@dphnAI

Dolphin

Dolphin is an AI lab developing uncensored LLMs and a distributed GPU inference network. The project has legitimate technical substance with 5M+ monthly HuggingFace downloads and multiple live model releases (X1 series, Venice collaborations). They've launched a token-incentivized distributed inference network on Base ($POD, formerly $DPHN) where GPU providers earn rewards for running models like Qwen 35B. The network is operational with 300+ GPUs and 24TB vRAM capacity.

𝕏 @dphnAI ↗ t.coAIpromising

＋ Watch

AI Analysispromising

Confidence

78%

Dolphin is an AI lab developing uncensored LLMs and a distributed GPU inference network.

The project has legitimate technical substance with 5M+ monthly HuggingFace downloads and multiple live model releases (X1 series, Venice collaborations).

They've launched a token-incentivized distributed inference network on Base ($POD, formerly $DPHN) where GPU providers earn rewards for running models like Qwen 35B.

The network is operational with 300+ GPUs and 24TB vRAM capacity.

Green flags: Verifiable technical work: 5M+ monthly downloads on HuggingFace, multiple model releases visible on public repos · Working product: distributed inference network live with 300+ GPUs, 9400 tokens/sec throughput, paying node operators · Strong partnerships: collaborations with Venice AI, sponsorships from DeepInfra, Targon Compute, Prime Intellect · Real token utility: $POD rewards distributed to node providers based on contribution metrics · Early-stage: 6K followers, account created Jan 2025, small market footprint despite technical progress

Red flags: Uncensored AI positioning could attract regulatory scrutiny or reputation risk · Token migration already occurred (Base CA change) which adds complexity for early holders · Limited visibility into team identities or backgrounds on public channels

Token

$POD

Chain

Base

Stage

mainnet+live