no tokens specifically allocated to the team or VCs; all tokens are generated through mining. This mechanism avoids the concentrated selling pressure that could come from teams and early investors holding large amounts of pre-allocated tokens, but it also means the project lacks liquidity support from professional market makers or institutional investors in its early stages. Trading demand in the public market relies more on miners selling tokens and naturally formed buy and sell orders. Over-the-counter trading has to some extent made up for the lack of public market liquidity — for example, the OTC market Pearl OTC has accumulated trading volume of nearly $33 million.
Turning AI inference into mining tasks, Pearl has been "called out" by multiple well-known institutions
As an L1 public chain focused on PoUW (Proof of Useful Work), Pearl attempts to combine AI computation with blockchain mining, exploring a new development path for the traditional PoW mechanism. Its mainnet went live in April this year.
Unlike Bitcoin, which maintains network security through hash computation, Pearl attempts to use matrix multiplication (GEMM), common in AI training and inference, as the core computational task of mining, and verifies the computation results through cryptographic proofs. In theory, this allows the same GPU computing power to both generate the computational proofs required for mining and provide computing services for tasks such as AI inference.
This approach targets the long-standing economic problem of traditional PoW networks. As Bitcoin block rewards gradually decrease, miners need to rely on transaction fees or other income to cover mining costs, and the rise of AI computing businesses provides a new monetization channel for GPUs. Pearl hopes to let miners earn token mining rewards while providing computing services to AI clients, using business revenue to share hardware and operating costs and reduce miners' dependence on selling newly mined tokens. For this reason, the community also regards Pearl as the "AI version of Bitcoin."
The team background is also one of the reasons Pearl has attracted attention. The founding team includes Omri Weinstein, a Princeton University mathematics PhD and professor of computational complexity theory at Columbia University, and cryptography researcher Ilan Komargodski, among others.
At the application level, Pearl's long-term goal is to build a useful computing power market, connecting computing power supply and demand and exploring on-chain settlement of computing contracts, so that the network gradually establishes a connection with real AI computing demand. Around this direction, Pearl has advanced several related business attempts, including cooperating with Together AI to provide inference services, launching an inference platform compatible with the OpenAI API, and providing cloud computing power mining services. Among them, in the cooperation with TogetherAI, Pearl is trying to attract users through price incentives, offering discounts of more than 25%, with part of the price advantage subsidized by expected PEARL token emissions.
In addition to competing on price, computational privacy and result verifiability are another technical route through which Pearl tries to differentiate itself. The protocol has adopted zk-SNARK to protect mining proofs, preventing matrix data from leaking during verification; at the same time, the project is working with Attestable to develop a privacy-preserving inference proof mechanism, aiming to let users verify whether inference services are executed as expected while protecting data privacy.
Recently, Pearl has also gained public attention from multiple well-known crypto institutions, further boosting the project's heat. For example, No Limit Holdings partner Brian Breslow regards Pearl as the "next-generation Bitcoin," calling it one of the investment targets he is most confident in, and even suggesting that the token price could reach above $30; Dragonfly partner Tom Schmidt also gave a positive evaluation of Pearl's mechanism design in related interviews, favoring the idea of introducing AI computation into PoW; Stripe's CEO also mentioned Pearl on social media; DWF Ventures pointed out that the project was a hot topic among institutional funds and retail traders during the TOKEN2049 Singapore conference, and its early characteristics of a certain mystique and relatively high participation threshold are reminiscent of Bittensor's early days.
From useful computation landing to token value capture, Pearl still faces multiple tests
Although Pearl has gained market attention with its innovative narrative of combining AI computation and blockchain mining, it still faces multiple challenges in technical implementation, token value capture, sustained inflation, and computing power concentration.
On one hand, whether useful computation can truly land remains the core technical controversy facing Pearl. Pearl's original design is to let miners complete mining while executing AI inference tasks, but the fact that the protocol can verify whether matrix multiplication is executed correctly does not mean these computations necessarily serve real AI demand. Miners can still use random matrices to complete computations and earn token rewards without providing inference services to paying clients. Therefore, although Pearl provides the possibility for useful computation in its mechanism design, it has not yet proven that most mining activity can be converted into AI services with real economic value.
On the other hand, a clear value capture mechanism has not yet been established between AI business growth and PEARL token demand. Pearl hopes to use PEARL to pay for AI inference fees through a computing service market in the future, but at this stage users mainly purchase services through USD-denominated prepaid credits, while miners obtain PEARL through mining — no clear token purchase, payment, or burn mechanism has yet been formed between the two. Therefore, even if AI inference business continues to grow in the future, it does not mean PEARL demand will necessarily increase in tandem. In the absence of a direct value capture mechanism, PEARL's current token valuation still relies to a large extent on market expectations of its technical narrative and the balance of miner selling.
At the same time, the new supply brought by continuous mining may exert long-term pressure on the market. PearlTrack data shows that, based on current block rewards and block production speed, Pearl's daily new issuance is about 1.2 million tokens. At PRL's current price, the nominal value of daily new tokens is about $2.628 million, equivalent to 26.7% of the $9.86 million trading volume over the past 24 hours. From an annualized perspective, PEARL's annual new issuance is about 438 million tokens, and at the current price, the corresponding annualized issuance value is close to $960 million, about 126% of its current $760 million market cap. It should be noted that new issuance value does not equal actual selling pressure. Miners may sell some tokens to cover costs such as electricity and hardware depreciation, or may choose to hold or sell in batches.
Looking at the longer-term issuance structure, Pearl's maximum supply is 2.1 billion tokens, all generated through mining. Because early mining difficulty was extremely low, the first 40,000 blocks were mined in just 6 days, during which about 121.7 million PEARL were issued. Based on the current cumulative mined amount of about 340 million tokens, this early issuance accounts for about 35.8% of the cumulative mined amount.
As of now, PEARL's cumulative mined amount accounts for about 16.2% of the maximum supply. This is because Pearl adopts a smooth polynomial decay mechanism rather than Bitcoin's periodic halving model. Based on the current issuance curve, about half of the token supply is expected to be mined by around the end of 2029. This means that a large number of tokens will continue to enter circulation over the next few years. If the growth of AI inference demand, actual PEARL usage demand, and market liquidity cannot keep up with the new supply, miners' selling pressure may persist. Once the market narrative cools, the supply-demand imbalance may further amplify price volatility.
In addition, Pearl's computing power concentration problem also brings concerns about its degree of decentralization and network security. Pearlhub.ai data shows that two major mining pools, Kryptex and PearlHash, occupy 72.2% of the market share, with network computing power concentrated in the hands of a few mining pool operators.
Overall, Pearl's innovation lies in attempting to combine AI inference with blockchain mining, but whether this concept can be transformed into a sustainable economic model still depends on factors such as the growth of real computing demand, the improvement of the token value capture mechanism, and the market's ability to absorb new supply.
Join the official Coincamps community:
Telegram: https://t.me/coin_camps
