Date: 2026-03-31 Quarter: Q4_FY25 (Dec-2025) Price at analysis: ~17.80(Mar25, 2026close)Marketcap: 24B | Shares: 1,346M diluted TTM adj net revenue: $3.591B | Rule of 40: 68
analyses/SOFI/SOFI_earnings-review_Q4_FY25.md — solid
six-factor scoring, I agree with the 4/5 convictionRevenue:
Segment Revenue (Q4 FY25):
Fee-Based Revenue:
Profitability:
Members & Products:
Balance Sheet:
Credit Quality:
Guidance — FY26:
Here's the thing...
SoFi is being valued as a bank. It's building a platform.
Let me break down the platform architecture:
Platform verdict: SoFi has built a four-layer platform stack. Consumer acquisition -> marketplace lending -> B2B fintech infrastructure -> crypto rails. Each layer feeds the others. The bank charter is the foundation that competitors cannot easily replicate. Nubank got conditional OCC approval (January 2026) and Revolut applied (March 2026) — the moat is narrowing, but SoFi has a multi-year head start in deposits, members, and platform breadth.
"2025 was a tremendous year on all fronts. Our member focus drove an unprecedented level of innovation across our business that led to the strongest financial performance in the history of our company." — CEO Anthony Noto
"Cross-buy continues at an exceptional pace with 40% of new products opened by existing SoFi members. Over the past year, our cross-buy rate has increased by 7 percentage points." — Noto (this is the platform flywheel in action)
"SoFi USD will be a game changer for our business as it enables us to be an infrastructure provider for banks, fintechs and enterprise platforms, positioning us at the center of the crypto ecosystem." — Noto
"The data continues to support our 7% to 8% net cumulative loss assumption for personal loans... The gap between the newer cohort curve and the 2017 cohort curve widened by 8 basis points during the fourth quarter. In fact, this gap has widened in each of the past 6 quarters since we began measurement." — CFO Chris Lapointe (credit bears, read this carefully)
"We expect to deliver compounded annual adjusted net revenue growth of at least 30% from 2025 to 2028. Additionally, we expect to deliver compounded annual adjusted earnings per share growth of 38% to 42% from 2025 to 2028." — Lapointe (this is a management team leaning in, not hedging)
"I really want us to get that escape velocity where the amount that we spend becomes more and more efficient." — Noto on brand/marketing leverage
1. The QoQ deceleration is noise. The YoY reacceleration is signal.
QoQ went from +12.5% (Q3) to +6.6% (Q4). Surface-level, this looks like deceleration. But look at the YoY trend: +19.7% -> +42.8% -> +37.9% -> +39.6% ^^. Q4 actually reaccelerated on a YoY basis. The QoQ slowdown is seasonal — Q4 is always sequentially slower for fintech (year-end, holiday period). The signal is in the YoY trajectory, and it's pointing up.
2. Financial Services is one quarter from becoming the largest segment.
FS: $456.7M. Lending (adj): $486.5M. Gap: $30M. FS is growing at 78% YoY vs Lending at 15%. FS will overtake Lending in Q1 or Q2 FY26. When that crossover happens, the narrative shifts permanently from "lending company" to "financial platform." That is a re-rating event. The market is slow to price mix shifts, and this one is happening in real-time.
3. LPB is a marketplace business hiding inside a bank.
$193.7M quarterly revenue, annualizing at $775M, growing 2.9x YoY. Zero credit risk. Pure fee revenue. The $3.6B in new committed flow from three independent institutional sectors is institutional validation of the platform model. This sub-segment alone, at its current trajectory, could be worth $5-8B as a standalone marketplace business. Instead it is buried inside a "bank" trading at 2.6x P/TBV.
4. Tech Platform accounts are down 23% but the REAL metrics are up.
Large client departed. Accounts dropped from 167.7M to 128.5M. But: revenue UP 19%, contribution margin UP from 31.2% to 39.2%, contribution profit UP 49.2% YoY. This is a textbook "addition by subtraction" — the departed client was likely low-margin, high-volume. The remaining portfolio is higher-quality and higher-margin. Galileo targeting $1B run rate tells you management sees this as a growth segment, not a shrinking one.
5. SoFi Plus paywall — real risk, but manageable with nuance.
This goes live tomorrow (March 31, 2026). The consumer math: 10/month(120/year) for a 0.25% APY bump that generates ~$50/year on $20K deposits. Net negative for light users. Doctor of Credit has already done the math publicly.
BUT — and this is the nuance the surfaced learnings correctly identified — the 40% cross-buy cohort (multi-product users) gets substantially more value: 2% IRA match, 5% grocery cashback via Smart Card, loan rate discounts, access to alternative investments. For members using 2+ products, the value proposition is net positive.
The churn risk is concentrated in the single-product deposit-only tail — members who only have a savings account for the APY. These are also the lowest-monetization members. If they leave, products-per-member and revenue-per-member may actually improve even as raw member count dips.
What to watch in Q1-Q2 FY26: Products per member (stable or rising = healthy churn). Raw member net adds (soft but not collapsing = manageable). Revenue per member (rising = the right people are staying).
6. Incremental margins reveal the operating leverage story.
EBITDA grew 60% on 37% revenue growth. Incremental EBITDA margin: 44%. Incremental gross margin: 99.7% (!). This company is past the fixed cost investment hump. Every incremental dollar of revenue is nearly pure gross profit, and almost half drops to EBITDA. At 30% revenue growth guided for FY26, EBITDA margin should push to 34% (which is exactly what management guided). The profitability ramp is structural, not one-time.
