- PEARL has no premine or founder/investor allocation, but the lab’s ~$130M financing leaves mined-supply ownership and GPU-fund participation unresolved.
- Launch difficulty lag minted 12.3% of PRL on day one; current issuance is ~1.15M PRL daily, or ~$2M and ~23% of volume, implying 83–97% supply growth over 12 months.
- Mining is concentrated: one pool found 52.6% of the last 500 blocks and the top five found 91.8%; PEARL trades at $3.7B FDV without disclosed holdings, locks, or token-linked inference revenue.
- The author closed the long and is waiting for disclosures, locks, GPU clarification, and token value capture; Coinbase’s roadmap listing is the next liquidity test, but has no date.
Investment pitches
filter:
85 70 - STANDARD’s bull case is a redesigned reflexive token system: ETH outflows cut banker pay immediately and fees fund buybacks; after five weeks, 26.4M tokens were destroyed versus 18M issued, leaving supply below genesis.
- The larger opportunity is market-making stock tokens: the reserve put 1,318 ETH into SpaceX, Alibaba, and SK hynix tokens, while its Alibaba and SK hynix pools are 17x and 12x deeper than Robinhood Chain’s next-largest pools.
- The author will keep banking and remains bullish, but key links are unproven: a cheaper SpaceX pool captures most volume, stock-pool fees go to a team multisig rather than directly to STANDARD, and attention may not persist.
83 - The bull case is that GPU matrix multiplications can secure the chain while serving inference, converting security spend into useful compute rather than wasted hashes.
- Distribution improved after Coinbase added PRL to its listing roadmap. Pearl’s API quotes DeepSeek V4.1 Flash at $0.05 input/$0.30 output versus Together AI’s $0.30/$1.20.
- The key gap is paid demand: the network discloses no share of hashrate serving paid workloads, while ~1M PRL is issued daily against an ~$815M float.
- Author remains long. Entropy’s io:PRL market grew OI from $3.05M to $9.3M, but its oracle remains 90% SafeTrade and longs paid 0.7–0.8% daily.
74 - ARX is framed as a $49M-mcap alternative to Venice’s $1.3B: Blackthorn removes Intel TDX, whose attestation was reportedly extracted and forged twice, while NVIDIA confidential GPU memory survived those attacks.
- Blackthorn runs plaintext models inside confidential GPU memory, splits session keys across MPC nodes, and offers API, rented-GPU, or self-hosted deployment; near-native speed is claimed but unbenchmarked.
- At $234M FDV versus Phala’s $59M, ARX already trades at a roughly 4x infrastructure premium, which the author reads as early pricing for a Venice-style application layer.
- The bet remains conditional: Blackthorn is gated alpha with no fees, whereas Venice has users and revenue; adoption and an app layer must arrive for the valuation gap to close.
74 - PRL’s core bet is proof-of-useful-work: chain-seeded noise forces miners to perform matrix multiplications, making each GPU run produce inference output and a block-winning ticket.
- Pearl’s public math, live chain, and Hawkeye verifier—accepted at MLSys 2026—support the mechanism; its API lists GLM-5.3 Flash at $0.07/$0.24 per 1M tokens versus Together’s $0.15/$0.50.
- Author bought the dip and is long: at $1.70, PRL’s ~$594M market cap is ~19% below its $2.10 ATH; upside depends on inference demand joining the mining economy.
- The unresolved test is whether paid workloads actually run through mining; without adoption, PRL must absorb ~1M new tokens daily, while thin spot books and fragile perp pricing amplify risk.