Self-Help

Self-Help

Fiserv can outlast the problem. Only $1.2 billion of debt falls due in 2026, against an $8.0 billion revolver running to 2030 and covenant room worth roughly 15% of EBITDA. What it has stopped doing is buying its own stock. Quarterly repurchases fell from $2.2 billion at about $223 a share to $200 million at about $61 — spending tracked the price down, not up. Management's stated priority is the leverage ratio.

The balance sheet against the problem's duration

Total debt was $28,997 million at December 31, 2025 — $1,239 million of short-term and current maturities plus $27,758 million long-term [1]. By March 31, 2026 that had edged to $29,182 million — $1,323 million short-term plus $27,859 million long-term [2]. Cash and equivalents were $798 million at year-end and $829 million at the end of March, of which only $335 million was classified as available — the rest sits in settlement advances, regulated entities, cash in transit and joint ventures [3].

memory access out of bounds

Source: FY2025 Annual Report (Form 10-K), Note 12 Debt — annual maturities of total debt [4].

Total principal payments of $29,166 million reconcile to $28,997 million of total debt after $169 million of unamortized discount and deferred financing costs [5]. Two presentation points matter for reading that table. The 2026 line of $1,239 million is entirely foreign lines of credit and current finance-lease obligations [6]; the $2.0 billion 3.200% notes contractually due in July 2026 are classified as long-term and sit in the 2030 column, because the company has the ability to refinance them under the revolver [7]. The same treatment moves $1,165 million of commercial paper into 2030 [8]. On a strictly contractual reading the next twelve months carry about $3.2 billion of paper to roll, not $1.2 billion — all of it refinanceable, but refinanceable is the operative word.

The revolver was replaced in August 2025 with a new $8.0 billion senior unsecured multicurrency facility maturing August 2030 [9]. Undrawn capacity net of borrowings, commercial paper backstop, near-term notes and letters of credit was $4.6 billion at December 2025 [10] and $3.8 billion at March 2026 [11]. The one financial covenant is a leverage test: consolidated indebtedness at each quarter-end no more than 3.75 times consolidated EBITDA as defined, with the company in compliance throughout 2025 and the first quarter of 2026 [12]. Management reported gross leverage below 3.2 times at March 31, 2026 [13]. Holding debt flat, covenant EBITDA would have to fall from roughly $9.1 billion to roughly $7.8 billion — about 15% — before the 3.75x test binds. That is real cushion, though the $9.1 billion base is an adjusted figure: the debt-to-adjusted-EBITDA measure management reports leverage against excludes merger and integration costs, severance costs, One Fiserv transformation expenses and share-based compensation [14]. Measured on reported EBITDA the cushion is narrower.

Refinancing is not free. The paper coming due was issued in a lower-rate world — 3.200% (July 2026), 5.150% and 2.250% (2027), 1.125% euro (2027) — and the replacements are pricing at 4.550% and 5.250% for the August 2025 dollar issue and 3.750% and 4.250% for the June 2026 euro issue [15] [16]. Net interest expense has already gone from $976 million in 2023 to $1,195 million in 2024 to $1,493 million in 2025 — a $517 million increase over two years against operating income of $5,818 million [17].

On the framework's own rule — levered at net debt of three times EBITDA or more — Fiserv sits on the line. fit_features.balance_sheet_class reports unknown, with the reason "debt or cash missing for FY 2025", so the deterministic classification is unavailable; management's own disclosed measure is 3.0 times at December 2025 and below 3.2 times gross at March 2026 [18] [19]. The bar this company faces is therefore either the 10% moderate line or the 25% levered line, and the answer to that is set out in Yield.

The plain answer on outlasting: yes. Nothing in the maturity schedule forces a decision in 2026 or 2027, the facility runs to 2030, the debt is entirely senior unsecured with no secured layer, and the covenant is a single leverage test with room. The constraint is not solvency. It is that management has chosen to spend the cash on the leverage ratio.

