Grove Data: Operating in Public

Grove Data began as a two-sided tool. Externally, a public dashboard that allowed parties insight into our allocations. Internally, tooling to monitor the various onchain and offchain assets we were deploying into, and a system to measure our borrowed capital from Sky. As we thought about developing this platform further, building in public felt natural. Our controller, allocations, and governance were already public, so why not build our protocol data in public too?

Since our first deployment in July 2025, the Grove Allocator has put billions to work across several asset types. That base has expanded steadily since our first Public Credit allocations. These assets differ in their source of underlying yield, in how they accrue yield, and in whether most data and context sit offchain or onchain. Today, our Grove spread-earning TVL sits at around $1.6 billion.

The expansion of asset types has encouraged us to develop an internal taxonomy: a consistent method of measuring performance and evaluating risk in real time across assets that have a variety of structures and differing NAV frequencies. This taxonomy underlies the operations of our credit strategy and helps us navigate daily flows.

Reporting: Reconciling debt, accruals, and risk capital

Grove Data now publishes the Financials on a monthly basis. Our function as a Sky Prime Agent involves a monthly settlement process that determines our revenue, the borrow expense to Sky, and subsequently our net income. The income statements on Grove Data break this down so anyone can tie these results to onchain activity or the Sky Forums.

Internally, we have built out tooling to measure our activity on both a cash basis and an accrual basis. Our experience in both traditional finance and crypto back offices encouraged us to build the best of both worlds. On the onchain side, a result-based ecosystem where we can measure debt, income, and index live allocations; on the offchain side, a classic accrual system to measure our more traditional holdings such as the Galaxy warehouse facility or our daily-NAV tokenized products such as JAAA and STAC.

Our Grove Data financial models now also power our latest Grove Quarterly Protocol Review.

Agentic Tooling: The role of agents in the Financial “Office”

We have found agentic tools particularly essential in this process and developed Grove Data with this new frontier in mind. Agentic infrastructure powers Grove Data, from ingesting transactions and validating onchain movements to experimenting with a variety of financial models. Agents also power our reconciliation process, applying our in-house expertise on credit structuring and accounting to review and track every dollar that we deploy.

A component of the agentic advances has been the ease with which we can experiment and shift gears. This has been facilitated especially by the fact that agents and models understand onchain data very well. Reviewing historical allocations, measuring returns over time, and evaluating risk are all enabled by traversing block history or indexed data. We believe the AI advances have taken the transparency of the blockchain to another level, delivering not only on the promise of low reconciliation cost, but also the ability to validate movements and accounting just as quickly. Not only can we utilize agentic tools in navigating our data, but agents now even review and propose reallocation workflows based on market movements. These workflows operate within the onchain rate limits of the Grove Allocator, following the guardrails set by the smart contracts in place.

What’s next: Operational excellence in public

We are just getting started in this new age of onchain and agentic infrastructure, building an asset allocator that is best in class across both traditional financial models and the onchain credit ecosystem. We expect to keep improving and updating the public version of Grove Data. Stay tuned for more historical statements that share our balance sheet, a public API with estimated daily accounting, and deeper insights into the asset classes we allocate to.

Follow along with our daily movements at grove_data on X