Dislocation
Bottom line
Fiserv fell 80.2% from a $237.79 close on 3 March 2025 to $47.18 on 22 June 2026, and trades at $55.67. Five single trading days carried a compounded 70.7 points of that fall. The largest, 29 October 2025, took the stock down 44.0% on 41x normal volume as management cut 2025 adjusted EPS guidance from $10.15–$10.30 to $8.50–$8.60 [1]. Traded volume peaked at 7.39x the pre-fall median.
The drawdown
Source: fit_features.capitulation_gauge.drawdown, derived from the daily price series; 476 is calendar days, equal to 327 trading days. Peak and trough are closing prices.
The fall ran 16 months and did not arrive as one shock. Six weeks of quiet drift preceded the first event; four dated events then delivered the bulk of the decline; a fifth, eight months later, made the low. From the 22 June 2026 trough the stock has recovered 18.0%, so the current price is not the point of maximum fear — it is 76.6% below the peak and 18.0% above the bottom.
memory access out of bounds
Source: daily closing prices from the run price series; month-end values. The 29 October 2025 guidance reset sits inside the October 2025 point [2].
The trigger
The identifiable, dated event is the third-quarter 2025 release of 29 October 2025. Fiserv reported adjusted EPS of $2.04 against a consensus of $2.65 and revenue of $4,919 million against $5,349 million, then reset the year: organic revenue growth of 3.5% to 4% and adjusted EPS of $8.50 to $8.60, from approximately 10% and $10.15 to $10.30 three months earlier [3] [4]. On the call, the CEO said that "Fiserv's growth and margin targets need to be reset," attributing the change to four factors: slowing cyclical growth in Argentina, the recalibration of optimistic growth assumptions in the original guidance, the impact of deferred investments, and the deprioritisation of short-term revenue and expense initiatives. He also said the analysis found "nothing at Fiserv is fundamentally broken" [5]. The same release announced two co-presidents, a new CFO effective 31 October, and three new directors [6].
The market reaction that day was the largest one-day decline in Fiserv's listed history back to 1990: minus 44.0%, on 103.5 million shares. The next day added minus 7.7% on 63.8 million.
memory access out of bounds
Sources: closes and volumes from the daily price series; volume multiples against the 2.50 million-share median of the 180 days before the peak. Events per the Q1 2025 release [7], the Q2 2025 release [8], the Q3 2025 release [9] and the 8-K reporting the CEO change [10]. The 15 May 2025 conference remarks were not filed on an 8-K and are attributed to contemporaneous press reports.
Compounded, those five days multiply to a 70.7% decline: $237.79 becomes $69.74. Every other trading day in the 16 months — 349 of them — accounts for the remaining minus 20.2% that takes the price to $55.67. Three of the six worst days in the company's entire price history back to 1990 fall inside this drawdown.
Two of the five were not earnings releases, and one of them rests on no filing at all. On 15 May 2025 the then-CFO said at an investor conference that Clover volume growth would be broadly similar to the first quarter, where Clover revenue had grown 27% on annualised payment volume growth of 8% [11]; the stock fell 16.2% that day. On 15 June 2026 the company disclosed that the CEO of 13 months had resigned on 12 June, with a successor appointed two days later [12]; the company's leadership-transition release the same day said he was leaving to become chief executive of another company. The stock fell 10.9% and made its low a week later.
Drift before the event
The first move down was drift, not a repricing. From the 3 March 2025 peak to 23 April 2025 the stock fell 8.7%, on median daily volume of 3.40 million shares — 1.36x the pre-peak median. That span includes the 4 April 2025 market-wide decline of 8.4%, which was not company-specific. Nothing in the corpus dates a Fiserv disclosure to that window, and the company affirmed its 10% to 12% organic growth and $10.10 to $10.30 adjusted EPS outlook on 24 April [13] — the same outlook it had set on 5 February [14]. The 18.5% fall on 24 April came with that affirmation intact: what changed was the composition of growth, not the headline number. The pattern repeated in July, when Fiserv reported adjusted EPS of $2.47 against a $2.43 consensus, raised the low end of its EPS range to $10.15, and still fell 13.8% — Clover revenue grew 30% on reported volume growth of 8%, 11% excluding the prior-year gateway conversion, and management promised a second-half acceleration [15] [16].
So the event leg separates cleanly from the drift: an 8.7% orderly slide on normal volume, then four disclosure-driven days on 6.8x to 41.4x volume that did the work.
