Stock Analysis · Bulls vs Bears

SoFi Technologies SOFI

🐂 Bulls vs Bears 🐻Report #213 Oct 2026
★ Willem owns this stock · 1.1% of the portfolio
45
moderate risk▼ −4 vs last
🐂 9vs12 🐻
Bears lead
25%
▲ Chance of the biggest tailwind the forecasts miss
25%
▼ Chance of the biggest headwind the forecasts miss
The verdict

The evidence leans slightly to the bears for the next two quarters: Revenue beats every guide, but diluted EPS has been stuck at $0.11–0.13 for four quarters. The capital buffer is being spent at about 2 points a quarter. Rates and the labour market are both moving against an unsecured, liability-sensitive lender. Muddy Waters' specific claims are still unanswered point by point. The franchise is real: $45.5B of…

🎙️ Bulls vs Bears podcast — SoFi Technologies
Alex and Sam debate every bull and bear argument · 13:23
0:0013:23
⤓ Download

Written by Claude (Anthropic’s AI) from public sources · Willem owns SOFI · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 9

  1. 1Cheap, insured deposit funding that rivals still lack at scaleStrong
  2. 2Every revenue guide since 2025 has been beatenMedium
  3. 3Losses are falling and running below SoFi's own assumptionsMedium
  4. 4Members and products keep compounding, mostly through cross-sellMedium
  5. 5Tangible book per share is up 56% and the multiple has halvedMedium
  6. 6The federal PLUS cuts are already showing up in student volumeMedium
  7. 7Capital is still above target and management says no raiseMedium
  8. 8Loan-platform buyers are broadening beyond one or two fundsWeak
  9. 9The CEO bought about $2.25M of stock in 2026Weak

🐻 Bear arguments 12

  1. 1Profits rest on $2B of fair-value premium a short seller contestsHigh
  2. 2Per-share profit is flat and Q3 needs a 26% EBITDA jumpHigh
  3. 3Liability-sensitive balance sheet in a hiking cycle with 5% yieldsHigh
  4. 4Capital ratio falls 2 points a quarter toward the target floorMedium
  5. 5Default assumptions rise every quarter as jobs stallMedium
  6. 6Shares up 16% in a year, with more dilution queuedMedium
  7. 7The loan-platform fee engine has stalled and buyers face redemptionsMedium
  8. 8Valuation still rich for a 7% ROTCE bank, and shorts doubledMedium
  9. 9Debt nearly doubled and $1.3B of convertible cash is due by 2027Medium
  10. 10More fintechs are getting bank chartersLow
  11. 11The technology platform is shrinking after losing its anchor clientLow
  12. 12Insiders keep selling and the CEO stopped buying in JuneLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case

🐂 Bull #1Cheap, insured deposit funding that rivals still lack at scale

  • The point: SoFi funds its loans mostly with its own retail deposits. It does not rent funding from warehouse lenders or the bond market. That lowers its cost per dollar lent, and it lets SoFi choose when to sell loans instead of being forced to.
  • Evidence:
    • Deposits were $45.5B at 2026-06-30, up $5.3B in Q2 and up $8.0B in H1 [1,2].
    • About 98% of deposits are FDIC-insured; uninsured deposits are $1.1B [2].
    • NIM was 5.98% in Q2. Management expects it "above 5% for the foreseeable future" (Q4-25 and Q2-26 calls) [3,6].
    • Remaining liquidity capacity was $13.68B at 2026-06-30 (up from $12.22B at 2025-09-30), including $5.69B of undrawn warehouse lines [2].
    • Q2 core personal-loan funding cost was 3.1% against a 12.9% coupon [3].
    • Upstart got only preliminary conditional OCC approval on 2026-07-23 [25]. Nubank's US launch on 2026-09-10 runs through a partner bank, with its own charter expected in 2027 [26].
  • Why it matters: Cheap deposit funding is the one cost edge in prime lending. It is why SoFi could keep $7.6B of Q2 personal loans on its own balance sheet and earn the full spread [3].
  • 😈 Devil's advocate: These are rate-shopping deposits. One-year time deposits rose from $1.2B to $4.4B in six months [2]. In a hiking cycle SoFi must pay up to keep them, which is why +100bp now cuts NII by $159M a year [2] (Bear #3). The edge is over other fintechs, not over large banks with free checking balances.
  • Verdict: Strong The funding base is large, insured and growing, and the devil's advocate attacks its cost rather than its existence.

🐂 Bull #2Every revenue guide since 2025 has been beaten

  • The point: Management guides conservatively and beats on revenue every quarter. The FY2025 guide was raised twice and still beaten.
  • Evidence:
    • FY2025: adjusted revenue was guided at $3.235–3.31B early in the year, raised to $3.375B in July and to $3.54B in October. The actual was $3.591B. Adjusted EBITDA came in at $1,053.9M against a final $1.035B guide [5,7,8].
    • Q1-26 beat its guide on revenue ($1.087B vs ~$1.04B) and EBITDA ($339.9M vs ~$300M) [4,6].
    • Q2-26 adjusted revenue grew +40% against a ~30% guide [1,4].
    • The FY2026 revenue guide was raised on 2026-07-29 from ~$4.655B to $4.75–4.85B [1].
    • Members: the guide of "at least 30%" growth has been held; actual growth is +35% [1].
    • Consensus Q3 revenue is $1.26B (16 analysts) [37].
  • Why it matters: Reliable top-line beats mean revenue forecasts are not the risk. The debate is what reaches EPS.
  • 😈 Devil's advocate:
    • The beats are on revenue. The FY26 EBITDA ($1.6B), adjusted net income ($825M) and EPS (~$0.60) guides were not raised in either Q1 or Q2 [1,4].
    • The EPS guide now absorbs a 22% tax rate instead of mid-teens, a ~$0.05 hit [3].
    • Diluted EPS was $0.11, $0.13, $0.12 and $0.12 over the last four quarters [1,4,5,7].
  • Verdict: Medium The record is real, but it is a revenue record. The profit line has only been held.

🐂 Bull #3Losses are falling and running below SoFi's own assumptions

  • The point: SoFi's prime borrowers are performing better, not worse, and the model already assumes higher defaults than SoFi is seeing.
  • Evidence:
    • Personal-loan losses, adding back loans sold while delinquent, ran ~4.2% in Q3-25, ~4.4% in Q4-25, ~4.4% in Q1-26 and ~3.7% in Q2-26 [1,4,5,7].
    • The reported annualised rate was 2.62% in Q2, −21bp year on year [1].
    • 90-day personal-loan delinquency was 40bp [2].
    • The 2022-Q4 to 2025-Q3 vintages show 4.68% cumulative net loss with 35% of balances left, against 6.43% for the 2017 vintage and a 7–8% maximum-loss assumption [1].
    • The fair-value model assumes 4.77% annual defaults, above the ~3.7% realised [1].
    • Student loans: 0.61% charge-offs and 11bp 90-day delinquency [3].
    • Peer check: Happen's (formerly LendingClub) held-for-investment charge-off ratio fell to 3.2% from 3.8% [54].
  • Why it matters: For a fair-value lender, assumptions running ahead of reality are the buffer. If losses stay below 4.77%, marks hold and profit is not overstated.
  • 😈 Devil's advocate:
    • Part of the Q2 improvement is the denominator: average loans grew fast, which flatters a rate. Management said so itself [1].
    • Muddy Waters says the true all-in rate was ~6.1% in 2025, counting losses on loans moved to unconsolidated vehicles [9].
    • The data predates the September payroll stall (+29k) and July's revision to −10k [18].
  • Verdict: Medium SoFi's disclosed data is good and its model assumption is conservative relative to it. But it is backward-looking, and the measure itself is disputed.

