Demand for AI training data and high-performance computing hardware is rising as the industry expands. In a research report published on October 23, Grayscale head of research Zach Pandl identified DATA Network (DATA), USD.AI (CHIP) and Pearl (PEARL) as projects targeting different parts of the AI infrastructure market. Their recent price performance, however, has diverged, underscoring the gap between project fundamentals and market valuations.
DATA Network Builds a Licensing Model for AI Training Data
AI models require large volumes of images, video and voice recordings for training. DATA Network aims to collect this material with the necessary permissions, license it to AI developers and reward the people who contribute the data. The model is designed to connect data collection and licensing with token-based incentives.
On October 12, the Sui Foundation announced a $30 million investment in the network. DATA also plans to migrate to the Sui blockchain, with existing tokens transferred on a one-to-one basis. The move gives the project additional funding and new on-chain infrastructure, but it has not yet produced a clear improvement in the token's price.
At the time of publication, DATA was trading at $0.15243, down 42.07% over 90 days and 34.65% over the past month. The performance suggests that investors are still assessing whether the funding, chain migration and underlying data business can provide lasting support for the token's value.
USD.AI Links Blockchain Lending to GPU Purchases
AI training and inference require large quantities of graphics processing units (GPUs). Because the hardware is expensive and becomes outdated quickly, computing providers often need substantial financing. USD.AI is designed to let computing companies raise funds through blockchain-based lending pools backed by the relevant hardware and future revenue.
In August, Bullish announced a $100 million financing facility to support USD.AI's GPU lending operations. The arrangement connects digital-asset financing tools with the capital requirements of AI computing infrastructure and has also placed the CHIP token under market scrutiny.
CHIP was trading at $0.04386, up 47.98% over 90 days. It remained about 69% below its April high, however. The recent gain has therefore not fully reversed the earlier decline. Future price volatility is likely to remain sensitive to how the market assesses the scale of GPU-backed lending, asset quality and the project's ability to generate revenue.
Pearl Ties Its Token to Computing Revenue
The third project highlighted by Grayscale is Pearl (PEARL). Unlike DATA, which focuses on training data, and USD.AI, which focuses on GPU financing, Pearl is positioned around earning revenue from computing capacity. Its model seeks to capture on-chain economic activity generated by demand for AI computing power.
The information disclosed so far places the three tokens in separate parts of the infrastructure chain: data supply, hardware financing and computing revenue. All are linked to AI infrastructure, but their business models differ, and their token prices have already moved in different directions. For market participants, the key test is whether each project can convert AI demand into verifiable data-licensing activity, lending business or computing revenue. The AI theme alone does not explain the differences in their market performance.