7. The dilution is real but the context matters.
Shares: 1,346M (+17% YoY). $3.2B in total equity raises in 2025. This is the single biggest criticism the market has. But: TBV/share went from $4.47 to $7.01 (+57%). Warehouse lines went to $0. Estimated $110M in annualized funding cost savings. Management says the raise "fully offsets dilution impact" on EPS. And the FY26 EPS guide of $0.60 is 54% above FY25's $0.39, incorporating the diluted share count. If they hit $0.60, the dilution complaint evaporates. CEO Noto put $1M of personal money in at $17.50-18.20 on March 2. That is conviction.
Revenue growth: ~37-40% YoY. Tier 2 (approaching high end).
Per my framework: Tier 2 = 40-50% growth, 5-12% position. SOFI is at 37-40% on adj revenue, right at the boundary. But here is why I would put it at the high end of Tier 2:
Green flags: Accelerating FS revenue ^^, platform optionality (4 layers), self-serve GTM (consumer direct), profitability swing (first $1B EBITDA year) !!, multi-product adoption (40% cross-buy, rising ^^), consumption/usage-based growth (deposits drive NII), cloud-native architecture (Galileo/Technisys), management confidence (medium-term guide, insider buying).
Red flags: Dilution (+17% shares YoY), credit cycle exposure (PL charge-off 2.80%), Tech Platform client loss, SoFi Plus churn risk (near-term), charter moat narrowing (Nubank, Revolut), Blind engineering culture concerns.
First muji analysis of SOFI. Establishing initial thesis.
| Dimension | Assessment |
|---|---|
| Platform architecture | Strong. Four-layer stack with genuine "building block" positioning. Galileo/Technisys powers competitors. LPB is marketplace infrastructure. SoFiUSD creates new crypto rails. This is not a typical bank. |
| GTM model | Self-serve consumer acquisition. No SI dependency. 35% member growth at 13.7M scale is impressive. Brand awareness only 9.6% = long runway. |
| Product cadence | Exceptional. Smart Card launched in 4.5 months. Crypto trading, stablecoin, SoFi Pay, options trading, alternative investments — all launched in 2025. The velocity here is genuinely impressive for a regulated bank. |
| Revenue quality | Improving rapidly. Fee-based revenue now 44% of total, growing 53% YoY. NII still dominant but share declining. The mix shift from balance-sheet-intensive to capital-light is exactly the right trajectory. |
| Management quality | Strong execution, needs watching on culture. Noto is credible — Goldman/Twitter pedigree, consistent delivery, personal conviction ($1M insider buy). But Blind reviews flag weak middle management and code debt. At scale, this matters for a bank. |
| Competitive moat | Real but narrowing. Charter advantage is genuine vs Chime/Cash App. But Nubank (conditional OCC) and Revolut (applied) are coming. SoFi needs to widen the moat through platform breadth and crypto infrastructure before charter-only advantage erodes. 12-18 month window. |
From the latest scuttlebutt stage (March 30, 2026):
Customer: Bifurcated. App Store 4.8/5 (routine users love it), BBB 1.29/5 (complaint-seekers). Loan origination is praised; account freezes and fraud resolution remain chronic pain points. $1.1M regulatory fine (2024) for automated fraud detection gaps. SoFi Plus paywall generating organized community pushback on Doctor of Credit, Reddit.
Employee: Glassdoor 4.2/5 (broad), Blind 3.3/5 (tech). Gap is structural and unchanged. Compensation adequate (3.5 on Blind), but tech leadership below Noto described as weak. "Move fast and break things" culture at a nationally chartered bank — this is a genuine risk.
Competitors: Charter moat narrowing. Nubank (conditional OCC, Jan 2026), Revolut (applied, Mar 2026). SoFi leads on breadth (212 app functions) but feature count does not equal market share dominance.
Management: Noto bought $1M in shares at $17.50-18.20 (March 2, 2026). Total holdings 11.6M+ shares. In public interviews, stated he believes stock is undervalued. No softening of ambition: 50M members / 150M products by 2030.
Hiring: Two layoff waves in Q1 2026 (collections automation + mortgage ops), offset by 175-195 open tech roles. Net: restructuring toward automation, not distress. Offshoring signal is new and worth monitoring given customer service friction.
This is a platform company being valued as a bank, and the market has not figured it out yet.
The numbers are exceptional: first $1B quarter, first $1B EBITDA year, Rule of 40 at 68, margins expanding on every line, member flywheel intact at 13.7M with 40% cross-buy, credit quality better than historical vintages. But the real story is the platform evolution:
LPB is a marketplace business — $775M annualized, 2.9x YoY, $3.6B in new committed institutional flow, zero credit risk. This alone could be worth more than the current stock gives credit for.
FS is about to become the largest segment — growing 78% vs Lending at 15%. When the crossover happens (Q1-Q2 FY26), the "bank" label starts to break.
Galileo/Technisys is embedded infrastructure — powers competitors, expanding internationally, contribution margin exploding to 39%. "Building blocks," not "process tools."
SoFiUSD is regulated crypto infrastructure — first-mover from a national bank. Optionality play with asymmetric upside.
The risks are real: SoFi Plus churn (watching Q1-Q2 member metrics), charter moat narrowing (12-18 month window), credit cycle (managed well so far but macro-sensitive), dilution (17% but accretive to book value), and engineering culture (Blind signals bear watching). None of these are thesis-breaking today.
At ~$18, this is ~30x FY26 P/E on 0.60guide.Ifmanagementhitsthe38 − 421.20 EPS in FY28, which means ~15x on today's price. For a company with Rule of 40 at 68, four platform layers, and genuine scale, that is compelling.
Tier: 2 (high end, approaching Tier 1 on platform optionality) Thesis: Initiating. Strengthening. Action: Compelling entry point. Would start a 5-8% position.
long SOFI (initiating coverage)
-muji