PEARL’s AI-Mining Pitch Faces a Testnet Reality Check

As reported on November 19, 2026, Pearl’s FP8 upgrade had gone live on testnet the previous day. PEARL had also risen over the preceding week despite a weaker broader crypto market. The question behind the price move is whether Pearl’s proposed AI-mining model can attract real inference work—not merely miners seeking token rewards.

Pearl is a Bitcoin code fork whose “Duplex Work” design aims to let AI service providers submit proofs of inference computations while earning block rewards. Bankless analyst David Christopher describes the FP8 upgrade as a way to reduce a technical barrier: Pearl previously supported INT7, requiring providers using other formats to convert their models. The analysis says more than 80% of model-inference traffic on platforms such as OpenRouter uses FP8. That figure is presented as part of the analysis, not independent evidence that those providers will use Pearl. FP8 may also require newer GPUs than INT7.

There is a more basic gap between the pitch and the current rules: according to the analysis, Pearl does not require miners to perform genuine AI inference to earn rewards. The FP8 change does not, by itself, prevent meaningless computation from qualifying as mining work. Christopher points to provider participation, mining-power concentration and the share of blocks tied to real user requests as tests of the model; the last would be difficult to observe externally.

Crypto researcher hitesh.eth offers a harsher assessment. He argues that the investment narrative depends on Bitcoin miners moving into AI computation for token incentives, and that inflation-funded rewards cannot sustain that shift. He also alleges coordinated influencer promotion and doubts that a higher token price will translate into product adoption. Those are his criticisms, not established findings. For now, a testnet upgrade and a token rally do not demonstrate demand for Pearl’s proposed service.

Further reading