The repurchase record — executed, not authorized

Fiserv has been a genuine repurchaser for a decade. Diluted weighted-average shares fell from 683.4 million in FY2020 to 549.0 million in FY2025, a five-year compound rate of −4.3% (fit_features.share_count_trend.cagr_5y_pct), and the trend flag is rising: false. The share count is not being inflated by stock compensation or serial acquisition — the framework's hard-fail condition on this pillar is not met.

memory access out of bounds

Sources: FY2025 10-K Consolidated Statements of Equity for 2023–2025 share counts and amounts [20]; FY2022 10-K equity statement for 2020–2022 [21]; FY2021 10-K equity statement for 2019–2020 [22]. Average price is repurchase spend divided by shares retired, on share counts rounded to millions in the equity statements.

Across FY2019 to FY2025 the company spent $22,987 million retiring roughly 174 million shares, an average of about $132 a share. The stock closed at $55.67 on 29 July 2026 (fit_features.market_cap.price). Those same 174 million shares would cost $9.7 billion today. The FY2025 program alone — 32.2 million shares for $5.6 billion, an average of $174 [23] — was executed at roughly three times the current price. The 4.1 million shares bought back from ValueAct in August 2023 went at $121.98 [24].

The quarterly path is where the framework question is actually decided.

memory access out of bounds

memory access out of bounds

Sources: Q1 FY2026 10-Q equity statement and Item 2 issuer purchases for 1Q25 and 1Q26 [25] [26]; Q2 and Q3 FY2025 earnings calls for 2Q25 and 3Q25 [27] [28]; Q4 FY2025 earnings call and FY2025 10-K Item 5 for 4Q25 [29] [30]. Average price is spend divided by shares repurchased.

In the first quarter of 2025, with the stock near its 3 March 2025 peak of $237.79 (fit_features.capitulation_gauge), Fiserv bought 9.7 million shares for $2.2 billion [31]. In the first quarter of 2026, with the stock in the low sixties, it bought 3.3 million shares for $200 million — no purchases at all in January, 2.5 million shares at $60.96 in February and 763,623 shares at $62.33 in March [32]. Spend fell 91% year on year while the price fell 72%. December 2025 saw no repurchases either [33].

Authorization is not the constraint. The board authorized 60.0 million shares in February 2025, the authorization does not expire, and 42.6 million shares remained available at March 31, 2026 [34]. At $55.67 the unused authorization is worth $2.4 billion of stock — a little over half of one year's free cash flow. The capacity exists and is not being used.

What management has said about the buyback

The buyback question has been asked and answered on every call through the drawdown, and the answer has moved.

In July 2025, with the stock already down but the guidance intact, the CFO raised the target: "We previously expected to return approximately 110% of free cash flow back to shareholders through share repurchase, and we now expect to return approximately 130%, which aligns with the upper end of our targeted leverage range" [35].

In October 2025, at the reset call, the new CEO put it conditionally: "Our capital management strategy hasn't changed; we plan to invest organically and make selective acquisitions, and we'll continue to buy back shares as needed. We're maintaining our leverage guidelines" [36]. Repurchases that quarter were 7 million shares for about $1 billion, and total debt stood at $30.2 billion with leverage at 3.0 times against a 2.5–3.0x target [37].

In February 2026 the ordering became explicit. The CFO: "We also repurchased 3 million shares during the quarter for approximately $200 million and paid down over $1 billion in debt after funding the acquisitions of StoneCastle and a portfolio of TD merchant contracts" — five times as much to debt as to stock, in the quarter the stock traded in the sixties. Buybacks were framed as the residual: "to the extent we generate any excess cash from business and asset optimization activities, we intend to deploy this additional cash to share repurchase" [38].

In May 2026 the CFO said it in one sentence: "We repurchased 3.3 million shares during the quarter for approximately $200 million. As we noted in February, we are focused on managing our leverage ratio and remain committed to returning capital to shareholders" [39].