🐂 Bull #4Members and products keep compounding, mostly through cross-sell

  • The point: SoFi adds about 1M members a quarter and sells them more products at near-zero acquisition cost, so revenue per member should rise faster than costs.
  • Evidence:
    • Members reached 15.81M (+35%), with a record 1.1M added in Q2 [1].
    • Products reached 24.38M (+42%). Twice as many products were added as members [3].
    • 51% of new products went to existing members, up from 43% in Q1 and 35% a year earlier [3].
    • SoFi Plus has 200k+ paid subscribers after one quarter. Management said it would be "disappointed" without 1M within a year [3].
    • Fee-based revenue was $472.3M, 39% of the total [1].
  • Why it matters: Cross-sell is the most SoFi-specific asset. Lending competitors cannot copy a 15.8M-member app quickly.
  • 😈 Devil's advocate: Contribution margin in Financial Services (46%) is below Lending (55%) [1]. Nothing in the filings shows profit per extra product. Many new products (SoFi Plus, Coach, crypto, SoFiUSD) are the "growth areas" management is reinvesting in instead of lifting the EPS guide [3].
  • Verdict: Medium The flywheel spins faster each quarter, but its profit is not yet visible per share.

🐂 Bull #5Tangible book per share is up 56% and the multiple has halved

  • The point: The stock fell ~42–43% in 2026 while the underlying book grew fast, so the price now asks much less of the business.
  • Evidence:
    • Tangible book value per share was $7.34 at 2026-06-30, up from $4.72 a year earlier and $7.01 at year-end [1,5].
    • At $15.23 (2026-10-12) [46], that is 2.07x tangible book (my arithmetic). Forward P/E is about 25x FY26 ($0.59–0.60) and 18.6x FY27 ($0.82) [37].
    • That compares with about 46x forward in January 2026 [13] and about 29x on 2026-03-24 [13].
    • Price-target initiations since August:
      • Piper Sandler: Overweight, $22, on 2026-08-17 [40].
      • Scotiabank: Outperform, $25, on 2026-09-02 [41].
      • Loop Capital: Hold, $22, on 2026-09-09 [39].
  • Why it matters: If the 38–42% EPS growth target for 2025–2028 holds even partly, a ~19x FY27 multiple leaves room [5,6].
  • 😈 Devil's advocate: Much of the TBV growth came from the $3.2B of 2025 equity raises, not earnings [2]. At an estimated ~6.7% annualised ROTCE (Q2 GAAP net income × 4 ÷ average TBV, estimate), 2.07x tangible book is still a premium multiple for a bank (Bear #8).
  • Verdict: Medium The de-rating is real and expectations are lower, but cheapness depends on EPS growth that has not yet appeared.

🐂 Bull #6The federal PLUS cuts are already showing up in student volume

  • The point: From 2026-07-01 new borrowers lost access to Grad PLUS, and Parent PLUS was capped. That pushes graduate and parent borrowing to private lenders, where SoFi has the brand and the funding [32].
  • Evidence:
    • SoFi student-loan originations were $2.69B in Q2 (+170%) [1,3].
    • Sallie Mae, the in-school leader, originated $716M in Q2. It reported peak-season indicators "at the higher end of our expectations or better" and estimates a $4.5–5B/yr private opportunity [22,23].
    • Estimates of the federal gap range from ~$8B/yr (NBER) to $15B (Ascent) [32].
    • BofA estimates that each +10 points of share is worth ~$1.4B of originations to SoFi (model estimate, secondary) [32].
  • Why it matters: Student loans are SoFi's best-performing credit (0.61% losses). New in-school volume would add high-quality balance-sheet growth.
  • 😈 Devil's advocate:
    • SoFi does not split in-school from refinancing. Q2 seasonality is weak for in-school, so most of the +170% is likely refinancing, which higher rates now hurt (CFO, 2026-09-08) [39].
    • Sallie Mae owns the school financial-aid channel.
    • Existing PLUS borrowers keep access until 2029 [32].
  • Verdict: Medium The policy shift is real, but SoFi's share is still unproven until Q3/Q4 disclosure.

🐂 Bull #7Capital is still above target and management says no raise

  • The point: SoFi has room to grow its balance sheet without new equity.
  • Evidence:
    • Holdco total capital was 18.8% (CET1 18.7%) at 2026-06-30, and Tier 1 leverage 16.5% [1].
    • The bank's total capital ratio was 15.3%, "well capitalized", against a 10.5% minimum with buffer [2].
    • CFO on 2026-07-29: "we're self-funded through our guidance period and the medium-term operating plan" [3]. He repeated "We don't see the need to raise any equity capital under our current operating plan" on 2026-09-08 [39].
    • No equity issuance, ATM or buyback was filed between June and September 2026 [2].
    • The Basel III re-proposal of 2026-03-19 would lower the risk weight on most retail exposures from 100% to 90% [20,21].
  • Why it matters: As long as no raise comes, earnings growth goes to existing holders.
  • 😈 Devil's advocate: The ratio fell from 22.9% to 18.8% in two quarters as risk-weighted assets grew from $37.2B to $48.7B [1,5] (Bear #4). The statement covers "the current operating plan", not a pledge, and management raised twice in 2025 [2].
  • Verdict: Medium True today. But the buffer is being spent on purpose, and the runway is a few quarters, not years.

🐂 Bull #8Loan-platform buyers are broadening beyond one or two funds

  • The point: The Loan Platform Business (SoFi originates loans for third-party buyers for a fee, without keeping the credit risk) is diversifying.
  • Evidence:
    • New buyers include Sixth Street (up to $1B of personal loans) and Basepoint ($3B of small-business loans over 3 years). Home-equity partners are expected at $100–200M a month [3].
    • Management says partner demand exceeds what SoFi fills [3].
    • The 10-Q says no individual buyer is ≥10% of consolidated revenue [2].
    • Q2 securitisations priced at 86 and 91bp spreads, "best-ever pricing" [3].
  • Why it matters: Capital-light fees (about 4.6% of volume in Q2: $143.3M ÷ $3.1B, my arithmetic) support growth without using capital.
  • 😈 Devil's advocate: LPB volume fell from $3.7B (Q4-25) to $3.0–3.1B in Q1–Q2, and its revenue share fell from 19.1% to 11.9% [1,4,5] (Bear #7). SoFi chose to hold more loans itself, which is exactly what uses capital.
  • Verdict: Weak to Medium. The buyer list is wider, but volume has stalled and SoFi is leaning on its balance sheet instead.

🐂 Bull #9The CEO bought about $2.25M of stock in 2026

  • The point: Anthony Noto buys with his own money in the open market.

  • Evidence: Noto bought about 130k shares across five dates in 2026, totalling about $2.25M [44,45]:

    Date Shares Price Value
    2026-03-02 56,000 $17.88 about $1.0M
    2026-03-17 28,900 $17.32 about $0.5M
    2026-05-08 15,878 $15.73 $250k
    2026-05-11 15,545 $16.00 $249k
    2026-06-16 13,888 $18.06 $251k

    The 2026-03-17 buy was on the day of the Muddy Waters report. Two other executives bought ~$100k each in February [44].

  • Why it matters: A CEO buying through a short attack signals that he does not expect accounting surprises.

  • 😈 Devil's advocate:

    • The amounts are small next to his ~12M-share holding.
    • Muddy Waters says Noto used prepaid variable forwards of ~$46.5M, which is not verified from filings [9,10].
    • There have been no buys since 2026-06-16, even at $15.23 [44] (Bear #12).

  • Verdict: Weak. It is a positive signal, but it is small and has gone quiet.

