EPAMNYSEThe short version

EPAM Systems, Inc.

EPAM builds custom software and digital-engineering programs for global enterprises, billing skilled engineers at a markup. Debt-free and cash-generative, its shares have fallen roughly 88% from a 2021 peak as growth stalled.

From a $717 close in November 2021 to about $86 now — a roughly 88% fall, almost all of it the earnings multiple compressing rather than the business shrinking.

$86.38

Share price

~$4.5B

Market capitalization

$5.46B

FY2025 revenue

~$1.3B

Net cash

SwipeScroll▾

The statements

A debt-free engineering franchise whose profit stalled as revenue kept climbing

FY2018 → FY2025as reported · $

Revenue$5.5B+15%

Operating margin9.5%−2.0pp

Net income$378M−17%

EPS$6.72−14%

Free cash flow$613M+16%

Open the full statements →

Consolidated statements, years ended December 31.

Revenue. Sales reached $5.46B in FY2025, up 15.4% — but two 2024 acquisitions and a weaker dollar did much of that; organic constant-currency growth was 4.9%, after a flat 2024 and a decline in 2023.
Earnings. GAAP operating margin fell to 9.5% from a 14.4% peak in 2021, so the company earned less profit in 2025 ($378M) than in 2021 ($482M) on 45% more revenue; the adjusted margin was 15.2%.
Cash. Free cash flow of $613M converted at more than 150% of reported net income on capital spending under 1% of sales, against $1.3B of cash and about $25M of debt.

Cyclical or structural

The stall had a cyclical shape, not the signature of AI substitution

Revenue growth by vertical: stall year (FY2024) vs recovery (FY2025)

FY2024 over FY2023 (stall); FY2025 over FY2024 (recovery).

EPAM's FY2025 10-K now names AI as a risk that 'may reduce demand for our services' and 'negatively impacted the price of our stock', but the stall it points to had a cyclical shape — Business Information & Media fell 10.5% and Consumer, Retail & Travel 5.6% in FY2024 while Life Sciences grew 17.3% —
and the only vertical behaving as AI substitution predicts, Software & Hi-Tech, is EPAM's once-largest and now among its softest at ~15% of FY2025 revenue.
The counter, in the same breath. The demand base is not hollowing out: aggregate delivery headcount rose to about 56,600 and utilization to roughly 77% over the same window — the opposite of AI substitution.

The de-rating

An ~88% fall from the 2021 peak, almost entirely the multiple compressing

Daily close, from the November 2021 peak to mid-July 2026.

Price, not earnings. GAAP diluted EPS was $8.15 in 2021 and $6.72 in 2025, and non-GAAP EPS is higher now than then; the shares fell from a $717 close to about $86 while the earnings held.
What is left. At $86 the stock trades near 13 times trailing GAAP earnings and close to seven times 2026 consensus — the multiples the market assigns to businesses it expects to shrink.
Two forces. The 2022 war disrupted the Central and Eastern European delivery base; and now AI, which management concedes could let clients perform its core service less expensively.

The moat, priced

The growth premium over the scale incumbents has closed to rough parity

EPAM versus the scale incumbents, latest fiscal year

Metric	EPAM	Accenture	Cognizant
Organic revenue growth	4.9%	7.4%	7.0%
Operating margin	9.5% / 15.2% adj	15.6%	16.1%
Voluntary attrition	8.5%	—	15.9%
Revenue per employee	~$87K	—	~$59K

EPAM operating margin shown GAAP / adjusted.

Growth converged. EPAM once grew two to three times faster than the incumbents; in 2025 organic growth of 4.9% ran just below Accenture's 7.4% and Cognizant's 7.0%, and 2026 guidance of 3–6% keeps it in that band.
Stickiness intact. Attrition near 8.5% is roughly half the peer rate, the top five clients are only 13.7% of revenue, and accounts outside the top twenty grew 15.5% — clients are being tested on price, not leaving.
The tell. The whole pure-play engineering cohort re-rated together — the closest listed comparable decelerated from 15.3% revenue growth to 1.6% in a single year — which points at the premium itself, not one firm's execution.

