How Capital Markets Finance Protects Balance Sheet Returns

How Capital Markets Finance Protects Balance Sheet Returns | Databricks Blog Skip to main content

Summary

Capital-markets finance teams need a continuous view of how funding, liquidity, capital consumption, and profitability move together as market conditions change.

Genie One brings governed business context to finance questions, helping teams investigate the desks, positions, definitions, and source data behind a changing metric.

With trusted analysis and coordinated follow-up work, finance can help the business respond faster while accountable leaders retain decision authority.

An insurer settles a claim months after the loss. An asset manager marks someone else's fund at day's end. The sell side doesn't get that grace period. A capital markets CFO can't wait for quarter-end to see where return on the balance sheet landed that — that balance sheet carries market, funding, and liquidity risk continuously, priced position by position, all day. Before the desk has time to react, the ground can shift under it: the financing cost on a position that was priced correctly last week, the capital consumed by that position as exposures move, or a funding gap opening faster than anticipated. Each number comes out of a different system — desk P&L, treasury funding curves, risk engines weighing exposure against netting sets and margin — and each is increasingly shaped by AI-informed pricing, funding, and risk workflows that reach across all of them. Why the balance sheet became finance's front line The sell side runs on its balance sheet. As markets move and spreads shift, funding costs, capital consumption, and liquidity demands move with them, each shaping how efficiently the firm can deploy scarce balance sheet and generate returns. Helping the business understand those tradeoffs and protect profitability has long been finance's job. That job has gotten harder. Volatility is up, trading books turn over faster, and Basel III Endgame together with the Fundamental Review of the Trading Book (FRTB) raise the capital required to support market risk. Industry analyses have warned that the FRTB standardized approach can materially affect market-risk capital requirements, depending on portfolio composition, jurisdiction, and modeling approach. Finance sits at the center of it, translating an increasingly complex set of funding, capital, and liquidity decisions for trading, treasury, and leadership. Now add AI agents embedded in pricing workflows, funding recommendations, reporting, surveillance, and other processes across the trading book. It isn't only that AI is informing more decisions, its that those decisions now touch funding, capital, liquidity, and valuation at the same time, and the pace of that shift will keep accelerating.

Markets, spreads, and funding move all day. Finance protects balance sheet returns as they do, but doing so increasingly requires a shared language for what those numbers actually mean. Finance is good at finding the number, even buried in complexity. But a good finance team is also the first to say the number alone doesn't tell the story.  What matters is what's behind it: the definitions, the desks and products, the funding and spread drivers, and how each is shifting as the business moves. Why finance needs business context, not just data That's the job of an ontology: capturing meaning and keeping it current as the business changes. On the sell side, meaning is inseparable from the firm’s own balance sheet — a position isn’t just a number, it's capital consumed, funding drawn, and risk the desk is carrying until it's hedged or unwound. For capital markets, firms can anchor that context in established standards such as FIBO, the Financial Industry Business Ontology, which defines how desks, products, and instruments relate to one another in terms a risk or finance system can reason over. Paired with the firm’s governed data and business definitions, that context lets finance read a number the same way across front office, treasury, risk, and regulatory reporting.

Accurate is the right figure. Correct is the same figure, rooted in the desk, the market, and the funding. An ontology holds up if it's kept current, though. Markets don’t wait for quarterly metadata updates, so context has to move as fast as the business itself. This is where Genie comes in — Databricks’ suite of data-smart AI coworkers, grounded in enterprise context. Genie   One helps finance leaders ask direct questions, investigate trusted business data, and carry that understanding into the next step of their work, pulling together the relevant definitions, desks, positions, source data, and permissions that matter. The goal isn't a readout of what happened; its an informed discussion of what to do next. People stay accountable for pricing, funding, risk, and capital decisions. That's not hypothetical. One top global bank, was untangling this fragmentation problem across multiple data sources and reporting standards; it cut regulatory data processing time for liquidity reporting from 10 hours to 8 minutes, letting Treasury respond faster as market conditions changed. Three questions sit on the mind of every capital markets finance team, each feeding into the next. Here's how Genie changes the work on each one. Where will funding and liquidity tighten before the desk sees it?

When markets move or inventory builds, funding needs, intraday liquidity, and leverage ratio headroom can shift quickly. Rather than waiting for a periodic report to surface the problem, finance can use Genie One to trace intraday funding usage and collateral movements back to specific positions, desks, and trades driving the change while there's still time to respond. Where are returns quietly eroding?

This one is harder to catch because it often happens a few basis points at a time: funding costs creep up, spreads compress, or the capital required against a position increases. Genie One helps finance follow the erosion to its source — the desks, products, positions, funding lines, and capital drivers behind it — instead of noticing it only after it shows up in a monthly readout . Can we explain every capital and valuation number on demand?

A number finance can't stand behind is a number it can't act on, and the moment that gets tested is usually the moment a regulator, auditor, or board member asks. Here Genie One does...