Bear case

🐻 Bear #1Profits rest on $2B of fair-value premium a short seller contests

  • The point: SoFi carries almost all its loans at a model price above face value. Changes in that price flow straight into revenue and equity. A credible short seller says the inputs are too generous and some structures hide debt.
  • Evidence:

    • The cumulative fair-value premium at 2026-06-30 was about $2.03B, about 18% of the $11.08B of equity (derived from the 10-Q/release tables) [1,2]:

      Loans Premium Mark
      Personal $1,222.8M 104.7%
      Student $704.6M 104.4%
      Home $99.6M n/a

    • Muddy Waters (2026-03-17) [9]:

      • All-in 2025 personal-loan losses were ~6.1%, not ~4.4%.
      • Personal-loan marks should be ~102%, not 106.4%, and it alleges $259M of unwarranted 2025 gains.
      • The student-loan discount rate was 27bp below the 10-year Treasury.
      • Its adjusted 2025 EBITDA estimate is ~$103M, against $1,054M reported.

    • Its follow-up (2026-03-30) cites a 2024-09-30 Utah UCC filing naming JPMorgan as "senior lender" and SoFi Bank as "mezzanine lender" on a $312M "sale" at par [10].

    • SoFi's 2026-03-17 response called the report a "fundamental lack of understanding" but gave no figures. No lawsuit against Muddy Waters has been found [11,12].

    • The Q2 10-Q shows $615.2M of secured loans to unconsolidated vehicles [2].

  • Why it matters: If marks fall 2 points on personal loans alone, that is roughly $520M off pre-tax income and book value (2% × $26.1B, estimate). Trust in the reported numbers is also what supports a 2x book multiple.
  • 🛡️ Counter-argument:
    • Deloitte audits the accounts, with no change in 2026 [2].
    • No restatement, material weakness, SEC inquiry or securities class action has been found [2,48].
    • The model's default assumption (4.77%) sits above the realised ~3.7% [1].
    • Delinquent-loan sales are disclosed in every release [1].
    • Muddy Waters said it would cover most of its short on publication [13].
  • Severity: High Fraud is not evident from anything primary. But the design makes reported profit and book sensitive to inputs, and the specific JPMorgan/UCC and unconsolidated-vehicle questions remain unanswered after seven months.

🐻 Bear #2Per-share profit is flat and Q3 needs a 26% EBITDA jump

  • The point: Revenue grows ~40% but per-share profit does not. The full-year guide now needs a steep H2 step-up.

  • Evidence:

    Q3-25 Q4-25 Q1-26 Q2-26
    Adjusted revenue $949.6M $1,012.8M $1,087.2M $1,205.6M
    GAAP net income $139.4M $173.5M $166.7M $156.6M
    Diluted EPS $0.11 $0.13 $0.12 $0.12
    Net margin 14.5% 16.9% 15.2% 12.9%

    Sources: [1,4,5,7].

    • Q2 net margin was the lowest of the four quarters.
    • Reaching the ~$1.6B FY26 EBITDA guide needs ~$902M in H2, ~$451M a quarter, against $357.8M in Q2 (+26%; my arithmetic from [1,4]).
    • Q3 consensus EPS is $0.17 against $0.12 in Q2 [37].
    • Estimated annualised ROTCE (return on tangible equity) is ~6.7%, against the 20–30% target [1,3].

  • Why it matters: At 25x this year's EPS, the stock needs the H2 ramp. A held-but-missed profit guide would break the 38–42% EPS-growth story.

  • 🛡️ Counter-argument: Management chose to reinvest ~$100–200M of revenue upside into marketing and new products, and says profit has upside "if there are no rate increases" [3]. In 2025 a similar H2 ramp was delivered [5,7].

  • Severity: High. It is the nearest test (2026-10-27), and the Fed hike the CFO named as the swing factor has already happened.

🐻 Bear #3Liability-sensitive balance sheet in a hiking cycle with 5% yields

  • The point: Higher rates hurt SoFi twice. Deposit costs rise faster than loan yields, and higher discount rates cut loan marks.
  • Evidence:
    • The 10-Q sensitivity is −$159.0M of NII per +100bp at 2026-06-30, against −$92.6M at 2025-12-31 [2].
    • The Fed hiked to 3.75–4.00% on 2026-09-16, voting 12-0 [16].
    • The median dot is 4.1% for both 2026 and 2027 [17].
    • Markets price ~17% odds of a hike on 2026-10-28 and expect the next in December [50].
    • The 10-year yield was ~5.18% on 2026-10-02 (single source) [50].
    • Q2 personal-loan marks fell 71bp, mostly because the discount rate rose from 4.61% to 4.97% [1,3].
  • Why it matters: Each 100bp costs ~$0.09 of EPS (−$159M × 0.78 ÷ 1.33B diluted shares, estimate), about 11% of FY27 consensus, plus one-off mark losses.
  • 🛡️ Counter-argument: Guidance already assumes 1–2 hikes in 2026, with "cushion" [3,39]. Higher rates also lift coupons on new loans, and the stock fell only 1.3% on hike day [46].
  • Severity: High One hike is in the guide, but a cycle that carries the funds rate above the dots is not.

🐻 Bear #4Capital ratio falls 2 points a quarter toward the target floor

  • The point: SoFi is spending its capital buffer quickly to hold loans on its balance sheet. Soon it must slow growth, sell more loans or raise equity.

  • Evidence:

    Date Holdco total capital ratio
    2025-09-30 20.2%
    2025-12-31 22.9%
    2026-03-31 21.3%
    2026-06-30 18.8%

    Sources: [1,4,5,7].

    • Risk-weighted assets grew from $37.2B to $48.7B in six months (+31%) [1,5].
    • The bank is already at 15.3% [2].
    • Management's long-term target is "low to mid-teens" [3].
    • Loop Capital (Hold, 2026-09-09) and Morgan Stanley ($15 target, 2026-07-30) cite more capital-intensive growth [39,42].

  • Why it matters: At −2 to −2.5 points a quarter, the holdco would reach 15% around Q4-26 to Q1-27 (estimate). After that, balance-sheet growth must be funded from earnings ($0.6B/yr) or new equity.

  • 🛡️ Counter-argument: Management says this glide path is planned and self-funded [3]. The Basel re-proposal could cut retail risk weights by 10% [21]. More loans can go through the platform business instead.

  • Severity: Medium-High. This is the mechanism behind any future raise (Bear #6).

🐻 Bear #5Default assumptions rise every quarter as jobs stall

  • The point: Realised losses are falling, but SoFi's own forward-looking default assumption has risen four quarters running, and the labour market is weakening.
  • Evidence:
    • The personal-loan default-rate assumption rose 4.33% → 4.46% → 4.57% → 4.77% from Q3-25 to Q2-26 [1,4,5,7].
    • September payrolls were +29k, July was revised to −10k, and unemployment rose to 4.2% [18].
    • Financial-activities jobs are down 129k since May 2025 [18].
    • Gasoline prices are up 27.4% year on year (August CPI) [19].
    • Credit-card charge-offs were 7.65% in Q2 [1].
    • TransUnion shows personal-loan originations +19.5% year on year, with subprime +29% [31].
  • Why it matters: SoFi's personal-loan book ($26.1B on balance sheet) is unsecured, and has never been through a recession at this size.
  • 🛡️ Counter-argument: Borrowers average a 746 FICO and $158k income (prior report [5]). The rising assumption is conservatism already in the marks. Prime tiers are growing slower than subprime [31].
  • Severity: Medium High. The assumption trend shows management sees the same risk. A recession would hit losses and marks together.

