RWT vs SAFE

Redwood Trust, Inc. vs Safehold Inc. New — Valuation Comparison 2026

RWT

Real Estate Investment Trusts
Redwood Trust, Inc.
Quality
5.3
out of 10
Value Trap
22
SAFE
Price
$5.42
Last close
Models
10/13
Active
VS

SAFE

Real Estate Investment Trusts
Safehold Inc. New
Quality
7.3
out of 10
Value Trap
12
SAFE
Price
$14.97
Last close
Models
10/13
Active

Model-by-Model Comparison

ModelType RWT Fair ValueRWT Upside SAFE Fair ValueSAFE Upside
Bayesian DCF Intrinsic $1.11 -79.5%
Earnings Power Value Intrinsic $3.38 -41.0%
EROIC Spread Intrinsic $•••.•• ••.•% $•••.•• ••.•%
First Chicago Scenario $•••.•• ••.•% $•••.•• ••.•%
Markov DDM Intrinsic $13.67 +152.3% $13.21 -11.8%
ML-RIV Intrinsic $12.30 +127.0% $83.55 +458.1%
PWERM Option-Based $•••.•• ••.•% $•••.•• ••.•%
Regime Cross-Sectional Relative $•••.•• ••.•% $•••.•• ••.•%
Sentiment SOTP Hybrid $•••.•• ••.•% $•••.•• ••.•%
CUCE Ensemble Ensemble $•••.•• ••.•% $•••.•• ••.•%
FTNN Topology Relative $•••.•• ••.•% $•••.•• ••.•%
RCMH-DCF Intrinsic $•••.•• ••.•% $•••.•• ••.•%
🔒

Unlock Full 13-Model Comparison

Access all valuation models for RWT vs SAFE — including EROIC Spread, First Chicago, Markov DDM, PWERM, and 7 more.

Access Full Analysis — From $27/mo →

RWT vs SAFE — Which Stock Is More Undervalued?

SAFE scores higher with a 7.3/10 quality rating vs RWT's 5.3/10. Both stocks are analyzed daily using SEC EDGAR filings across 13 independent models.

Comparing Redwood Trust, Inc. (RWT) and Safehold Inc. New (SAFE) across 13 institutional-grade valuation models reveals how each company's intrinsic value stacks up against its market price. CirclFi's engine processes SEC EDGAR 10-K and 10-Q filings, FRED macroeconomic data, and GDELT news sentiment to generate independent fair value estimates daily.

RWT currently trades at $5.42 with a QOC of 5.3/10, while SAFE trades at $14.97 with a QOC of 7.3/10.

Both companies are analyzed with models spanning intrinsic (Bayesian DCF, EPV), scenario-based (First Chicago), regime-switching (Markov DDM, RCMH-DCF), machine learning (ML-RIV, FTNN Topology), and ensemble methods (CUCE).