SoFi Technologies, Inc.

▲ Bull Case

• ◆Bank layer: NIM holds at 5.8%+, personal loan charge-offs stay below 2%, and loan book grows 25%+ through FY2027 on robust HENRY borrower demand — ROTCE reaches 13% by FY2027, triggering P/TBV expansion from 2.0x to 3.0x.
• ◆Membership layer: 20M+ members by FY2027 with products per member reaching 2.8 (vs. 2.2 today) as HENRY cohorts from 2022–2024 age into peak earning years, take mortgages, increase investing AUM, and deepen their SoFi financial stack, driving 15%+/yr revenue-per-member growth organically.
• ◆Infrastructure layer: Galileo wins 3–5 large enterprise non-fintech clients in 2026–2027, revenue reaches $600M, and the combined Galileo/Technisys stack attracts mid-tier bank core replacement deals in a $10B+ TAM — catalyst stack of S&P 500 inclusion + Galileo re-acceleration + credit resilience drives $32/share in 18–24 months.

▼ Bear Case

• ◆Credit quality deteriorates: 2023–2025 personal loan vintages ($35B+ originated during peak confidence) show NCO deterioration as unemployment drifts above 5.5%; at 3.2% NCO on a $50B personal loan book, annual losses of $1.6B exceed FY2027E earnings, requiring defensive reserve builds that shrink earnings.
• ◆NIM compresses: Fed cuts 150bps through 2026–2027 compress NIM from 5.94% toward 5.2%; on a $60B loan book, 74bps NIM compression equals ~$444M annual NII reduction, with asset-sensitive balance sheet declining faster in yield than deposit costs.
• ◆Galileo stalls and multiple compresses to bank parity: account growth remains at 1–2% YoY, a large client migrates to in-house infrastructure, revenue is flat at $450–480M, goodwill impairment charges hit, and the market applies LC-equivalent 1.3x P/TBV as ROTCE falls below WACC — stock rerates toward $13.

“The primary analytical debate is whether SoFi is a bank (1.2–2.0x P/TBV) or a bank with an embedded technology business (2.5–4.0x P/TBV). The bear view holds that at $16.72 SoFi trades at 2.0x TBV despite ROTCE of only 6–7% (well below WACC of 9–10%), a premium unjustified until ROTCE exceeds WACC, with Galileo's stalling growth and Muddy Waters LPB questions reducing conviction the technology premium will materialize. The bull view argues the market is systematically wrong to apply bank-only multiples — the bank charter saves ~$2B/year vs. pre-charter warehouse financing, Galileo's 160M accounts generate 32% contribution margins on zero capital, and ROTCE crossing WACC is a question of 'when' not 'if,' with FY2027–2028 the expected crossing. The debate resolves empirically on two data points: (1) whether Galileo revenue accelerates above 15% in FY2026, and (2) whether ROTCE crosses 10% in FY2027 — if both occur, $28–32 is achievable; if Galileo stagnates and ROTCE stays at 7–8%, the bank-only framing dominates and $13–16 is the range.”

• ◆S&P 500 inclusion — eligibility confirmed (4+ GAAP-profitable quarters, $18B+ market cap); passive fund mandatory buying could add $2–4 to stock price on announcement
• ◆FY2026 EPS delivery vs. $0.60 adj. guidance — beating by >10% would demonstrate earnings power and narrow P/E discount to growth
• ◆Galileo revenue re-acceleration — any quarter showing >15% YoY Tech Platform growth signals $1B target is achievable and re-rates SOTP value
• ◆Personal loan credit quality — every quarter without NCO deterioration above 2.5% confirms the HENRY underwriting moat in a slowing economy
• ◆ROTCE crossing WACC (~9–10%) — expected late FY2027; when ROTCE exceeds WACC the bank should trade above 2.0x TBV, justifying P/TBV expansion to 2.5–3.0x
• ◆Products per member reaching 3.0 — validates the CLV flywheel and switching-cost moat; management is actively targeting this milestone
• ◆SoFiUSD stablecoin adoption — capital-light fee revenue stream potentially worth $50–200M over 3 years; currently unquantified upside option not in base case
• ◆Chime post-IPO strategy announcement — if Chime publicly commits to remaining on Galileo for 3+ years, largest client concentration risk resolves

• ◆Personal loan NCO rate >3.5% for 2+ quarters (KS-1): at 3.5% NCO on $25B+ unsecured book, annual losses of $875M+ approach earnings levels; probability 20–25%
• ◆LPB revenue quality investigation (KS-6): Muddy Waters raised concerns about Loan Platform Business revenue sustainability; an SEC inquiry or restatement would destroy market confidence; probability ~8%
• ◆NIM compression to <5.0% (KS-2): 150bps+ Fed cuts compress NIM from 5.94% toward 4.8%; on $60B+ loans, 114bps compression equals ~$684M NII reduction; probability 20–25% in recession
• ◆Galileo revenue stagnation (KS-3): Tech Platform flat/declining for 2+ quarters triggers goodwill impairment and removes the technology premium from valuation; probability 30–35%
• ◆Chime migration off Galileo (KS-7): Chime (estimated 30%+ of Galileo revenue) announces proprietary infrastructure build post-IPO; SOTP Layer 3 value drops $1–2/share overnight; probability 20%
• ◆Robinhood/Apple captures HENRY wallet: Robinhood Gold (banking + brokerage + credit card) directly targets SoFi's member demographic; if member growth decelerates below 15% the flywheel slows; probability 30% and growing