🐻 Bear #6Shares up 16% in a year, with more dilution queued

  • The point: Growth has been funded partly with new shares, and more are coming from stock pay and convertibles.
  • Evidence:
    • Shares outstanding were 1,290.3M at 2026-06-30, against 1,113.4M a year earlier (+15.9%). Diluted weighted shares rose 14.3% [1,2,8].
    • 2025 raises: 82.7M shares at $20.85 (July) and 54.5M at $27.50 (December) [2,51].
    • Stock-based compensation was $76.9M in Q2, about 6.3% of revenue [1].
    • Awards outstanding total ~63.4M, 4.9% of shares (13.4M options at $7.98, 45.5M RSUs, 4.5M PSUs) [2].
    • The 2029 notes ($862.5M) convert at ~$9.45 [14]. SoFi pays cash up to principal and shares above it. At $15.23 that is about 35M net shares, ~2.7% (my arithmetic). 43.3M convert shares were already in Q2 diluted EPS [1].
  • Why it matters: EPS growth of 38–42% a year needs the share count to stop rising. Muddy Waters estimates ~15%/yr dilution [10].
  • 🛡️ Counter-argument: Both raises came at prices well above today. Capped calls offset part of the 2029 conversion. No raise has happened since December [2].
  • Severity: Medium The past dilution is done. The ongoing ~5–8% overhang is modest, but another raise would hurt at $15.

🐻 Bear #7The loan-platform fee engine has stalled and buyers face redemptions

  • The point: The capital-light fee business that was meant to fund growth without capital is shrinking as a share of revenue. Its buyers, private-credit funds, are under redemption pressure.

  • Evidence:

    Q3-25 Q4-25 Q1-26 Q2-26
    LPB volume $3.4B $3.7B $3.0B $3.1B
    LPB revenue $167.9M $193.7M $140.8M $143.3M
    Share of adjusted revenue 17.7% 19.1% 12.9% 11.9%

    Sources: [1,4,5,7].

    • Blue Owl's credit income fund got redemption requests of 21.9% in Q1-26 and paid 5% [29].
    • Blue Owl's $5B SoFi agreement (2025-03-13) was reported as two years, implying renewal around March 2027 (secondary) [30].
    • Needham and Truist cut targets on weaker LPB volume [42,43].

  • Why it matters: Lower LPB means either slower growth or more capital use (Bear #4).

  • 🛡️ Counter-argument: Redemptions were driven by software-loan exposure, not consumer loans [29]. New buyers are signed (Bull #8). Management says it chooses volume based on unit economics [3].

  • Severity: Medium.

🐻 Bear #8Valuation still rich for a 7% ROTCE bank, and shorts doubled

  • The point: Even after a 42% fall, SoFi trades on growth multiples while earning bank-like or lower returns, and short sellers are adding.
  • Evidence:
    • At $15.23 the market cap is ~$19.7B [36,46]. That is 25x FY26 EPS, 18.6x FY27 and 2.07x tangible book (my arithmetic).
    • Ally trades at ~7x forward and Dave at ~17x (TIKR, secondary) [39].
    • Short interest was 214.9M shares (16.6% of outstanding, ~4.2 days to cover) at 2026-09-30, up from 100.3M a year earlier [35].
    • The analyst view is a consensus Hold: 9 Buy, 11 Hold, 3 Sell, with an average target $22.50 [36].
  • Why it matters: If EPS growth stalls, a re-rating toward bank multiples (10–12x) would mean a far lower price. A good Q3 could equally squeeze shorts.
  • 🛡️ Counter-argument: On FY28 consensus ($1.06) it is 14.4x [38]. Growth of 30%+ is not bank-like, and the options market implies only ±11% for Q3 [47].
  • Severity: Medium

🐻 Bear #9Debt nearly doubled and $1.3B of convertible cash is due by 2027

  • The point: SoFi re-drew warehouse lines it said it had "fully paid down" [6], and faces two cash calls from convertibles.
  • Evidence:
    • Total debt was $3.30B at 2026-06-30, against $1.82B at year-end, including $1.49B drawn on warehouses (zero at year-end) and a $486M revolver [2].
    • The 2026 notes ($428.0M, conversion price $22.41) mature on 2026-10-15. They are far out of the money, so expect cash repayment [2].
    • The 2029 notes ($862.5M) can be called from 2027-03-15 if the stock is above ~$12.29 (130% of $9.45). If called, holders convert and SoFi pays principal in cash [14,39].
    • The 10-Q says SoFi Bank "would have no funds free of restrictions available for dividend payments" [2].
    • Holdco-only cash: not found.
  • Why it matters: About $1.29B of cash is due in roughly 17 months, payable by the holding company, not the bank. If it is refinanced at ~7% instead of 0–1.25%, the extra interest costs ~$0.03 EPS a year (estimate).
  • 🛡️ Counter-argument:
    • Consolidated cash was $3.13B and investment securities $4.23B [1].
    • The revolver and warehouse lines are covenant-compliant [2].
    • The 2029 call is SoFi's option, not an obligation.
  • Severity: Medium Low. It is manageable, but it adds to the reasons capital is tight.

🐻 Bear #10More fintechs are getting bank charters

  • The point: SoFi's charter advantage is eroding as the OCC approves rivals.
  • Evidence:
    • Upstart got preliminary conditional OCC approval on 2026-07-23 [25].
    • Nubank got conditional approval on 2026-01-29 and launched US deposits at 3.50% on 2026-09-10 [26].
    • Revolut got conditional approval on 2026-09-03 [27].
    • Robinhood Banking reached ~$3–4B of deposits and has applied for a charter [28].
    • The OCC received 40 charter applications in 18 months [33].
  • Why it matters: Competition for the same high-earning depositors and borrowers compresses SoFi's funding and lending margins.
  • 🛡️ Counter-argument: None of these lends at SoFi's scale ($10.7B of personal loans a quarter). Most need 12–18 months to open their banks [25,26].
  • Severity: Low Medium over 36 months.

🐻 Bear #11The technology platform is shrinking after losing its anchor client

  • The point: The Galileo/Technisys segment that underpins the "tech company" narrative is contracting.
  • Evidence:
    • Tech Platform revenue was $122.4M (Q4-25), $75.1M (Q1-26) and $84.5M (Q2-26), −23% year on year [1,4,5].
    • Accounts were 134.8M, −16% [1].
    • Contribution margin was 14% [1].
    • A large client "fully transitioned off the platform" before 2025-12-31 [1]. Third parties name it as Chime, which SoFi has not confirmed [49].
  • Why it matters: It removes a source of non-credit revenue and weakens the case for a tech multiple.
  • 🛡️ Counter-argument: No client is ≥10% of revenue anymore [2]. The FY26 guide is ~$325M [52]. SoFi Bank's own migration onto the core is a reference win.
  • Severity: Low

🐻 Bear #12Insiders keep selling and the CEO stopped buying in June

  • The point: Apart from Noto's small buys, insider flow is selling.
  • Evidence:
    • CTO Rishel sold ~$3.5M in March and June and 18.6k shares in September, all under 10b5-1 plans [44].
    • EVP Keough sells ~10k shares every month under a 10b5-1 plan [44].
    • Director Freiberg sold 94k shares at $20.31 after an option exercise [44].
    • MarketBeat shows $0 bought and $891k sold in the last three months [36].
  • Why it matters: It is a mild sentiment signal. Nobody is adding at the lows.
  • 🛡️ Counter-argument: Almost all sales are pre-planned 10b5-1 sales or follow option exercises. They are routine.
  • Severity: Low

Tailwinds NOT in the forecasts

Consensus base used for all sizing: FY2027 revenue $6.09B (21 analysts) and EPS $0.82 (9 analysts) [37]. Diluted adjusted shares ~1.33B and tax rate 22% [1,3].

  • The Fed reverses and cuts at least 100bp by end-2027 — 25% chance by 2027-12-31

    • Impact on revenue: +$159M, +2.6% vs consensus FY2027.
    • Impact on EPS: +$0.09, +11.4% vs consensus FY2027 (run-rate; excludes one-off mark gains).
    • How we got the number (estimate): $159M NII sensitivity per 100bp (assumed symmetric) × (1 − 22% tax) ÷ 1.33B shares ≈ +$0.093.
    • Expected value: 25% × $0.093 ≈ +$0.023.
    • What it is: The oil-driven hiking cycle ends quickly and the Fed cuts back below 3%.
    • Why it's not priced in: The Fed's own dots show 4.1% at end-2027, and markets expect another hike in December [17,50]. SoFi's guide assumes 1–2 hikes [3].
    • Trigger to watch: Negative payroll prints, core CPI staying near 2.4% [19], and the December dot plot.
    • Why this probability: Core inflation is already at 2.4% and jobs are stalling, which argues for cuts. But headline 3.4% and 5% long yields make a full reversal within 15 months a minority case.

