Back to the tracker
SoFi Technologies logo

SOFI

Position open

SoFi Technologies · Consumer fintech · $23.91B

A digital bank: lending, brokerage and deposits in one app.

The short case

Published positions, quoted as they were argued.

  • Muddy Waters logo
    Muddy WatersPosition openMarch 2026

    Carson Block’s research outfit, known for due-diligence short reports since Sino-Forest in 2011.

    Short report the company publicly rebutted, floating legal action against the firm in response.

    Read the source

Short interest & squeeze model

Short interest

FINRA settlement data

As of 2026-08-14. Fetched 3 Sept 2026, 22:37 UTC.

$3.19B of stock has been borrowed and sold, and has to be bought back. That is the position at today’s price. Everything below is context for it.

Shares short
172.52M
% of float
13.56%
% of shares out
13.36%
Days to cover
3.41

as reported

Float
1271.72M

shares, not dollars

Shares out
1291.57M

shares, not dollars

Avg daily volume
70.60M

stockanalysis-history-3mo

Institutional own.
50.94%

Funds & institutional holders

SEC 13F puts

Named funds disclosing a PUT position against this ticker in their most recent SEC 13F-HR filings, matched on CUSIP rather than issuer name (a multi-filer filing can carry an unrelated company's name over the right CUSIP) and attributed to the filer named in that filing's own primary document. This is a quarterly snapshot filed up to 45 days late, so the position may already be gone. A put can be a hedge on a long position, not a directional bearish bet. And 13F excludes short stock outright, which is how most short interest is actually expressed, so this is a small, self-selected slice of the bearish side shown alongside the aggregate short interest above, never an explanation of it. Absence from this list is not evidence a fund is not short.

FundUnderlying sharesValuePeriod endingFiling
MIRAE ASSET SECURITIES (USA) INC.250.0K$4.0M2026-03-3113F filing
MIZUHO MARKETS AMERICAS LLC1.25M$2.8M2026-06-3013F filing
Volterra Technologies LP37.4K$670.6K2026-06-3013F filing

Call holders

for reference, not the short story
FundUnderlying sharesValuePeriod endingFiling
Longitude (Cayman) Ltd.1.95M$35.0M2026-06-3013F filing
MIZUHO MARKETS AMERICAS LLC1.25M$6.3M2026-06-3013F filing
Volterra Technologies LP42.6K$763.8K2026-06-3013F filing
IFS Group,LLC3.0K$47.6K2026-03-3113F filing
Mutual Advisors, LLC100$1.8K2026-06-3013F filing

Scanned 20 of the most recent 13F-HR filings mentioning this ticker (cap 20), filed between 2026-02-15 and 2026-09-03. Fetched 3 Sept 2026, 22:37 UTC.

What a squeeze costs

Model, not a prediction

A size check, not a forecast: how much net buying it would plausibly take to move the price by a given amount, using the square-root market-impact law ΔP/P = Y·σ·√(Q/ADV) (Almgren et al. 2005; consistent with Kyle's lambda). Y = 1 here, a commonly-cited value for the impact coefficient, exposed as a named constant rather than baked in. σ is trailing daily volatility over the last 65 trading days; ADV is average daily volume. None of this knows anything about actual short entry prices, margin calls, or borrow availability. It is a mechanical answer to "how big a trade moves this how far", nothing more.

+25%
$66.43B

beyond the model’s range: asks for 3589M shares, more than exist

+50%
$265.72B

beyond the model’s range: asks for 14355M shares, more than exist

+100%
$1.06T

beyond the model’s range: asks for 57421M shares, more than exist

+200%
$4.25T

beyond the model’s range: asks for 229684M shares, more than exist

The square-root law is fitted to trades that are a fraction of one day’s volume. Extrapolated to the larger moves above it asks for more shares than have ever been issued, which is arithmetic rather than a market outcome. Those targets are marked, and should be read as “further than this model can see” rather than as a price.

Squeeze trigger (20% move)
$42.51B

capital assumed to put the average short meaningfully underwater, a stated assumption rather than an observed cost basis

Cover-cascade target
$19.52

if the entire short position ($3.19B) had to cover at once, modelled as +5.5%

Days of full volume
34.4

trigger capital ÷ average daily dollar volume, i.e. how many days of the entire market's volume that is

Options & gamma exposure

Options & gamma exposure

CBOE, delayed

Feed quote time 2026-09-03 22:36:23. This is delayed data, not a live tape. Cached 3 Sept 2026, 22:37 UTC. Spot used for these figures: $18.48.

Gamma exposure is an intraday, mechanical hedging flow: how many shares a delta-hedged options book would need to trade for a 1% move in the underlying, computed from today's open interest and the venue's own per-contract greeks (no Black-Scholes re-derivation here). It is a different mechanism from the short-interest days-to-cover model above, which plays out over days on a figure that only updates a couple of times a month. Treat these as two separate stories, not one combined "squeeze" number.

Every "net" figure below assumes dealers are net short the calls and net long the puts their customers hold. That is a common convention (SqueezeMetrics' GEX methodology; the assumption most retail options-flow trackers use), which makes call gamma an accelerant on the way up (+1) and put gamma an offsetting term (-1). This is an assumption about who is on the other side of the open interest, not a disclosure. Nobody publishes actual dealer books. The raw, unsigned call and put totals are shown alongside the net figure so a reader who rejects the assumption still has usable numbers. Open interest also names nobody: unlike the 13F table above, options data cannot show who wrote a contract at all (13F itself never covers short stock either, see the caveats on that table).

Call gamma / 1%
3.26M sh

$60.3M

Put gamma / 1%
1.45M sh

$26.8M

Net gamma / 1%
1.81M sh

$33.5M, assumed sign convention

Call gamma vs ADV
4.62%

call gamma/1% ÷ average daily volume

Call open interest
2.68M
Put open interest
1.33M
Put / call ratio
0.49
Chain size
978 contracts

19 expiries

Where the gamma sits

Model, not a prediction

Strikes above spot carrying the most call gamma. These are the levels where dealer hedging, under the assumption above, would plausibly accelerate a move on the way up.

$20.00
550.4K sh/1%
$19.00
504.5K sh/1%
$18.50
369.5K sh/1%
$22.00
295.4K sh/1%
$25.00
224.1K sh/1%

Gamma-flip strike: $18.53

The strike where cumulative net gamma (summed from the lowest strike up, under the same sign convention) crosses zero, if the chain has one. This is a fragile marker, not a level: it depends on every assumption above, moves with which strikes and expiries happen to sit in the chain, and can shift meaningfully between one feed pull and the next on thin or stale quotes near the crossing. Read it as a rough boundary, not a price target.

How the pair is supposed to work

Step 1

🚀

The token runs

A meme paired to the stock catches a bid and brings people with it.

Step 2

🔗

The arb passes it on

The tokenized share is arbitraged against the listed one, so the buying lands there.

Step 3

📈

SOFI follows

The smaller the company, the further the same money pushes the price.

Step 4

🩸

Shorts feel it

A crowded short gets more expensive to hold the higher it goes.

Steps one and two are the parts anyone here controls. Steps three and four are the bet. A meme token has no claim on the stock, the two prices can drift apart for as long as they like, and a short position can stay open longer than a group chat stays interested.