Three new reviews. One diagnostic skill they all teach.
This week I shipped three forensic single-product reviews — and the most useful thing in any of them isn't a yield number. It's a diagnostic skill that generalizes.
The three reviews:
- Fundrise Flagship Real Estate Fund — $1.22B AUM, 1.33% 2025 return, $100M of new SOFR+525bps leverage drawn in February 2026 — disclosed only in the Subsequent Events footnote of the audited Form N-CSR.
- EquityMultiple Alpine Notes — marketed at "$235M+ raised from 1,850+ investors," but the most recent SEC Form D/A filed September 29, 2025 reports $23,034,380 sold to 341 accredited investors. A ~90% gap.
- Groundfloor Notes — 8.25% Signature Notes, first-priority security interest in the issuer's loan pool, perfect repayment record since 2018 — and both the parent AND the Notes issuer received going-concern qualifications in their FY2024 1-K filings.
The diagnostic skill is this: before allocating capital to any SEC-registered private investment, read three specific filings, in this order.
-
The Subsequent Events footnote (typically the last numbered note in any 10-K, N-CSR, 1-K, or 1-A POS). This is where the material things that happened after the fiscal year end but before the filing date appear. The Fundrise Flagship $100M leverage was here. It would not show up in the headline financials for another full year.
-
The Form D / Form D/A on SEC EDGAR for the specific issuer. This shows the actual capital raised in the current offering — not the marketing aggregate. The EquityMultiple Alpine Notes $23M vs $235M gap was here.
-
The auditor's opinion paragraph at the top of the financial statements. Look for "explanatory paragraph" or "going concern" language. If those words appear, the auditor is telling you something the marketing won't.
Three filings. Fifteen minutes total. Two of this week's three reviews — and dozens of others I've published — found their entire forensic angle in one of these three places.
That's the actual takeaway from this batch. Not the yields. The diagnostic.
If you want the worked examples, the three reviews above show what the data looks like when you actually run the diagnostic on real platforms.
— Jorge