  • Loan-platform volume rebounds to about $4B a quarter in 2027 — 25% chance by FY2027

    • Impact on revenue: +$166M, +2.7% vs consensus FY2027.
    • Impact on EPS: +$0.068, +8.3% vs consensus FY2027.
    • How we got the number (estimate): +$0.9B/qtr × 4 × 4.6% fee yield (Q2 $143.3M ÷ $3.1B) ≈ $166M × 70% incremental margin (estimate, capital-light) × 0.78 ÷ 1.33B ≈ +$0.068.
    • Expected value: 25% × $0.068 ≈ +$0.017.
    • What it is: New buyers (Sixth Street, Basepoint, home-equity partners, the three unnamed partners with $3.6B of commitments) plus a Blue Owl renewal lift volume back above the Q4-25 peak.
    • Why it's not priced in: Volume has run at ~$3.0–3.1B for two quarters, and analysts cut targets on it [42,43]. As capital tightens (Bear #4), more volume may have to go through LPB anyway.
    • Trigger to watch: Q3/Q4 LPB volume, a named Blue Owl renewal before ~March 2027.
    • Why this probability: Capital pressure pushes SoFi toward LPB, but private-credit redemptions cap buyer appetite.

  • Basel III final rule cuts retail risk weights to 90% — 40% chance by end-2027

    • Impact on revenue: not quantifiable. It frees capital rather than adding revenue directly.
    • Impact on EPS: about +$0.024, +2.9% vs consensus FY2027, if freed capital avoids a raise or funds a buyback.
    • How we got the number (estimate): ~10% of ~$43B retail RWA = ~$4.3B RWA × 13% target ratio ≈ $0.56B capital ÷ $15.23 ≈ 37M shares ≈ 2.8% of diluted shares ⇒ +2.9% EPS.
    • Expected value: 40% × $0.024 ≈ +$0.010.
    • What it is: The 2026-03-19 standardized-approach proposal moves "Other Assets", including most retail exposures, from 100% to 90% risk weight [21].
    • Why it's not priced in: It is not final, and there is no effective date [21].
    • Trigger to watch: A final rule, and whether SoFi lowers its capital floor or announces buybacks.
    • Why this probability: The direction is deregulatory, and comments closed on 2026-06-18. But how much reaches consumer lenders is uncertain, and final rules often change.

  • Student loans become a multi-billion in-school line beyond the guide — 40% chance by FY2027

    • Impact on revenue: +$88M, +1.4% vs consensus FY2027.
    • Impact on EPS: +$0.028, +3.4% vs consensus FY2027.
    • How we got the number (estimate): +$1.4B originations (BofA +10pt share) × ~4% day-one fair-value gain ($56M) + ~$0.7B average balance × ~4.5% risk-adjusted spread ($32M) ≈ $88M × 55% Lending contribution margin × 0.78 ÷ 1.33B ≈ +$0.028.
    • Expected value: 40% × $0.028 ≈ +$0.011.
    • What it is: SoFi captures a real share of former Grad PLUS and Parent PLUS borrowing.
    • Why it's not priced in: SoFi does not disclose in-school volume, and the Lending guide predates the first peak season.
    • Trigger to watch: Q3 student originations, and any in-school split on 2026-10-27. Also Sallie Mae's late-October Q3 peak-season data.
    • Why this probability: The policy gap is sourced at $8–15B a year [32], but Sallie Mae dominates the school channel.

  • S&P 500 inclusion — 35% chance within 36 months

    • Impact on revenue: none.
    • Impact on EPS: not quantifiable. It affects the multiple, not earnings.
    • How we got the number (estimate): n/a.
    • Expected value: not quantifiable.
    • What it is: Index funds would be forced to buy.
    • Why it's not priced in: SoFi was passed over in March and June, and was not added in September (Bloom Energy, Illumina and others were) [52].
    • Trigger to watch: The December 2026 rebalance, which needs a higher market cap and four quarters of GAAP profit (already met).
    • Why this probability: It meets the profit test, but its ~$19.7B cap is below recent adds, and the committee has discretion.

  • SoFiUSD becomes a disclosed settlement and float business — 15% chance within 36 months

    • Impact on revenue: not quantifiable. Only $52.5M was outstanding at 2026-06-30 (search summary of company data) [56], and no revenue is disclosed.
    • Impact on EPS: not quantifiable.
    • How we got the number (estimate): n/a.
    • Expected value: not quantifiable.
    • What it is: SoFiUSD settles SoFi's ~$25B card programme on Mastercard and is listed by Payward [56]. It earns NII on the reserves, according to Noto [3].
    • Why it's not priced in: Consensus models it at zero, and the GENIUS Act final rules are not out (the OCC targets "by November") [33].
    • Trigger to watch: Circulation above $1B, and named merchant deals.
    • Why this probability: The total stablecoin market is ~$305B, dominated by USDT and USDC. A bank coin at $52.5M is very small.

Headwinds NOT in the forecasts

  • Consumer downturn: all-in personal-loan losses rise to about 5.5% — 25% chance by FY2027

    • Impact on revenue: −$470M, −7.7% vs consensus FY2027 (fair-value accounting runs losses through revenue).
    • Impact on EPS: −$0.28, −34% vs consensus FY2027, plus a one-off mark-down of ~$440M if marks fall ~1.7 points.
    • How we got the number (estimate): +1.8pt losses × $26.1B on-balance-sheet personal-loan UPB ≈ $470M × 0.78 ÷ 1.33B ≈ −$0.276.
    • Expected value: 25% × −$0.276 ≈ −$0.069.
    • What it is: The payroll stall becomes job losses and real-income pressure (gasoline +27%) reaches prime borrowers.
    • Why it's not priced in: The FY27 consensus of $0.82 implies +38% EPS growth with stable credit [37].
    • Trigger to watch:
      • Unemployment above 4.6%.
      • SoFi's add-back loss rate rising for two quarters.
      • Its default assumption moving above 5%.
    • Why this probability: Jobs are near zero and July was negative [18]. But SoFi's borrowers are prime and losses are still falling, so it is below one in three.

  • A new equity raise of about $1.5B — 25% chance within 36 months

    • Impact on revenue: roughly neutral; the capital supports growth already in consensus.
    • Impact on EPS: −$0.057, −7% vs consensus FY2027 (before earnings on new capital).
    • How we got the number (estimate): $1.5B ÷ $15.23 ≈ 98M shares ≈ 7.4% of 1.33B ⇒ $0.82 × (1 − 1/1.074) ≈ −$0.057.
    • Expected value: 25% × −$0.057 ≈ −$0.014.
    • What it is: SoFi raises equity again as the holdco ratio nears the low-to-mid teens.
    • Why it's not priced in: Management says "self-funded" [3], and consensus has no raise.
    • Trigger to watch:
      • Q3 holdco total capital ≤17%.
      • Any change to the capital target.
      • A shelf takedown after a rally.
    • Why this probability: The buffer runs out in ~2–4 quarters at the current pace (Bear #4), and SoFi raised twice in 2025. But at $15 a raise is expensive, and LPB is an alternative.

