Compound builds financial models from built-in market data: SEC filings, standardized and as-reported financial statements, earnings call transcripts and daily prices, with nothing to upload or license. Compound’s own spreadsheet engine implements 477 Excel functions, and on 148.9 million formula cells from public workbooks, 98.5% of its results are bit-for-bit identical to Excel’s. The worked examples below model one deal, the $8.4 billion take-private of Clearwater Analytics, from its filings and merger proxy.
Three-statement model from the 10-Ks
Compound links the income statement, balance sheet and cash flow statement so each forecast year’s cash and debt follow from the operating case. The three-statement model example takes Clearwater Analytics from five years of 10-Ks to a forecast to 2030, and shows that FY2025 revenue growth of 61.9% was 18.8% organic once its two acquisitions are taken out.
DCF model from the merger proxy
Compound’s AI DCF model takes management’s projections from the merger proxy, finds the depreciation, capex and working capital lines in a banker’s deck, and ties free cash flow to the proxy’s own figure within $1.5 million in every year. On management’s plan it values Clearwater at $19.78 a share with stock pay as a cash cost and $24.20 with it added back, against the $24.55 offer, and a stress case gives $14.68.
LBO model: sources and uses to returns
The LBO model example runs the take-private on management’s plan and on an outside-in case that Compound built only from what Clearwater had reported. The outside-in case returns a 21.1% IRR and 2.4x MOIC on a 2030 exit at the 25x forward multiple paid, against 32.4% on management’s plan, and the gap between them is revenue growth.
Debt schedule and sensitivity tables
Compound’s debt schedule example rolls the $2.7 billion term loan forward a year at a time from 9.6x leverage at close, with paid-in-kind interest, the cash sweep and the springing covenant. The two-variable sensitivity tables rank eleven IRR drivers and build four grids as live Excel data tables, which show the IRR stays above 20% only while revenue growth averages at least 11.5% a year.
Why buy-side teams build models in Compound
- Every input traces to its source. Click a cited cell to open the filing page, transcript passage or workbook cells it came from; the downloaded .xlsx keeps each citation as a cell note.
- Your firm’s templates. Compound works in your firm’s model and deck templates, keeps their layouts and formulas, and fills them for the new deal.
- A comment becomes the change. A VP comments on a cell or a slide in the deal’s shared agent, and the agent makes the edit and saves it as a new version.
- Many models at once. Start a task per company or per document, and subagents run them in parallel on Compound’s servers while the team is away.