The fear gauge
The measured spike is 7.39x: the maximum 20-day average volume in the peak-to-trough leg divided by the median daily volume of the 180 days before the peak. That maximum window ended 25 November 2025 — it is the four weeks that begin with the 29 October reset. Inside it, 369.2 million shares changed hands, equal to 67.2% of the 549 million shares outstanding, in 20 trading days. The single day of 29 October traded 41.4x the pre-peak median and stands as the largest volume anomaly in Fiserv's price history relative to its own 50-day average.
memory access out of bounds
Source: daily volumes from the run price series, divided by the 2,496,382-share median of the 180 calendar days before the 3 March 2025 peak; the same denominator used by fit_features.capitulation_gauge.volume_spike. July 2026 excludes a partial final session.
Two readings follow, and they pull in different directions. The spike is genuinely emotional in shape: it arrived on a single disclosure, at a point where the stock had already fallen 47% from the peak, and it turned two-thirds of the register in a month — the profile of holders exiting on the news rather than on the arithmetic. Against that, volume has never gone back to normal. Every month since April 2025 has run between 1.7x and 4.9x the pre-fall median, and the most recent month is still 2.4x. A capitulation that exhausts itself leaves quiet behind it; 15 months of doubled turnover reads more like a continuing transfer of the register than a single flush. The June 2026 low came on a second, smaller spike of 10.7x, eight months after the first.
Who was selling
The disclosed record identifies one large seller, two large buyers, and no forced seller.
memory access out of bounds
Sources: the 2025 proxy statement ownership table as of 28 February 2025 [17]; the 2026 proxy statement ownership table as of 27 February 2026 [18]; Schedule 13G/A filings by T. Rowe Price dated 14 August 2025 and by Dodge and Cox dated 13 November 2025 and 13 February 2026, staged under data/ownership_by_owner/.
The seller. T. Rowe Price held 35,160,318 shares, 6.3% of the class, at 28 February 2025 [19]. Its Schedule 13G/A reporting 30 June 2025 shows 18,707,183 shares, 3.4% — 16.5 million shares sold, 46.8% of the position, in four months. That selling happened during the April and May legs and before the October reset, and because it crossed below 5% the firm has no further disclosure obligation, so the eventual size of the exit is unknown. It does not appear in the 2026 proxy's 5% table [20].
The buyers. Dodge and Cox went from 34,866,570 shares at 28 February 2025 to 38,634,009 at 30 September 2025 and 49,474,622 at 31 December 2025 — an addition of 10.8 million shares, 28.1% of the position, in the quarter containing the crash. Vanguard rose from 50,085,002 shares to 63,408,549 over the year; some of that percentage gain is mechanical, since shares outstanding fell from 558,574,211 at the 17 March 2025 record date [21] to 533,948,657 at 27 February 2026 [22]. BlackRock's disclosed holding is the identical 39,890,598 shares in both proxies, which means the same underlying report was carried forward; it carries no information about flows.
Insiders. Two executives sold into the top: the COO sold 45,000 shares for $10.6 million on 20 February 2025 at $235 to $236, and the head of the financial institutions group sold 13,560 shares for $3.2 million at $238.15 on 3 March 2025, the peak day itself. No insider has made an open-market sale since 28 May 2025. Buying then ran the other way, in three clusters: a director bought $651,800 at $65.18 on 30 October 2025, the day after the crash; the new CFO bought $1,060,970 at $62.41 on 1 December and the chief legal officer $499,201 at $63.19 on 2 December 2025; and six insiders including the CFO bought $1,723,401 between $48.41 and $50.59 on 16 and 17 June 2026, days after the CEO resignation and days before the low. The total is $3.94 million bought against $13.8 million sold before the fall — small against a $30.6 billion market value, but directionally unambiguous and made at prices that have since held.
Forced and structural sellers. None is documented. The exchange move was voluntary: on 28 October 2025 Fiserv notified the NYSE that it would withdraw its listing and transfer to Nasdaq effective on or about 11 November 2025 [23], which does not affect its S&P 500 membership and produces no index-mandated selling. No index deletion, fund liquidation or margin event appears in the corpus or in the web record. Short interest is not a factor on any available read: the official FINRA series returned no rows for this ticker, and third-party trackers put the position near 20.6 million shares, 3.9% of shares outstanding, against 23.2 million the prior month — a level that neither drove the fall nor sets up a squeeze. An activist, Jana Partners, disclosed a roughly 2.2 million-share position in February 2026, about 0.4% of the register.