  • The 2029 converts are called and refinanced at market rates — 55% chance by end-2027

    • Impact on revenue: none. Interest expense sits below the revenue line here.
    • Impact on EPS: −$0.029, −3.6% vs consensus FY2027.
    • How we got the number (estimate): $862.5M × (7% − 1.25%) ≈ $50M × 0.78 ÷ 1.33B ≈ −$0.029.
    • Expected value: 55% × −$0.029 ≈ −$0.016.
    • What it is: From 2027-03-15 SoFi can call the notes if the stock is above ~$12.29. Holders would then convert, and SoFi pays principal in cash and the excess in shares [14].
    • Why it's not priced in: The net shares are in diluted EPS, but the loss of 1.25% funding is not obvious in consensus models (not verified).
    • Trigger to watch: A call notice after 2027-03-15, and any new note issue.
    • Why this probability: The stock is already ~60% above the call threshold, and calling removes dilution. But SoFi could leave the notes to 2029 if cash is tight.

  • The Fed hikes at least 50bp beyond the dots by mid-2027 — 30% chance by 2027-06-30

    • Impact on revenue: −$80M, −1.3% vs consensus FY2027.
    • Impact on EPS: −$0.047, −5.7% vs consensus FY2027, plus a one-off mark loss of ~$0.5B (half of the $1.04B per-100bp fair-value sensitivity in the prior report).
    • How we got the number (estimate): ½ × $159M × 0.78 ÷ 1.33B ≈ −$0.047.
    • Expected value: 30% × −$0.047 ≈ −$0.014.
    • What it is: The oil shock keeps headline inflation above 3%, and the Fed goes beyond the 4.1% median.
    • Why it's not priced in: The guide covers only 1–2 hikes in 2026 [3].
    • Trigger to watch: A hike on 2026-10-28 (priced at ~17%), December dots, and the 10-year above 5.25%.
    • Why this probability: Futures gave 37% odds of two more hikes on hike day [39], but core CPI at 2.4% argues against a long cycle.

  • Loan-platform buyers pull back (volume −25%) — 20% chance within 36 months

    • Impact on revenue: −$143M, −2.3% vs consensus FY2027.
    • Impact on EPS: −$0.059, −7.2% vs consensus FY2027.
    • How we got the number (estimate): $3.1B × 4 × 25% × 4.6% ≈ $143M × 70% × 0.78 ÷ 1.33B ≈ −$0.059.
    • Expected value: 20% × −$0.059 ≈ −$0.012.
    • What it is: Private-credit redemptions force buyers to cut forward-flow purchases, or the Blue Owl agreement is not renewed around March 2027.
    • Why it's not priced in: Management says demand exceeds supply [3].
    • Trigger to watch: LPB volume below $2.5B a quarter, and Blue Owl/Fortress renewal news.
    • Why this probability: Redemption stress is real [29], but it comes from software loans and the buyer list is broadening.

  • An accounting or regulatory event from the Muddy Waters allegations — 7% chance within 36 months

    • Impact on revenue: not quantifiable. It depends on what would be restated.
    • Impact on EPS: not quantifiable. Muddy Waters' own claim (EBITDA ~$1B lower) is an allegation, not a basis for a number.
    • How we got the number (estimate): n/a.
    • Expected value: not quantifiable. The probability is low, but the impact would be severe.
    • What it is: An SEC inquiry, restatement, material weakness or securities suit that survives dismissal.
    • Why it's not priced in: None has been disclosed after seven months [2,48].
    • Trigger to watch: 10-Q legal and controls sections, an 8-K on inquiries, an auditor change, or a UCC lien release on the JPMorgan structure.
    • Why this probability: The allegations are specific, but the core practices are disclosed and audited. Short reports on regulated banks rarely lead to restatements.

Investment Risk Score: 45/100 (moderate)

# Factor Weight Measured value Score Why
1 Liquidity & funding runway 15% Profitable; deposits $45.5B (98% insured); $13.68B remaining liquidity incl. $5.69B undrawn warehouses; consolidated cash $3.13B + securities $4.23B vs $428M converts due 10-15; FCF not meaningful for a bank (H1 operating CF −$6.2B = loan originations held for sale) 28 No runway issue, and funding grows faster than loans. Held off the lowest band because holdco cash is undisclosed, the bank cannot pay dividends, and ~$1.29B of convert cash is due by 2027.
2 Leverage 10% Holdco total capital 18.8%, Tier 1 leverage 16.5%; corporate debt $3.30B (≈2.6x LTM adj EBITDA $1.29B gross; roughly offset by $3.13B cash) 30 Bank-style leverage is comfortable, but debt nearly doubled in six months as warehouses were re-drawn.
3 Dilution (last 12 months + overhang) 10% Shares 1,113.4M → 1,290.3M (+15.9%); overhang ~63.4M awards (4.9%) + ~35M net 2029-convert shares (2.7%); SBC $77M/qtr 68 The +10–25% band, driven by the July and December 2025 raises, with a modest live overhang on top.
4 Profitability 10% GAAP net margin 12.9–16.9% over the last 4 quarters (Q2 12.9%); adj EBITDA margin 30%; ROTCE ~6.7% (estimate) 36 Net margin sits in the 5–15% band. Low ROTCE and dependence on fair-value marks keep it from scoring better.
5 Revenue quality & concentration 10% No customer or buyer ≥10% of revenue; NII 65% of revenue (recurring); LPB from a "limited pool" of buyers; fair-value premium $2.03B ≈ 18% of equity 38 Diversified and recurring, but part of each quarter's revenue is a model-driven mark.
6 Execution & track record 10% Q1 and Q2 2026 quarterly guides beaten, FY25 guide raised twice then beaten; FY26 profit guide held twice; warehouses "fully paid down" then $1.49B re-drawn; no restatement or material weakness found 35 Mostly hits, with small credibility misses on profit and capital messaging.
7 Competitive position & disruption 10% ~15% of US prime personal loans (CEO claim); Upstart/Nubank/Revolut charters, Robinhood banking ~$3–4B deposits; Galileo revenue −23% 50 Contested. A strong scale position in a commoditised product, with the charter moat eroding.
8 Cyclicality & commodity/macro sensitivity 10% NII −$159M per +100bp (was −$92.6M); unsecured personal loans $26.1B; PL default assumption 4.33→4.77% in 4 qtrs; payrolls +29k, July −10k 68 Highly cyclical. Rates and consumer credit hit NII, marks and losses at the same time.
9 Regulatory, legal & geopolitical 5% Muddy Waters allegations unanswered point by point; Smith cash-sweep class action (size not estimable); no SEC inquiry found; GENIUS final rules pending 45 Material but contained. No existential ruling is pending.
10 Valuation & expectations 10% 25x FY26 / 18.6x FY27 EPS, 2.07x TBV at ~7% ROTCE; short interest 214.9M (16.6%), doubled y/y; Q3 consensus needs +42% q/q EPS 62 Rich for the returns, with a crowded short and a high Q3 bar.

Overall: weighted average = 0.15×28 + 0.10×30 + 0.10×68 + 0.10×36 + 0.10×38 + 0.10×35 + 0.10×50 + 0.10×68 + 0.05×45 + 0.10×62 = 4.20 + 3.00 + 6.80 + 3.60 + 3.80 + 3.50 + 5.00 + 6.80 + 2.25 + 6.20 = 45.15 → 45 (moderate).

Change vs last: −4 (49 → 45). This is mainly the switch from the 6-area to the 10-factor method:

  • The new method scores funding/liquidity and leverage low, which suits a deposit-funded, well-capitalised bank.
  • It scores dilution (+15.9%) and macro sensitivity separately and high (68 each).
  • Facts moved both ways. Losses improved (3.7%), but the capital ratio fell, debt rose and short interest doubled.

Ranking check: At 45 SOFI ties VLN (45) and sits just below PGY (47). It is safer than Pagaya, which has retained risk paper of 1.75x equity and 63% related-party revenue on a much smaller capital base. It is clearly riskier than NU (33), which earns ~28% ROTCE with flat share count against SoFi's ~7% and +16%.

