The comfortable reading is that SoFi was cut in half because the business broke. That reading is wrong. SoFi closed Friday, 23 October 2026, at $14.615. Through 31 December the last 90 daily log-returns put a bull at $18.22, a base at $13.22 and a bear at $10.68. The base sits $1.395 under that close because the mean of those returns is negative.
What the coverage skips is the calendar. The third-quarter release is set for about 7 a.m. Eastern on Tuesday, 27 October, and SoFi Earn was announced on 22 October, after the quarter had ended. The 4.40 percent variable yield is not in the print. The open question is the fair-value mark on the loan book, which Keefe, Bruyette & Woods has described as more than $1 billion before hedges.
Key facts
- Nasdaq official close, Friday 23 October 2026: $14.615, down $0.155, or 1.05 percent, from Thursday’s $14.77. Low $14.235. Nasdaq file downloaded 24 October 2026.
- Nasdaq historical file, 24 October 2026: 20-day average $15.2757, so the close is 4.33 percent under it. 50-day average $16.7405, so the close is 12.70 percent under it. Prior five sessions, 16 to 22 October: low $14.18, high $15.54.
- File high, 12 November 2025: close $32.21, intraday high $32.73. Friday is 54.63 percent under that close. The 10 July 2026 close was $18.78. Nasdaq file, 24 October 2026.
- Company notice, 29 September 2026: third-quarter results at about 7 a.m. Eastern on Tuesday, 27 October 2026, conference call at 8 a.m. Eastern. SoFi investor relations.
- SoFi Earn release, 22 October 2026: each SOFIE is $1 at SoFi Bank, 4.40 percent APY, FDIC insurance up to applicable limits, convertible 1:1 into SoFiUSD. Launch release.
- Second-quarter slides, 29 July 2026: adjusted net revenue $1.2 billion, up 40 percent. Adjusted EPS $0.12. Deposits $46 billion. 2026 revenue guide $4.75 billion to $4.85 billion. EPS guide about $0.60. Investor presentation.
- TipRanks, fetched 24 October: Tim Switzer at Keefe, Bruyette & Woods kept a Sell and cut the target to $15 from $16 on a pre-hedge rate hit above $1 billion. PriceTargets, same day: consensus $22.05, high $35, low $15. Mizuho’s Dan Dolev cut to $22 from $29 on 31 July.
- House path computed 24 October 2026 from the $14.615 Nasdaq close, not a broker target: bull $18.22, base $13.22, bear $10.68. N is 47. Mean log-return -0.00213423. Sample standard deviation 0.03118738. Not the July pair of $31 and $12.
What just happened, and why the obvious reading is wrong
Two things landed in the same week, and they are not the same event. On Thursday, 22 October, SoFi announced SoFi Earn, ticker SOFIE, and called it the first national bank-issued tokenized deposit for consumers. On Friday the shares closed at $14.615, down 1.05 percent from $14.77. The closes used here are Friday’s completed prints. A midday screen and an after-hours print are not the spot.
The other event has not happened. On 29 September the company said it would release third-quarter results at about 7 a.m. Eastern on 27 October and hold the call at 8 a.m. Earnings Whispers, in a Friday calendar post, lists the ticker among reporters for the week of 26 October. The clock on the release is the company’s. The quarter the call will discuss ended in September. SOFIE did not exist for that quarter.
The half-off reading uses the right tape and the wrong cause. The file high is the 12 November 2025 close of $32.21 and the intraday high of $32.73. Friday’s $14.615 is 54.63 percent under that close. Half of $32.21 is $16.105, and Friday is through it. The last reported quarter did not miss on the company’s slides.
The 29 July slides show adjusted net revenue of $1.2 billion, up 40 percent, and adjusted earnings of $0.12. They raised full-year adjusted net revenue to $4.75 billion to $4.85 billion, from a prior guide near $4.655 billion, and left adjusted earnings per share near $0.60. The session that took that news closed at $15.25, down 8.90 percent from $16.74. Friday is under that print. A token announced on 22 October does not explain a high printed on 12 November 2025.