The May 2026 Investor Day formalised it. The medium-term capital-allocation commitment reads: "Majority of free cash flow directed to stock buybacks while targeting low-end of 2.5-3.0x" [40], with the leverage pillar described as "Remain committed to investment-grade ratings, targeting lower end of leverage range by 2029" and capital return as "Return excess cash to shareholders, through stock buybacks with value-focused deployment" [41]. Free cash flow of over $13.5 billion is targeted across 2027–2029 [42].

Two things are true at once. The majority of free cash flow is promised to buybacks, which is a substantive commitment. And it is promised while moving leverage from 3.0x toward 2.5x on an EBITDA base guided to shrink in 2026 — which means debt reduction has a claim on the same cash, at the moment the repurchase tailwind is largest. That is the framework's stated falsifier — capital allocation pivoting to debt paydown over repurchases — and it is presently observable in the cash-flow record, not merely a risk.

The action taken at the price low was on the debt side. On 16 June 2026, with the stock at its 22 June trough of $47.18, the company launched a cash tender for any and all of its 5.150% notes due 2027 and 4.400% notes due 2049, conditional on receiving proceeds from a new euro notes offering [43]. That offering closed on 23 June at €500 million of 3.750% notes due 2030 and €500 million of 4.250% notes due 2034 [44]. Retiring a discounted 2049 bond is defensible capital allocation. It is not the buyback flywheel.

Insiders did buy, and they bought at the bottom. Between 16 and 17 June 2026 the CFO purchased 10,060 shares at $49.70, the chief administrative and legal officer 10,150 shares at $49.33, and four directors a further 14,571 shares between $48.41 and $50.59 — roughly $1.7 million in aggregate open-market purchases [45]. Genuine, well-timed, and small: $1.7 million against a $30.6 billion market capitalisation.

The levered exception

Where the balance sheet is levered, the framework tolerates the debt only when three things hold together. All three have to compute.

Leg one — a yield of roughly 25% or more. fit_features.adjusted_fcf_yield is not_computable, because the feature file carries no share-based compensation for any year from FY2018 to FY2025 and therefore no complete adjusted-FCF series. The 10-K does disclose the components. FY2025 free cash flow of $4,299 million (operating cash flow $6,062 million less capital expenditures of $1,763 million), less share-based compensation of $357 million, less the trailing five-year average of acquisition payments of $534 million — 2021 $848 million, 2022 $988 million, 2023 $13 million, 2024 nil, 2025 $820 million — gives adjusted free cash flow of $3,408 million [46] [47]. On the market capitalisation of $30,563 million that is 11.2%. Against a 25% levered bar, short by roughly 14 percentage points. This leg fails.

Leg two — a demonstrated multi-year reduction in share count. Diluted weighted-average shares fell 683.4 million to 549.0 million between FY2020 and FY2025, −19.7% in five years. This leg holds.

Leg three — free cash flow relative to revenue not deteriorating. Reported free cash flow was 19.8% of revenue in FY2023, 24.7% in FY2024 and 20.3% in FY2025, on revenue of $19,093 million, $20,456 million and $21,193 million [48] [49]. The FY2025 ratio is below FY2024 by 440 basis points and roughly level with FY2023. Guidance for 2026 is free cash flow conversion of about 90% of adjusted net income on adjusted EPS of $8.00 to $8.30, against $8.64 delivered in 2025 [50]. This leg is contested at best.

One leg of three holds. The levered exception does not compute here, and nothing in this tab should be read as clearing it.

The absurdity check

fit_features.float_retirement_years is not_computable for the same reason the yield is: no adjusted-FCF series. On the filings-based derivation above, the arithmetic is $30,563 million of market capitalisation divided by $3,408 million of adjusted free cash flow — 9.0 years of adjusted free cash flow to retire the entire float at today's price. Using the company's own defined free cash flow of $4,435 million for 2025 [51], less the same compensation and acquisition adjustments, the figure is 8.6 years.

