# $SOFI vs $NU: why Nu gets 5.3x book and SoFi gets 1.7x (valuation-only) — retained record

> Retained reddit record from reddit:ValueInvesting.

Canonical: https://highsignal.app/data/records/fd9177536a0327de

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ValueInvesting)  published 2026-10-21 00:25 ingested 2026-10-22 02:43

[original source ↗](https://www.reddit.com/r/ValueInvesting/comments/1xb8r7q/sofi_vs_nu_why_nu_gets_53x_book_and_sofi_gets_17x/)

## retained text



```
Scope note first. This is a valuation-only comp: book value, ROE, and what the multiples imply. It's not the full case for either company. I'm bullish on both. I own $SOFI and I don't own $NU.

\*\*Where they sit (Yahoo Finance, Oct 19 close)\*\*

\- $SOFI $14.30, down 45.4% YTD. $NU $14.51, down 13.3%.  
\- Price to book: 1.67x vs 5.29x  
\- ROE, trailing 12 months: 7.1% vs 31.6%  
\- Forward P/E: 17.3x vs 12.9x  
\- Deposits (Q2): $45.5B vs $45.3B. Basically the same size.

Nu's book multiple is about 3x SoFi's. Its ROE is about 4.5x. So most of the valuation gap is a returns gap. Honestly the market's pricing makes sense to me. It's paying up for proven returns.

\*\*The bear case on SoFi, stated fairly\*\*

On returns alone, $SOFI is the pricier stock. You pay about 0.24x book per point of ROE. Nu is about 0.17x. Nu is cheaper on forward earnings too. And Q2 by itself annualizes to about 5.9% ROE at SoFi on my math, so the inflection hasn't shown up yet.

\*\*What ROE is the price assuming?\*\*

Justified P/B = (ROE - g) / (COE - g). My assumptions: 11% cost of equity and 5% growth for both. Treat it as an illustration. Change the assumptions and the answers move a lot.

\- $SOFI at 1.67x implies about a 15% sustainable ROE. It earns 7.1% today.  
\- $SOFI at 2.5x book would need about 20%.  
\- $NU at 5.29x implies about 37%. It earns 31.6%.

So the market is already giving SoFi credit for roughly doubling its ROE. That cuts against my bull case and I'll admit it. Management's long-term ROTCE target is 20% to 30%, so 2.5x book is the bottom of their own range.

\*\*Why I think SoFi's ROE can move\*\*

Deposits were over 90% of average liabilities in Q2, so I think most of the funding remix is done. The next leg has to be operating leverage. Adjusted EBITDA margin was about 31% in Q1 and 30% in Q2. The full-year guide (about $1.6B on $4.75B to $4.85B of adjusted net revenue) needs roughly 35% to 37% in H2 on my math.

The risk is rates. The Fed hiked Sep 16, and SoFi's own 10-Q says +100bp costs about $159M of NII because liabilities reprice faster than assets.

\*\*Nu's returns come with Brazil attached\*\*

\- 90+ day NPLs went from 6.5% to 6.9% over the past year. Nu says part of that is intentional.  
\- A 22.9% NIM sits on a 13.75% Selic, more than 3x the US base rate.  
\- It earns in reais and reports in dollars.  
\- Brazil's runoff is Oct 25, and the US launch is only about six weeks old.

\*\*What I want on Oct 27\*\*

SoFi's ROTCE was about 6.7% annualized in Q2 on my math. Net income around $190M would put Q3 near 8%. That's the start of the inflection I'm looking for. Tangible book was $7.34/share at Q2, so at $14.30 you're paying about 1.95x tangible book.

If SoFi prints \~8% ROTCE next week, does the market pay 2x book for it, or does it want to see 15% first?

Longer version with all the charts and sources is on my Substack if you want it: https://nightanddaycpa.substack.com/p/sofi-vs-nu-the-roe-gap

NFA. I'm long $SOFI, no position in $NU.
```



document fetched 2026-10-22 00:47 · canonical https://reddit.com/r/ValueInvesting/comments/1xb8r7q/sofi_vs_nu_why_nu_gets_53x_book_and_sofi_gets_17x