Who is exposed
The slides put members at 15.814 million, up 35 percent, and deposits at $46 billion, up $5.3 billion. The funding page says the book was about 93 percent deposit funded, with a net interest margin of 5.98 percent. Products per member were 1.54, and 51 percent of new products were opened by existing members.
Lending adjusted net revenue was $711.672 million, against $446.798 million a year earlier. Financial services were $466.278 million. Technology-platform revenue fell to $84.505 million from $109.833 million, and contribution profit fell to $11.772 million. Accounts on that line were 134.804 million, against 160.046 million. The loan book grew. The enterprise line did not. The highlights page still prints $1.2 billion.
Originations were a record $14.8 billion, including $3.1 billion of personal loans written for third parties, against $2.4 billion a year earlier. Those partners take the credit. SoFi takes a fee. Borrowers, depositors, and those partners are the operating exposure. Holders from the 12 November close of $32.21 are the price exposure.
SoFiUSD, as of those July slides, had roughly $300 million in circulation. That is the rail SOFIE converts into. It is not a third-quarter earnings line, and it was not large next to $1.2 billion of adjusted net revenue. A July account of the quarter is here: crypto revenue against the stablecoin push. The earlier steps are the stablecoin launch and Big Business Banking.
What the tape and the slides show
The chart is the Nasdaq official daily close. It runs from 1 October 2025 through Friday’s $14.615, with daily lows on the top panel, so the 12 November high of $32.73 and the 15 October low of $13.86 both fit. The lowest close in the file is 14 October, at $14.22. Friday’s low was $14.235.
The close is $0.661 under the 20-day average of $15.2757, or 4.33 percent, and $2.126 under the 50-day average of $16.7405, or 12.70 percent. It is $0.925 under the prior five-session high of $15.54 and $0.435 above the prior five-session low of $14.18. Both averages are overhead.
The largest log-return in the 90 is 3 August, up 10.55 percent from $16.31 to $18.03, and it stayed in the sample. Sigma is not one session. The Nasdaq info page still stamped this $14.615 sale as 22 October. The historical row dates it 23 October, and that row is the spot.
| Reading | Figure | What it is, and what it is not |
|---|---|---|
| Official close | $14.615 on 23 October | Nasdaq daily close. The formula spot. |
| Prior close | $14.77 on 22 October | The session of the SOFIE announcement. Down $0.155 the next day. |
| November high | $32.21 close, $32.73 high, on 12 November 2025 | The top of this file. Friday is 54.63 percent under the close. |
| Last reported quarter | Adjusted net revenue $1.2 billion, EPS $0.12 | 29 July slides. Not the 27 October print. |
| Street, separate | Consensus $22.05. Mizuho $22. KBW $15. | Fetched pages. Not this bull or this bear. |
| House bull, base, bear | $18.22, $13.22, $10.68 | 47 sessions and the last 90 log-returns. Not $31 and $12. |
From the 10 July close of $18.78, Friday is down 22.18 percent. The older English page’s $31 bull and $12 bear, set against a $21.10 consensus, are not this title.
What SoFi said, and what it has not reported
Anthony Noto, chief executive, in the 22 October release: “Too much money in the digital asset ecosystem is sitting idle, leaving massive potential earnings on the table for millions of people.” He followed it with the product claim: “We believe your money should keep working for you until the moment you need it. SoFi Earn combines the earning power and protections of a bank deposit with the speed and utility of digital assets, giving people more value, flexibility and control over their money.”
He posted the same morning, at 13:37 GMT: “Today, we launched @SoFi Earn (SOFIE), the first and only national bank-issued tokenized deposit for consumers. We’re bringing the best of banking to the digital money ecosystem, because your money should keep working for you until the moment you need it.” The status is the source. The company account wrote: “Meet SoFi Earn (SOFIE), with 4.40% APY* and FDIC protection.” That post calls it the first bank deposit token for SoFi Crypto. None of those sentences quotes a share price.