Conclusion

  • What you are really betting on: SoFi turns a deposit-funded, 15.8M-member franchise into rising per-share profit before the credit and rate cycle tests its fair-value marks, without another equity raise.
  • The 2–3 things that will decide the outcome:
    1. The H2 profit ramp. Q3 (2026-10-27) needs EBITDA of ~$451M and EPS ~$0.17, against $358M and $0.12 in Q2. That has to happen after a Fed hike, and on the same days as the 2026-10-27/28 FOMC meeting.
    2. Credit vs assumptions. Realised all-in personal-loan losses (3.7%) must stay below the 4.77% model default rate while payrolls stall.
    3. Capital runway. The holdco ratio is falling ~2 points a quarter. SoFi must either rebuild LPB volume or slow balance-sheet growth to avoid a raise.
  • Signposts:
    • Confirm bull:
      • Q3 adj EBITDA ≥$430M and the FY26 profit guide raised.
      • Add-back personal-loan losses ≤4% with the default assumption flat.
      • LPB volume back above $3.5B, or a named Blue Owl renewal.
      • Holdco capital ≥17% at Q3 and no raise.
    • Break bull:
      • Q3 EBITDA below ~$400M, or the FY26 EBITDA/EPS guide cut.
      • Personal-loan marks below 103.5%, or the default assumption above 5%.
      • Any equity issue, or a change in capital target.
      • A detailed SEC inquiry, restatement or material weakness linked to the secured-loan/VIE structures.
  • New since the last report:
    • The FY26 revenue guide was in fact raised (to $4.75–4.85B). Only profit lines were held, and the hidden cost is the tax-rate change (22% vs mid-teens ≈ $0.05).
    • Quantified, from the filings:
      • SoFi's own default assumption has risen four quarters in a row (4.33 → 4.77%) while realised losses fell.
      • The fair-value premium is ~$2.03B (18% of equity).
      • Debt nearly doubled to $3.30B after the warehouses were re-drawn.
    • The 2029 notes convert at ~$9.45 and are callable from 2027-03-15, a ~$862.5M cash call. Short interest doubled to 214.9M. The "17% short interest" claim was correct, and so was the March S&P pass-over, but the second pass-over was in June, not September.