Set against the buyers is a live overhang. Three federal securities class actions are pending, the first covering purchasers from 22 July 2024 to 24 July 2025, with lead plaintiffs appointed in November 2025, the others covering 23 July 2025 to at latest 29 October 2025; all allege that statements about Clover's growth were false or misleading [24]. That is a reason for some institutional holders to stay out that has nothing to do with the operating arithmetic.
Estimates against price
The forward earnings base and the price did not move together. Measured from the peak, the near-term guidance fell 16.2% while the price fell 70.3%.
memory access out of bounds
Sources: the guidance midpoints as published on 5 February 2025 [25], 24 April 2025 [26], 23 July 2025 [27], 29 October 2025 [28] and 10 February 2026 [29]; the final bar is the 2026 guidance midpoint. Closes from the daily price series.
The sequence has three phases.
Between 3 March and 23 July 2025 the price fell 39.9%, from $237.79 to $143.00, while the company's own adjusted EPS guidance midpoint went from $10.20 to $10.225 — it rose. The price moved roughly six months and 40 points ahead of any published cut.
On 29 October 2025 the two converged in a single session. Consensus for the quarter had stood at $2.65 of adjusted EPS and $5,349 million of revenue; the reported figures were $2.04 and $4,919 million, misses of 22.9% and 8.0%. The full-year guidance midpoint moved from $10.225 to $8.55, a cut of 16.4%; the stock fell 44.0% the same day. That is the only day on which the estimate cut and the price cut arrived together.
Since then the price has fallen again with the estimates roughly still. Over the six months from 31 January 2026, consensus normalised EPS for FY2027 moved from $9.15 to $8.93, minus 2.3%, and for FY2028 from $9.75 to $10.02, plus 2.8%; consensus revenue for both years moved less than 1%. The price over the same window fell from $63.73 to $55.67, minus 12.6%, touching minus 26.0% at the June low.
Put through the multiple, the whole 16 months decomposes exactly. At the peak, $237.79 against the then-guided FY2025 adjusted EPS midpoint of $10.20 is 23.3 times. Today, $55.67 against the FY2026 consensus of $8.12 is 6.9 times. The forward earnings base is down 20.4%; the multiple is down 70.6%; and 0.796 times 0.294 is 0.234, which is the 76.6% peak-to-current fall. Roughly four-fifths of the decline is re-rating, one-fifth is the earnings cut itself.
The strongest fact against reading that as anchoring to a one-year number is the shape of the consensus curve beyond it. This was not framed as a single bad year: the medium-term organic growth target of 9% to 12% was withdrawn with the guidance [30], FY2025 adjusted EPS came in at $8.64 against $8.80 the prior year — a decline, not a pause [31] — and 2026 was guided to 1% to 3% organic growth and $8.00 to $8.30 of adjusted EPS [32], a second consecutive decline. Consensus normalised EPS does not regain the $10.20 that was guided for FY2025 by FY2028, where it sits at $10.02. The sell side is modelling three-plus years to get back to the starting line, not one.
Consensus forward free cash flow tells a steadier story than consensus EPS. On the current market value of $30.6 billion, the consensus mean stands at $3.88 billion for FY2026, $4.28 billion for FY2027 and $4.68 billion for FY2028 — rising, not falling, across the forecast horizon. What those figures imply against the framework's yield reference lines belongs to Yield; what the earnings cut did to intrinsic value, and whether it is temporary, belongs to Damage Math.
What this does not settle
The drawdown is fully dated and the trigger is a filed document, but three things sit outside what the record here can answer.
Reported short interest is unavailable from the official source for this ticker — the FINRA feed staged for this run returned no position rows, plausibly a consequence of the November 2025 ticker change from FI to FISV — so no through-the-fall series can be shown, only a current third-party level near 3.9% of shares outstanding, and two trackers disagree on it by a third. Institutional flow is visible only at the 5% disclosure threshold and at proxy dates, which means the middle of the register, where most of the 2.5 billion shares traded since the peak actually moved, is not observable. And one of the five decisive days rests on unfiled oral remarks at an investor conference rather than on a document, so the mechanism of that leg is attributed to contemporaneous reporting rather than to a primary filing.