Chris Lapointe, chief financial officer, is not on the SOFIE release. At a Goldman Sachs conference on 8 September, as reported by TIKR on 20 October, he said fee streams let SoFi “use less equity to drive even more revenue.” TIKR does not print the transcript. What management has not said is the third-quarter fair-value mark, the hedge offset, or whether the $4.75 billion to $4.85 billion revenue range and the $0.60 guide still stand. The slides say that $0.60 now assumes a 22 percent tax rate, against a mid-teens rate when it was first set.
The rate mark, the charter, and a planned sale
SOFIE is a deposit at SoFi Bank, N.A., not a stablecoin with a coupon. The rate sheet effective 20 October lists 4.31368 percent interest and a 4.40 percent APY. The rate is variable. The sheet says the balance “is not currently eligible for and will not be placed in the SoFi Insured Deposit Program.” It stays at SoFi Bank, with FDIC insurance up to applicable limits. Conversion into SoFiUSD leaves the deposit.
That is why the product does not answer the KBW note. TipRanks, updated 23 October, says Tim Switzer kept a Sell and cut the target to $15 from $16, on a third-quarter rate hit of more than $1 billion before hedging. TIKR dates the cut to 12 October and calls the rating Underperform. Both pages put the target at $15. The figure is theirs, not a loss SoFi has reported.
The last published capital ratio is still June. CET1 was $9.113 billion, 18.7 percent, against a 7.0 percent minimum, and Tier 1 leverage was 16.5 percent against 4.0 percent. A pre-hedge hit above $1 billion matters only if the hedge does not cover it. TIKR says the July guide assumed one to two rate hikes, and that traders put the odds of a hike on 28 October at about 50 percent. That is an input to the mark, not a replacement for $18.22 or $10.68.
The insider sale is the other false tell. StockTitan’s account of the Form 4 says Kelli Keough, executive vice president, sold 10,203 shares on 20 October at a weighted average of $14.7226 under a Rule 10b5-1 plan adopted on 30 July 2025, and held 397,870 shares afterward. The sale is about $150,215 against Friday volume of 38.9 million shares. It does not set the close.
Bull, base, and bear
The title numbers are this distribution, not a broker target. The English pair already printed is a $31 bull and a $12 bear on the 10 July close of $18.78. This bull is $18.22, $12.78 away from $31. This bear is $10.68, $1.32 away from $12. The year-end horizon stays. PriceTargets’ high of $35 and low of $15 are a different object.
The spot is Friday’s $14.615 close. The last 90 daily log-returns run from 17 June 2026 through 23 October 2026. Mu is -0.0021342339. Sigma is 0.0311873768. N is 47: five sessions from 26 to 30 October, 20 November weekdays once 26 November is out, and 22 December weekdays once 25 December is out. The 27 November and 24 December early closes count. The square root of 47 is 6.85565460.
Base is the spot times e to the power of N times mu. Forty-seven times -0.0021342339 is -0.10030899. e to that power is about 0.9046, and 0.9046 times $14.615 rounds to $13.22, down $1.395, or 9.54 percent. Bull adds 1.5 times sigma times the square root of N. Sigma times 6.85565460 is 0.21380988, and 1.5 times that is 0.32071482. The exponent is -0.10030899 plus 0.32071482, or 0.22040583. On $14.615 that rounds to $18.22, up $3.605, or 24.67 percent. Bear subtracts one times 0.21380988. The exponent is -0.31411888. On $14.615 that rounds to $10.68, down $3.935, or 26.92 percent. The bull stays above the spot, so the negative drift is kept. It is $2.94 above the 20-day average and $1.48 above the 50-day average. The base is under both averages and under the close. This is not financial advice.
| Case | Price | Versus the $14.615 spot | What has to be true |
|---|---|---|---|
| Bull | $18.22 | Up $3.605, or 24.67 percent | Drift continues and 31 December finishes 1.5 standard deviations above it. |
| Base | $13.22 | Down $1.395, or 9.54 percent | Only the mean continues for 47 sessions, the path down from $32.21. |
| Bear | $10.68 | Down $3.935, or 26.92 percent | The same drift minus one standard deviation, through the 27 October print. |
$18.22, $13.22 and $10.68 are that arithmetic, not Switzer’s $15, not Dan Dolev’s $22, and not the $22.05 average. PriceTargets, as of 5:00 p.m. Eastern on 23 October, had that average on 25 analysts, a high of $35, a low of $15, and a Hold split of 3 sell, 11 hold, 9 buy and 2 strong buy. The average is 50.87 percent above the close. The bear sits $4.32 under the street low of $15. That gap is the drift.
What happens next
Tuesday, 27 October 2026, is the first session that can price the quarter. The release is about 7 a.m. Eastern and the call is at 8 a.m., so that day’s cash close is the first official bar after the numbers. TipRanks puts the street at earnings of $0.17 and revenue of about $1.25 billion. A held guide with a hedge that covers the rate mark would say the negative drift is a multiple story. A pre-hedge hit above $1 billion that the hedge does not cover would make the June CET1 ratio of 18.7 percent the last published floor. Neither result changes $18.22 or $10.68 until a completed close replaces Friday.
The next test is the 20-day average, not the 4.40 percent APY. The close is $0.661 under $15.2757. Closes back through that average, after the call, would say the tape rejected the drift. A close under Friday’s $14.235 low, and then under the 15 October low of $13.86, would say the de-rating is still the path. A SOFIE balance does not, by itself, move that average.
On 31 December the base of $13.22 is the mean continued for 47 sessions. The bull of $18.22 adds 1.5 standard deviations. The bear of $10.68 is one deviation under that drift, not a forecast that deposits of $46 billion unwind. A close back above the 50-day average of $16.7405 would say the drift in this file has stopped. TIKR’s reported coin-flip on a 28 October rate decision is one input. It is not a second set of targets.
Frequently asked questions
Where do the sessions through 31 December 2026 put a SoFi bull and bear, from Friday’s official close?
The bull is $18.22 and the bear is $10.68, with a base of $13.22, from the $14.615 Nasdaq close on 23 October 2026. The path uses 47 sessions, a mean log-return of -0.0021342339 and a sample standard deviation of 0.0311873768. The base is under the close because that mean is negative. SoFi Earn is not an input to the formula.
Has the stock been cut in half from the November 2025 high?
Yes, on this file. The 12 November 2025 close was $32.21 and the intraday high was $32.73. Friday’s $14.615 is 54.63 percent under that close, past halfway at $16.105. The last reported quarter was not a miss on the company’s slides. The cut is the multiple since that November bar, not a third-quarter number that has not been released.
Is the 4.40 percent SoFi Earn yield inside the third-quarter results?
No. SoFi announced SoFi Earn on 22 October 2026, and the quarter being reported ended in September. Results are set for about 7 a.m. Eastern on 27 October. The 20 October rate sheet shows 4.31368 percent interest and a 4.40 percent variable APY, and the balance is outside the SoFi Insured Deposit Program. Conversion into SoFiUSD leaves the deposit. That is not a September revenue line.
Are the bull and the bear Street targets?
No. The $18.22 bull, the $13.22 base and the $10.68 bear are the distribution above. PriceTargets had a $22.05 average, a $35 high and a $15 low as of 23 October. Mizuho’s Dan Dolev was at $22, cut from $29, on 31 July. TipRanks reported Tim Switzer at $15, cut from $16. Those are separate numbers. The July FinanceFeeds pair of $31 and $12 is not reused.
Did Kelli Keough’s sale on 20 October set the close?
No. StockTitan’s account of the Form 4 says she sold 10,203 shares at a weighted average of $14.7226 under a Rule 10b5-1 plan adopted on 30 July 2025, and held 397,870 shares afterward. The sale is about $150,215. Friday’s volume was 38.9 million shares. A plan that old, in that size, is not a view on Tuesday’s guide. The formula spot is the Nasdaq close, not the insider print.
What would make the $13.22 base the wrong year-end price?
A break in the drift. The base only compounds the recent mean for 47 sessions. Closes back above the 20-day average of $15.2757, then the 50-day average of $16.7405, would mean that mean had stopped. A hedge that offsets the rate mark, with the $0.60 guide intact, would be one reason. It would still not make $18.22 a promise or $22.05 the house number.