Open questions & data gaps

  • The Q2-26 10-Q Part II (risk factors, legal) and the FY2025 10-K were not read in full. So a disclosed SEC inquiry, the parent-only cash balance and the treatment of the JPMorgan/UCC structure remain unverified. This matters most for Bear #1 and Bear #9.
  • 2026 notes settlement (2026-10-15): cash repayment is inferred, not confirmed by an 8-K.
  • 2029 notes: whether any holder converted in the Jul–Sep 2026 window is not found.
  • Holdco-only cash: not found. Parent cash was only $30.8M at 2024-12-31 (FY2024 10-K) [55], so convertible cash likely needs upstream dividends from non-bank subsidiaries or the revolver. This matters for the $1.29B of cash calls.
  • Numeric 30+/90+ delinquency by product, beyond the 40bp personal-loan figure: not found.
  • In-school vs refinancing split of student originations: not disclosed.
  • Blue Owl agreement term and renewal date: only a secondary "two-year" description.
  • Muddy Waters' prepaid-variable-forward claim against the Form 4/proxy disclosures: not checked.
  • Why management moved to a 22% tax rate, and whether FY27 consensus uses it: not found.
  • Brent level and the exact 10-year yield: single or conflicting sources. They are used only as context.
Sources
  1. SoFi Q2 2026 Earnings Release (8-K Ex. 99.1) — SoFi / SEC EDGAR — 2026-07-29 — https://www.sec.gov/Archives/edgar/data/0001818874/000181887426000050/a2026q2earningsrelease.htm
  2. SoFi Technologies Form 10-Q for the quarter ended 2026-06-30 (read in part) — SEC EDGAR — 2026-08-06 — https://www.sec.gov/Archives/edgar/data/0001818874/000181887426000054/sofi-20260630.htm
  3. SoFi Q2 2026 Earnings Call Transcript (Motley Fool copy; last ~20% of Q&A not read) — Mitrade — 2026-07-29 (copy 2026-08-08) — https://www.mitrade.com/insights/news/live-news/article-8-1978081-20260808
  4. SoFi Q1 2026 Earnings Release and Form 10-Q — SoFi / SEC — 2026-04-29 / 2026-05-07 — https://s27.q4cdn.com/749715820/files/doc_financials/2026/q1/2026-Q1-Earnings-Release_FINAL.pdf
  5. SoFi Q4 2025 Earnings Release — SoFi — 2026-01-30 — https://s27.q4cdn.com/749715820/files/doc_financials/2025/q4/2025-Q4-Earnings-Release.pdf
  6. SoFi Q4 FY2025 Earnings Conference Call Transcript — SoFi IR — 2026-01-30 — https://s27.q4cdn.com/749715820/files/doc_financials/2025/q4/SoFi-Q4-FY-2025-Earnings-Conference-Call-Transcript.pdf
  7. SoFi Q3 2025 Earnings Release — SoFi — 2025-10-28 — https://s27.q4cdn.com/749715820/files/doc_financials/2025/q3/2025-Q3-Earnings-Release_FINAL.pdf
  8. SoFi Q2 2025 Earnings Release — SoFi — 2025-07-29 — https://s27.q4cdn.com/749715820/files/doc_financials/2025/q2/2025-Q2-Earnings-Release.pdf
  9. Muddy Waters short report on SoFi — Muddy Waters Research — 2026-03-17 — https://muddywatersresearch.com/wp-content/uploads/2026/03/MWR_20260317.pdf
  10. How SOFI's Borrowing Money and Booking it as a Sale Enriched Management — Muddy Waters Research — 2026-03-30 — https://muddywatersresearch.com/wp-content/uploads/2026/03/MW_SOFI_BorrowingSales_20260330.pdf
  11. SoFi Responds to Inaccurate Short Seller Report — SoFi / Business Wire — 2026-03-17 — https://s27.q4cdn.com/749715820/files/doc_news/SoFi-Responds-to-Inaccurate-Short-Seller-Report-2026.pdf
  12. Muddy Waters, SoFi stock and accounting — Fortune — 2026-03-25 — https://www.fortune.com/2026/03/25/muddy-waters-sofi-stock-accounting
  13. SoFi: The Short Thesis — Motley Fool — 2026-03-24 — https://www.fool.com/investing/2026/03/24/sofi-the-short-thesis/
  14. SoFi Technologies Prices $750 Million Convertible Senior Notes Offering Due 2029 — Business Wire — 2024-03-05 — https://www.businesswire.com/news/home/20240305635218/en/SoFi-Technologies-Inc.-Prices-750-Million-Convertible-Senior-Notes-Offering-Due-2029
  15. SoFi Schedules Conference Call to Discuss Q3 2026 Results — SoFi / Business Wire — 2026-09-29 — https://s27.q4cdn.com/749715820/files/doc_news/SoFi-Schedules-Conference-Call-to-Discuss-Q3-2026-Results-2026.pdf
  16. FOMC statement — Federal Reserve — 2026-09-16 — https://www.federalreserve.gov/newsevents/pressreleases/monetary20260916a.htm
  17. Summary of Economic Projections — Federal Reserve — 2026-09-16 — https://www.federalreserve.gov/monetarypolicy/fomcprojtabl20260916.htm
  18. The Employment Situation — September 2026 — BLS — 2026-10-02 — https://www.bls.gov/news.release/archives/empsit_10022026.htm
  19. Consumer Price Index — August 2026 — BLS — 2026-09-11 — https://www.bls.gov/news.release/cpi.nr0.htm
  20. Agencies request comment on capital proposals — Federal Reserve — 2026-03-19 — https://www.federalreserve.gov/newsevents/pressreleases/bcreg20260319a.htm
  21. Federal banking agencies re-propose Basel III endgame — Debevoise & Plimpton — 2026-03 — https://www.debevoise.com/insights/publications/2026/03/federal-banking-agencies-basel-iii-endgame-mullig
  22. SLM Corp Q2 2026 earnings release (8-K) — SEC EDGAR — 2026-07-23 — https://www.sec.gov/Archives/edgar/data/0001032033/000103203326000041/slm072326ex991.htm
  23. SLM Q2 2026 earnings call transcript — Earnings Whispers — 2026-07 — https://earningswhispers.com/transcript/SLM/Q22026
  24. Upstart Q2 2026 results (8-K) — SEC EDGAR — 2026-08-04 — https://www.sec.gov/Archives/edgar/data/0001647639/000164763926000062/upst991prq22026.htm
  25. OCC Corporate Decision #1382 (Upstart) — OCC — 2026-07-23 — https://www.occ.gov/topics/charters-and-licensing/interpretations-and-decisions/2026/cd1382.pdf
  26. Nubank introduces US deposit accounts via partner bank — Crowdfund Insider — 2026-09 — https://www.crowdfundinsider.com/2026/09/310325-nubank-introduces-us-deposit-accounts-via-partner-bank/
  27. Revolut gets conditional OCC approval — The Block — 2026-09-03 — https://theblock.co/news/regulation/2026-09-03-revolut-takes-important-step-toward-becoming-a-us-bank-with-conditional-occ-approval-413474
  28. Robinhood Banking surpasses $4B assets — Crypto Briefing — 2026 — https://cryptobriefing.com/robinhood-banking-surpasses-4b-assets/
  29. Blue Owl credit funds face heavy redemption requests — Semafor — 2026-04-02 — https://www.semafor.com/article/04/02/2026/blue-owl-credit-funds-face-heavy-redemption-requests-as-private-credit-jitters-persist
  30. SoFi Expands Loan Platform Business with $5 Billion Agreement with Blue Owl — Business Wire — 2025-03-13 — https://www.businesswire.com/news/home/20250313368726/en/SoFi-Expands-Loan-Platform-Business-with-5-Billion-Agreement-with-Blue-Owl-Capital-Funds
  31. TransUnion Q2 2026 CIIR: consumer credit divergence (summary of TransUnion report) — Stacking Capital — 2026 — https://www.stacking.capital/articles/transunion-q2-2026-ciir-consumer-credit-divergence-subprime-rising-balance-level-stable-small-business-funding-approval-implications
  32. Can private markets fill the gap? Grad PLUS elimination (NBER w35683) — NBER — 2026 — https://www.nber.org/papers/w35683
  33. OCC promises final rule for GENIUS Act by November — PYMNTS — 2026-08-19 — https://www.pymnts.com/legal/2026/occ-promises-final-rule-for-genius-act-by-november/
  34. SoFi launches first bank-issued stablecoin SoFiUSD — Loeb & Loeb — 2026-06 — https://www.loeb.com/en/insights/passle/2026/06/sofi-launches-first-bankissued-stablecoin-sofiusd-on-consumer-banking-app
  35. SOFI short interest (FINRA data) — Equibles — through 2026-09-30 — https://equibles.com/stocks/sofi/shortinterest
  36. SoFi Technologies given average rating of Hold by analysts (data as of 2026-10-12) — MarketBeat — 2026-08-02 / 2026-10-12 — https://www.marketbeat.com/instant-alerts/sofi-technologies-inc-nasdaqsofi-given-average-rating-of-hold-by-analysts-2026-08-02/
  37. SOFI analyst estimates — Yahoo Finance — retrieved 2026-10-13 — https://finance.yahoo.com/quote/SOFI/analysis/
  38. SoFi's unstoppable growth points to strong upside — TipRanks — 2026-08-18 — https://www.tipranks.com/news/sofis-unstoppable-growth-points-to-strong-upside
  39. SoFi's guidance already assumed Fed rate hikes: the capital math — TIKR — 2026-09-28 — https://www.tikr.com/blog/sofis-guidance-already-assumed-fed-rate-hikes-heres-the-capital-math-investors-should-watch
  40. SoFi stock: Piper Sandler Overweight — TIKR — 2026-08-19 — https://www.tikr.com/blog/sofi-technologies-stock-piper-sandler-overweight-rating
  41. Scotiabank initiates coverage of SoFi with Sector Outperform — Nasdaq / Fintel — 2026-09-02 — https://www.nasdaq.com/articles/scotiabank-initiates-coverage-sofi-technologies-sofi-sector-outperform-recommendation
  42. SOFI stock surges to its best day of 2026 after record Q2 revenue; Wall Street slashes targets — Stocktwits — 2026-07-30 — https://stocktwits.com/news-articles/markets/equity/sofi-stock-surges-to-its-best-day-of-2026-after-record-q2-revenue-why-is-wall-street-slashing-price-targets/cZNEM4zRJ2S
  43. SoFi stock price forecast: what analysts expect — TipRanks — 2026-06-08 — https://www.tipranks.com/news/sofi-stock-price-forecast-what-financial-analysts-expect-right-now
  44. SOFI SEC filings and insider transactions (Form 4 compilation) — Chartrow — retrieved 2026-10-13 — https://chartrow.com/quote/sofi/filings
  45. SoFi insider activity (Form 4 compilation) — Dataroma — retrieved 2026-10-13 — https://www.dataroma.com/m/stock.php?sym=SOFI
  46. SOFI historical prices — FinancialContent — through 2026-10-12 — https://www.financialcontent.com/quote/NQ%3ASOFI/historical
  47. SoFi stock before earnings (implied move) — EarningsWatcher — 2026-10-10 — https://earnings-watcher.com/wiki/sofi-stock-before-earnings
  48. SoFi securities investigations and lawsuits 2026 — ClaimDepot — 2026 — https://www.claimdepot.com/investigations/sofi-securities-2026
  49. Galileo's total accounts drop explained — fintechdd (Substack) — 2026 — https://fintechdd.substack.com/p/galileos-total-accounts-drop-explained
  50. Weak September jobs report lowers odds of an October Fed hike — MPA Mag — 2026-10 — https://www.mpamag.com/us/mortgage-industry/industry-trends/weak-september-jobs-report-lowers-odds-of-an-october-fed-hike/592115
  51. SoFi pricing of public offering (8-K Ex. 99.2) — SEC EDGAR — 2025-12-04 — https://www.sec.gov/Archives/edgar/data/1818874/000181887425000225/ex992pricingpressreleasede.htm
  52. S&P 500 September 2026 rebalance — iTiger — 2026-09 — https://www-web.itiger.com/news/1115350276 (Galileo FY26 ~$325M guide per TIKR Q1 review: https://www.tikr.com/blog/sofi-technologies-beat-earnings-and-still-fell-13-is-sofi-stock-a-buy-in-2026)
  53. SoFi stock's worst quarter; Muddy Waters JPMorgan warning — Stocktwits — 2026-03-31 — https://stocktwits.com/news-articles/markets/equity/sofi-stock-worst-quarter-muddy-waters-jpmorgan-warning-1b-ebitda-risk/cZEPV9IRINx
  54. LendingClub (Happen) Q2 earnings call highlights — MarketBeat — 2026-07-27 — https://www.marketbeat.com/instant-alerts/lendingclub-q2-earnings-call-highlights-2026-07-27/ (search summary only)
  55. SoFi Technologies Form 10-K FY2024, parent-company condensed balance sheet (parent cash $30.8M at 2024-12-31) — SEC EDGAR — 2025-02 — https://www.sec.gov/Archives/edgar/data/1818874/000181887425000016/Financial_Report.xlsx (search summary only)
  56. SoFi company news feed (Payward/SoFiUSD listing; SoFiUSD $52.5M outstanding at 2026-06-30) — Autobound (aggregator, search summary only, UNCONFIRMED) — 2026 — https://www.autobound.ai/companies/sofi/news
All SOFI Bulls vs Bears reports
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports