Chapter 4

Buybacks and Debt

Fiserv has never paid a dividend; its entire return of capital is share repurchase [1]. Since 2020 it has retired about a fifth of its shares, and in the last three years the pace ran ahead of free cash flow, funded partly by debt that now totals roughly $29 billion and costs $1.5 billion a year in interest [2] [3]. Most of that stock was bought at $120–$227 a share; it now trades near $52.

A share count that keeps shrinking

The buyback is real and it is large. Fiserv holds 784 million common shares issued, and the share it keeps in treasury — bought back and set aside — has climbed from 121 million at the end of 2020 to 250 million by the end of 2025 [4] [5]. Net shares outstanding have fallen from 668 million to 534 million — a 20% reduction in five years.

Loading...

Source: shares issued less treasury shares, Consolidated Statements of Changes in Equity, FY2022 10-K [6] and FY2025 10-K [7].

That shrinkage does mechanical work on per-share numbers. In 2025 net income attributable to Fiserv rose 11%, to $3,480 million, but diluted earnings per share rose 18%, from $5.38 to $6.34, because the diluted share count fell about 6% [8]. Roughly seven points of the 2025 EPS growth — more than a third of it — came from the buyback rather than from the business. With the adjusted-earnings and cash-conversion question already on the table (Earnings Quality), it is worth knowing how much of the reported per-share progress is arithmetic.

The pace outran free cash flow

The company converts most of its earnings to cash, but in 2023–2025 it spent more on buybacks alone than it generated in free cash flow, before a single acquisition.

Loading...

Source: Consolidated Statements of Cash Flows, FY2022 10-K and FY2025 10-K; buybacks are purchases of treasury stock including shares withheld for taxes [9]; acquisitions are payments for businesses, net of cash acquired.

In 2025, free cash flow was about $4.3 billion. Against that, Fiserv spent $5.9 billion buying back stock, $0.8 billion on eight bolt-on acquisitions (StoneCastle, CCV, Payfare, TD Merchant Canada, CardFree and three others) [10], and a further $0.4 billion buying in minority stakes [11]. The roughly $2.9 billion gap was bridged with debt: net long-term borrowings rose about $4 billion over the year, and the $2.0 billion of senior notes issued in August 2025 were earmarked, in the company's own words, for "the repayment of a portion of the Company's commercial paper notes and for share repurchases" [12].

Total debt reached roughly $29 billion at the end of 2025 — $27.8 billion long-term plus $1.2 billion short-term — against $0.8 billion of cash, for net debt near $28 billion [13] [14]. The cost of carrying it is now visible in the income statement: net interest expense rose from $976 million in 2023 to $1,195 million in 2024 to $1,493 million in 2025 — up 25% in the last year as new notes were issued at 4.55%–5.25% [15] [16]. Interest now absorbs about 26% of operating income, up from roughly 19% two years earlier.

Shares Retired Since 2020

20%

Net Debt ($B)

$28.2

Interest Expense 2025 ($B)

$1.5

Dividend per Share

$0.00

Source: derived from FY2025 10-K — equity statement [17], debt note [18], interest note [19], and dividend policy [20].

What it paid, and when

The harder question is not how much stock Fiserv bought, but at what price. The average paid climbed steadily as the shares rose, peaking as the multiple did — and then the buyback shut off just as the stock collapsed.

Loading...

Source: purchases of treasury stock divided by shares repurchased, Consolidated Statements of Changes in Equity, FY2022 10-K [21] and FY2025 10-K [22]; Q1 2026 from the Q1 2026 10-Q [23].

The clearest read is the quarterly one. In the first quarter of 2025, Fiserv repurchased 9.7 million shares for $2.2 billion — about $227 a share. In the first quarter of 2026, after the October 2025 de-rating, it bought 3.3 million shares for $200 million — about $61 a share [24]. The company spent eleven times as much buying its stock near the top as it did after the price had fallen more than 70%. Across 2023–2025 it deployed roughly $16 billion at an average near $150 a share; those same shares trade at $52.33 as of early July 2026.

There is a fair counter-case, and it rests on the same leverage number. Management frames repurchases as opportunistic within an investment-grade guardrail, and slowing purchases while leverage is at the high end of the 2.5–3.0x range is defensible balance-sheet discipline rather than simple mistiming [25]. A new leadership team inheriting a guidance reset (What Fiserv Is) has reason to preserve capital. And the buyback did permanently retire a fifth of the shares — that accretion is banked, whatever the stock does next. The record still shows the company was a heavy buyer at $160–$227 and a light one at $61, which is the opposite of buying value.

The plan from here

At its May 2026 Investor Day, Fiserv committed to a "disciplined approach" in which excess cash is returned "through stock buybacks with value-focused deployment," a "majority" of future free cash flow is directed to repurchases, and leverage is steered toward the low end of the 2.5–3.0x range by 2029 [26] [27]. The plan is coherent: it keeps the buyback as the primary lever while promising to de-lever. Its two frictions are that the same free cash flow is being asked to both repay debt and repurchase shares, and that the balance sheet enters that plan carrying near-$28 billion of net debt and only $0.8 billion of cash [28].

On the evidence, Fiserv's capital allocation has compounded share count effectively but managed price and leverage poorly through the 2024–2025 peak; the single most useful thing to watch is whether the 2026 repurchase pace re-accelerates at the depressed price while leverage still falls. If both happen, the "value-focused" language is being honored. If instead buybacks stay muted and leverage drifts above 3.0x — or high-multiple M&A resumes — the concern that the return-of-capital engine is debt-financed engineering on a decelerating base gets harder to answer.