Google Cloud pilot automates syndicated loan lifecycle

Nammu21 has developed a proof of concept with Google Cloud to automate the syndicated loan lifecycle on Google Cloud Universal Ledger. The project modelled a USD $1.945 billion Term Loan B.

The work targets a part of corporate finance that still relies heavily on manual processes, document interpretation and coordination across multiple parties. In the syndicated loan and private credit markets, those frictions can stretch settlement into weeks and leave participants repeatedly reconciling records across agents, borrowers and lender groups.

According to the Loan Market Association, secondary market settlement times for par and near-par loans averaged 42 days in the fourth quarter of 2024, with a median of 27 days. Those delays can lock up capital and increase counterparty exposure while back-office teams work through approvals and match records.

Nammu21's approach is to convert loan agreements into what it calls a Digital Credit Instrument, or DCI. The model captures terms in the legal document, including conditions, consents, eligibility rules and operating mechanics, and links each element back to the source provision in the agreement.

The proof of concept then uses those structured terms to create smart contracts on Google Cloud Universal Ledger, a permissioned distributed ledger designed for institutional financial workflows. In the model, the contract acts as a digital representation of a specific loan facility and governs actions against its live state.

How it works

In the process outlined by the companies, an agent bank selects a credit agreement that has been transformed by Nammu21. The platform provides structured terms from the document, which are then used to create a gculpy contract on the ledger.

That contract establishes an on-ledger version of the loan facility. The next stage is participant onboarding, in which the borrower and lenders each review the request and approve it through their own interface before the structure is activated.

Once all parties are onboarded, the agent bank records an immutable hash of the loan's key terms on the ledger. This is intended to confirm that the on-ledger data matches the canonical off-ledger record.

Later events in the loan lifecycle, including secondary market transfers and repayments, are managed as multi-step workflows based on participant consent. A loan position transfer, for example, requires consent from the new lender, the borrower and the agent bank before it can be executed on the ledger.

Payment flows, from interest payments to loan closeouts, settle on the ledger using tokenised commercial bank money, removing the need for external payment rails within the proof-of-concept architecture. The ledger is also designed to restrict asset interactions to those explicitly authorised by the user, addressing a compliance issue that can arise on public blockchain networks when unwanted tokens are sent to an organisation's wallet.

Market focus

The demonstration centred on one segment of a broader multi-facility financing: a USD $1.945 billion Term Loan B. In that model, the agent bank, borrower, transferor and transferee all operated from a permissioned view of the same on-ledger record.

Google Cloud described the architecture as a way to create a single immutable source of truth for the parties involved. The reasoning behind the design is that many inefficiencies in loan markets stem not only from duplicated data entry, but also from different interpretations of legal terms by different participants.

Nammu21 serves two functions in that set-up. First, it renders the DCI directly from the legal agreement without manual document reading to establish terms. It then derives the gculpy contracts that define the loan obligation and its permitted actions on the ledger.

GCUL's contract language, gculpy, is a statically typed subset of Python intended to support readable and auditable financial logic. Google Cloud said the language removes some of the development constraints found in other ledger programming models, allowing engineers to focus on contract logic tied to financial agreements.

Broader implications

The companies are targeting a market where legal agreements are often highly tailored, with complex covenants, dynamic tranches and approval structures that are not easily standardised. That has limited the pace of digitisation compared with other areas of capital markets.

The proof of concept suggests that if governing terms can be structured from source documents and enforced through shared ledger logic, settlement times could be reduced and operational workloads lowered. It also points to a model in which counterparties rely on common execution rules rather than separate downstream checks of what each party believes the documents allow.

The same framework could also be applied beyond syndicated loans to other forms of credit in which bespoke legal terms still govern transfers, repayments and participant rights. In that sense, the project reflects a wider push by financial technology firms and cloud providers to move more post-trade processes into programmable, permissioned environments built for regulated institutions.

According to the project description, every on-ledger action in the model completes near-instantly and is anchored in data derived from the original credit agreement, replacing a reconciliation-heavy process with real-time verification.