The framework's reference point for a price making a claim that cannot survive is around three years. Nine is a different statement: cheap on cash generation, not absurd. And on the 2026 plan the company will spend a minority of that year's free cash flow on stock, so the practical retirement rate is slower than the arithmetic allows.

Dividend safety

Fiserv has never paid a dividend on its common stock and does not anticipate paying one [52]; the stated policy is to use operating cash flow for capital expenditures, merchant and settlement advances, share repurchases, acquisitions and debt repayment rather than dividends [53]. No part of the return case rests on a dividend, and the coverage test does not apply.

Promise against delivery

The framework treats big claims paired with repeated misses and thin ownership as disqualifying, so the record needs naming rather than characterising. Five material commitments made between February and July 2025, and what happened to each:

memory access out of bounds

Sources: Q4 FY2024 earnings call, 5 Feb 2025, for the 2025 guidance and the fortieth-consecutive-year statement [54] [55] [56]; Q2 FY2025 call for the buyback commitment [57]; Q4 FY2025 call for FY2025 outcomes [58].

The sequence within a single year is the point. On 5 February 2025, guidance was 10–12% organic growth, adjusted EPS of $10.10 to $10.30, and a fortieth consecutive year of double-digit adjusted EPS growth [59] [60], with free cash flow of about $5.5 billion and Clover reaching its $3.5 billion revenue target [61]. On 23 July 2025 organic growth was refined to about 10%, the Clover target reaffirmed, and the EPS floor raised to $10.15 [62] [63]. On 29 October 2025 the same year was reset to 3.5–4% organic growth and $8.50 to $8.60 of adjusted EPS [64]. Three months separated a reaffirmation from a two-thirds cut to the growth rate.

The company's own diagnosis is the most direct evidence on the question. The incoming CEO, on 29 October 2025: "Fiserv's recent results have increasingly relied on short-term initiatives. These initiatives place too much emphasis on pursuing in-quarter results as opposed to building long-term relationships" [65]. And on the underlying growth rate stripped of Argentina: "if we look at the growth rates for 2023, 2024, and year-to-date in 2025, we see 6% growth, 6% growth, and 3% growth" [66] — against the 16% headline organic growth reported for 2024 [67]. A company that had presented a 16% organic growth number was, on its own later analysis, growing at 6%.

Economic ownership is thin. As of 27 February 2026, all current directors and executive officers as a group — fifteen people — beneficially owned 324,298 shares [68], against 533,948,657 shares outstanding [69] — 0.06% of the company, about $18 million at today's price. The then-CEO held 18,388 shares; the former CEO who ran the 2024 guidance cycle held none [70]. Directors are subject to a $1,320,000 holding requirement and were compliant at the June 2025 valuation date [71], so the thinness is a matter of scale rather than policy breach.

Three counter-facts belong in the same paragraph, because they cut against reading this as a settled promotional pattern. First, the reset guidance has been met: FY2025 delivered adjusted EPS of $8.64 against the October range of $8.50–8.60 and free cash flow of $4.44 billion against $4.25 billion guided [72]. Second, the people who made the missed promises are gone — and so, now, are their replacements: Michael Lyons resigned as CEO on 12 June 2026, without severance or accelerated vesting [73], thirteen months after taking the role on 6 May 2025 [74], and President Dhivya Suryadevara resigned for good reason on 7 July 2026 [75]. Third, insiders put their own money in at the low.

What that leaves is a documented pattern of over-promising under the prior management, an explicit company admission that reported growth was flattered by in-quarter initiatives and Argentine inflation, near-zero insider economic ownership, and a management team roughly six weeks old whose promise record is a blank page. The falsifier to watch is dated and specific: whether the majority-of-free-cash-flow buyback commitment made on 14 May 2026 shows up as more than $200 million a quarter in the 2026 filings, or whether the leverage ratio keeps first claim. The drawdown itself is set out in Dislocation; the durability of the cash flows funding any of this is in Durability.