Grayscale Expands Beyond AI Tokens to Computing Data and Credit
Grayscale's October 23 research note looks at a more specific crypto-and-AI proposition than a general bet on smarter software: who supplies training data, who finances hardware and who receives payment when computing capacity is used.
Released on October 23, a research note from Grayscale examines a more targeted crypto-and-AI investment thesis than a broad wager on intelligent software: examining the suppliers of training data, the financiers of hardware, and the beneficiaries of computing capacity payouts.
TL;DR
- The report highlights DATA, USD.AI and Pearl as examples of different infrastructure roles.
- Hardware financing introduces credit questions alongside blockchain mechanics.
- A useful service and a valuable token remain separate propositions.
Three business models sit behind one broad AI label
Within the analysis, research head Zach Pandl outlines DATA and its focus on contributed training data, USD.AI and its GPU-backed lending, and Pearl’s efforts to merge AI computing with blockchain rewards. These represent Grayscale’s specific examples and forecasts rather than independently verified outcomes across the entire industry.
Every model serves a distinct customer base. Data networks require parties willing to license datasets, financing protocols need borrowers capable of servicing debt, and computing networks require practical workloads paired with dependable delivery measurement. Grouping them all beneath a single AI-token category can obscure critical details.
The credit component is stated clearly in USD.AI’s documentation, which outlines loans backed by GPUs and related cash flows, where recourse is typically restricted to the collateral alongside defined exceptions. Its yield-generating asset derives from loan repayments and income generated by reserves.
Such a structure introduces familiar underwriting inquiries via a novel distribution channel. What occurs if equipment depreciates more rapidly than anticipated? How concentrated is the customer base paying for usage? Who possesses the authority to recover and liquidate collateral if a borrower defaults? Although a blockchain ledger can enhance transparency regarding a claim, it does not automatically resolve these questions.
This division between operating and financing activities is likewise evident in Bitdeer’s disclosures regarding mining and AI operations. While hardware, funded capacity, and actualized customer revenue are interconnected metrics, they cannot be used interchangeably.
The token must connect to the underlying activity
The presence of successful infrastructure does not guarantee an equivalent benefit for every linked token. Governance tokens, claims on lending revenue, and utility tokens used for service payments can carry vastly differing rights.
For instance, the USD.AI depositor guide distinguishes between the non-yielding USDai asset, the income-generating sUSDai asset, and CHIP governance tokens. Merely citing the protocol does not clarify the exact exposure held by a token owner.
A parallel standard governs data operations. While compensating contributors can stimulate valuable supply, long-term demand relies on whether purchasers view the data as lawful, pertinent, and scarce. Token rewards can jumpstart initial participation without guaranteeing that customers will maintain payments after those incentives decline.
Similarly, the expansion of tokenized assets must be evaluated through the specific rights granted by the token. Furthermore, consumer-facing AI applications ultimately depend on retaining users drawn to genuine product utility.
Grayscale’s publication outlines several distinct business models for evaluation. Ultimately, broader adoption, sound unit economics, and legally enforceable claims will offer the most reliable proof that these models are evolving into sustainable enterprises.
Short FAQs
- What is the focus of Grayscale’s October 23 research note?
- The note explores the specific infrastructure behind AI and crypto, looking at who supplies training data, finances hardware, and gets paid for computing capacity.
- Which projects are highlighted in the report?
- The report highlights DATA, USD.AI, and Pearl as examples of different AI and crypto infrastructure roles.
- How does USD.AI handle credit and loans?
- USD.AI features loans secured by GPUs and related cash flows, with recourse typically limited to the collateral and specified exceptions.
- Do successful infrastructure projects guarantee token value?
- No. A successful service and a valuable token remain separate propositions, as governance tokens, lending income claims, and payment tokens carry very different rights.