Chapter 1

EPAM Systems: the business and its drawdown

EPAM Systems builds software for large enterprises — digital engineering, cloud and AI work delivered by roughly 56,600 engineers, mostly in lower-cost geographies [1]. The business still grows and still converts profit to cash. The stock does not reflect that: at $86, EPAM trades about 87% below its late-2021 peak, near $13 of trailing earnings and under 8x free cash flow, with net cash worth a quarter of its market value. This chapter establishes what the company is, how it earns, and what the collapse has left on the table.

What EPAM does

EPAM describes itself as a global provider of digital engineering, cloud, and AI-enabled transformation services, pairing consulting with more than three decades of software-delivery execution [2]. In plain terms, when a bank, retailer, or drug company needs custom software built, modernized, or run, EPAM supplies the multidisciplinary engineering teams to do it, billing largely for the time of those engineers.

The predecessor business was founded in 1993 and incorporated in Delaware in 2002; it is headquartered in Newtown, Pennsylvania [3]. Founder Arkadiy Dobkin ran it as CEO from the start until September 1, 2025, when Balazs Fejes became President and CEO and Dobkin moved to Executive Chairman [4]. Dobkin remains the largest individual holder, with about 1.5 million shares, close to 2.9% of the company — meaningful founder alignment, though far from control.

Its clients are concentrated in five verticals — financial services; software and hi-tech; consumer goods, retail and travel; life sciences and healthcare; and business information and media — and they tend to stay: 64.4% of 2025 revenue came from clients of five years or more, and no single client is more than a low-double-digit share of the total, with the top five at 13.7% [1]. It competes with the large IT-services names — Accenture, Cognizant, Infosys, TCS, Capgemini — and with closer digital-engineering peers such as Globant, Endava, and Grid Dynamics [5].

How the money is made

The revenue sits in developed markets while the cost of delivering it sits in emerging ones — the labor-arbitrage core of the model. In 2025, the Americas produced 58.7% of revenue and EMEA most of the rest [6], but the engineers are largely elsewhere: India is now the largest delivery location at about 12,200 professionals, and Ukraine remains significant at roughly 8,750 despite the war [1]. The spread between what clients pay in New York or London and what engineers cost in Hyderabad or Kraków is where the margin lives.

Source: FY2025 Annual Report (Form 10-K), MD&A revenues by location [6].

That structure carries two exposures worth stating up front, because both are central to why the stock fell. The first is geographic: after Russia's February 2022 invasion, EPAM had to relocate large parts of a delivery base then concentrated in Ukraine, Belarus, and Russia, and it has since rebuilt around India and elsewhere [7]. The second is structural, and EPAM names it directly: increased adoption of AI-based software tools "may reduce demand for our services," because clients can use large-language-model and agentic tools to build and maintain software themselves rather than buy engineering hours [7]. A business that sells engineering time faces an obvious question when the tool it sells begins to write code.

The drawdown

EPAM was one of the market's fastest-growing IT names on the way up: revenue compounded from about $1.2 billion in 2016 to $3.8 billion in 2021, and the shares closed 2021 near $668, up more than thirtyfold from the 2016 year-end. The re-rating then reversed in stages — the 2022 delivery shock, a demand slowdown through 2023 and the first half of 2024, and, most recently, an accelerating AI-disruption fear that took the stock from roughly $205 at the end of 2025 to $86 by mid-2026.

Source: NYSE daily price history (2026 point is the July 13, 2026 close of $86.38); company filings, as reported.

The distinction that matters for a value investor is whether the earnings collapsed with the price. They did not, at least not to the same degree.

The financials behind the fall

Revenue did stall. After growing 28% in 2022, it fell in 2023 and was roughly flat in 2024 as clients deferred spending, before reaccelerating to $5.46 billion in 2025 — up about 15%, helped by acquisitions [8]. Management attributes the 2023–early-2024 weakness primarily to macroeconomic uncertainty rather than a structural loss of demand [7]. Operating margin, however, has compressed — from around 14% in 2021 to about 9.5% in 2025 — as the company absorbed relocation, hiring in higher-cost geographies, and reinvestment [8].

Source: FY2025 Annual Report (Form 10-K), Consolidated Statements of Income; earlier years from prior 10-Ks, as reported [8].

Net income of $378 million in 2025 is below the $482 million EPAM earned in 2021 on 45% less revenue — a real deterioration in profitability, and the honest counter-point to any "nothing changed" reading. What did not deteriorate is cash generation. Operating cash flow reached $655 million in 2025 and free cash flow was about $613 million, both records, because this is an asset-light business with minimal capital spending [9]. The balance sheet carries $1.30 billion of cash against about $25 million of debt [10]. For a reader whose first fear is bankruptcy, EPAM is close to the opposite case: net cash, no meaningful borrowings, and positive free cash flow every year through the downturn.

Source: FY2025 Annual Report (Form 10-K), Consolidated Statements of Cash Flows [9].

What the price implies

At $86.38, EPAM's equity is worth about $4.9 billion. Strip out the $1.27 billion of net cash — 26% of the market value — and the operating business is priced at roughly $3.6 billion, against $613 million of trailing free cash flow: an enterprise value near 6x that cash flow. On reported (GAAP) earnings the trailing multiple is about 13x; on the consensus non-GAAP earnings analysts actually model, the forward multiple is roughly 7x.

Source: valuation derived from the July 13, 2026 close and FY2025 reported financials [8]; net cash per the FY2025 balance sheet [10]; price per NYSE history.

The sell-side is not modeling collapse. Consensus expects revenue to keep growing modestly — around 5% in 2026 and 6% in 2027 — and non-GAAP earnings to rise to roughly $13 and $14 a share. The mean analyst price target is about $139, some 60% above the current quote; notably, the lowest target in the range, $85, sits essentially at today's price. That configuration — a stock trading at the floor of the analyst range, on a mid-single-digit forward multiple, while estimates still show growth — is the "pessimism already priced in" setup, and it is what makes EPAM worth a careful look rather than a glance.

Source: FY2025 actual per the 10-K [8]; FY2026–FY2027 figures are consensus estimates, as reported.

The question this report answers

EPAM is a fallen star with a clean balance sheet, a still-loyal client base, a founder who built it over thirty years, and a valuation that already assumes the story is broken. The central question this report exists to answer is this: does EPAM's roughly 87% fall from its 2021 peak mark a durable, cash-generative, founder-anchored engineering franchise that the market has overshot to the downside — or the early, rational repricing of a labor-based software-services model that AI and geographic dislocation are structurally impairing? Everything that follows tests one side of that question against the other — the economics of the delivery model, what AI does to billable hours, how management is spending the cash, and how much margin of safety the price genuinely provides.