Delivery Footprint

Delivery Footprint

EPAM entered 2022 with more than half of its delivery staff in Ukraine, Belarus and Russia. The invasion forced a full exit from Russia and a relocation of the base: the war-zone share of delivery professionals fell from 58% to 21% by the end of 2025, with India now the largest single location. The direct financial hit was contained — roughly $42 million of one-time charges against $4.8 billion of revenue — but the higher-cost rebuilt footprint sits inside the gross-margin erosion, and about 12,000 people still work in a live conflict zone.

Where a company's engineers physically sit is not usually a first-order investment question. For EPAM it is. The firm was built on a single, unusual delivery base — Central and Eastern Europe rather than India — and in February 2022 that base became the front line of a war. How EPAM absorbed that shock, what it cost, and what the rebuilt footprint means for margins and risk is a distinct thread from the demand debate (AI and Billable Hours) and the moat (Engineering Premium), and it is the most company-specific fact in the case.

A base concentrated in three countries

At the end of 2021 EPAM employed 52,617 delivery professionals. Its three largest delivery locations were Ukraine (12,389), Belarus (9,416) and Russia (8,933) — 30,738 people, or 58% of the delivery base, concentrated in three countries that shared a border and, within months, a war [1]. India, by contrast, held 4,349 and Poland 3,055 [2]. The company's own risk disclosure put it plainly: more than half of its global delivery, administrative and support personnel were located in Ukraine, Belarus and Russia [3].

That concentration was a feature, not an accident. It supplied the deep engineering talent, established technical education and government tax incentives that underwrote EPAM's premium — the same premium the moat chapter traces (Engineering Premium) — at a labor cost below Western Europe and the United States [4]. It also concentrated the firm's operating risk in a region its filings had long flagged for political and economic instability [5].

Sources: delivery-staff mix from FY2021 [6] and FY2025 10-Ks [7]; exit charges from the FY2022 10-K, Note 2 [8].

The 2022 shock, and what it actually cost

EPAM's response was fast and total on the Russia side. On 4 March 2022 it announced it would discontinue services to customers in Russia; on 7 April it began a phased exit of Russian operations; and on 7 September it signed an agreement to sell substantially all its remaining Russian holdings [9]. The accounting consequences landed in 2022 but were modest against the size of the company: asset impairments of $15.1 million (property and equipment), $3.8 million (leases) and $0.7 million (goodwill); $5.1 million of bad-debt expense on Russian receivables; and $17.1 million of employee-separation costs — roughly $42 million in total one-time charges [10]. Alongside, EPAM committed $100 million in humanitarian aid to its people in Ukraine [11].

The lost Russian business was the larger line. Revenue from customers in the CEE geography — mostly Russia — fell $88.8 million, or 52.9%, in 2022, with Russian customers alone accounting for a $90.4 million decline [12]. The Russia reporting segment swung from a $32.5 million operating profit in 2021 to a $13.5 million operating loss in 2022, and shrank from 4.4% of segment revenue to 1.5% on its way to zero [13].

Revenue grew even through the exit. As it wound down Russia and relocated tens of thousands of people out of a war zone, EPAM still grew total 2022 revenue about 28% to $4.82 billion [14]. The growth stall came a year later, in 2023, and the demand chapter attributes it to macro conditions, not the war (AI and Billable Hours). On the evidence, the invasion was a contained, largely one-time shock to a still-growing business — an execution credit for management — rather than the origin of the current slowdown.

Rebuilding the map

The harder task was rebuilding the delivery base. Through 2023 EPAM sustained hiring across Central and Eastern Europe, Central and Western Asia, India and Latin America, and by the end of that year employed delivery professionals in more than 50 countries [15]. The total delivery headcount actually contracted first — from 52,617 in 2021 to about 47,350 by the end of 2023 — as the Russia exit and the 2023 demand slump both bit, before recovering to roughly 56,600 by the end of 2025 [16] [17].

By the end of 2025 the geography had inverted. India was the largest single delivery location at about 12,200 professionals, having added roughly 2,150 during the year; Ukraine held about 8,750 (essentially flat on 2024's 8,764); Poland 5,050; Belarus 3,400; and Mexico 2,950 — with Russia at zero [18]. Management now describes four delivery hubs — Europe, India, Latin America and Western Central Asia — where it had effectively one region before [19].

Source: delivery-staff counts by location, FY2021 10-K [20] and FY2025 10-K [21].

The combined Ukraine-Belarus-Russia share of delivery staff fell from 58% in 2021 to about 21% in 2025, while India rose from 8% to 22%. Central and Eastern Europe remains, in management's words, a "cornerstone" that anchors long-tenured clients, and Ukrainian headcount has stabilized — Q4 2024 marked the first sequential net additions there since the invasion began [22] [23]. The result is a more diversified base than the one that existed in 2021, a durability improvement the war forced rather than one management chose.

What the rebuild left behind

Two residuals matter for the investment case. The first is cost. As early as the Q4 2022 call, an analyst pressed management on exactly this — that "the average cost of delivery has gone up" and asked whether the old Eastern-European cost edge was gone; the CEO's answer pointed to India and Latin America as the fastest-growing, and higher-cost-to-establish, locations [24]. The mechanism shows up in the gross margin: it has fallen from 33.9% in 2021 to 28.8% in 2025, roughly five points, as the base shifted toward higher-cost geographies and relocated Ukrainian staff while the Belarusian and Russian tax-incentivized cost pools shrank.

Source: derived from reported financials, FY2021–FY2025 10-Ks (revenue and cost of revenues).

The footprint is one contributor to that erosion, not the whole story — wage inflation the company could not fully offset through pricing is the other, as the moat and earnings-quality chapters set out (Engineering Premium, Earnings Quality). But the direction is unambiguous: the delivery base that survived the war is structurally more expensive to run than the one that entered it, and it competes on more of the same offshore geography as Cognizant, Infosys and TCS than the old CEE-only model did.

The second residual is a live tail. About 12,000 delivery, administrative and support personnel still work in Ukraine and Belarus — roughly 12,600 at end-2023 and around 12,150 at end-2025 — inside or adjacent to an active war [25] [26]. EPAM owns an office building in Kyiv and leases space in several Ukrainian cities that could be damaged, and its own disclosure warns that any escalation drawing in Belarus's military could jeopardize its people and operations there [27]. This is a lower-probability, higher-severity risk than it was in 2022 — the base is smaller and the firm has proven it can relocate under fire — but it has not gone to zero, and it is not one a peer running an India-and-Latin-America model carries.

The footprint story cuts both ways for the central question of whether EPAM is a cyclical trough or a structural decline. It removes the war as an explanation for the current stall — 2022 grew 28% — and it leaves the company more diversified than before. Against that, it is one reason the old cost-and-engineering distinctiveness has narrowed toward peer terms, and it embeds a geopolitical tail that a value buyer must price into the margin of safety rather than assume away.