Delivery footprint

The war forced a rebuild: the front-line share of staff fell from 58% to 21%

Delivery professionals by location: 2021 vs 2025

India is now the single largest location; Russia is zero.

Base inverted. EPAM was built on Ukraine, Belarus and Russia — 58% of delivery staff in 2021. After a full Russia exit and relocations under fire, India is now the largest single location and the war-zone share is about 21%.
Contained cost. The direct hit was roughly $42M of one-time 2022 charges against $4.8B of revenue, and total revenue still grew about 28% that year — a one-time shock, not the cause of the later stall.
What it left. A structurally more expensive base — gross margin fell about five points to 28.8% since 2021 — and roughly 12,000 people still working in Ukraine and Belarus, a live geopolitical tail.

The floor

A net-cash balance sheet takes the worst outcome off the table

$1.3B

Cash

~$25M

Total debt

$3.7B

Shareholders' equity

$613M

Free cash flow, FY2025

FY2025 balance sheet and cash flow.

No solvency question. Against $1.3B of cash sits about $25M of finance-lease debt, so roughly $1.3B of net cash — near 28% of the market value — sits inside the equity. For a buyer whose first rule is avoiding bankruptcy, ruin is off the table.
Cash that keeps coming. Free cash flow of $613M has been positive and growing for a decade, on capital spending under 1% of revenue — the business is a payroll wrapped in client relationships, not a capital sink.
One caveat. To fund first-quarter 2026 buybacks EPAM drew $140M on its revolver for the first time — small and revolving, but the first time it has leaned on borrowing to repurchase stock.

Earnings quality

Charge stock comp in full and the owner's yield is nearer 10% than 13.5%

From reported free cash flow to owner cash flow

Owner cash flow	Amount	Yield
Reported free cash flow	$613M	13.5%
Less: stock-based compensation	–$177M	—
Free cash flow to owners	$436M	9.6%

FY2025; yield on ~$4.5B of equity value.

The headline flatters. Free cash flow converting at more than 150% of net income is an artifact: acquisition amortization tripled and depressed reported earnings. Against the profit management guides on, conversion is about 95%.
The real charge. Stock compensation of $177M is a genuine, recurring cost borne by shareholders through dilution; subtract it and free cash flow to owners is closer to $436M — and it is guided up to $202M in 2026.
Clean beneath. Collections are steady at 72 days, interest paid is negligible, and capital spending is trivial — the cash is real; the question is framing, not integrity.

Capital allocation

Every buyback tranche to date was struck well above today's price

Average repurchase price by period vs current market

Buyback averages 2023–1Q2026 against the mid-July 2026 price.

Bought high. About 6.1M shares repurchased across 2023–2025 cost roughly $1.22B; at $86 they are worth about $525M. The buying accelerated as the price fell — the right direction — but the record so far transferred value to sellers.
The base is shrinking. Shares outstanding fell about 9% to 52.4M by March 2026, so repurchases now outrun stock-comp dilution and lift per-share earnings rather than merely absorbing issuance.
Acquisitions, on soft assets. The 2024 deals added about $1.0B, almost all goodwill and intangibles — now $1.6B together, or 44% of equity — closed just as margins compressed. No impairment yet.

Governance

Pay fell with the stock, and the board is shedding its own defenses

Value of $100 invested — EPAM vs peer group

Proxy pay-versus-performance disclosure, total shareholder return.

Real alignment. Executive pay marked to the share price swung from +$26.7M in 2021 to –$13.9M in 2022; a $100 stake in EPAM became $57 by 2025 while the peer group grew to $258. Pay shared the pain.
Not the reason to walk. The board is nine of eleven independent and is voluntarily unwinding its staggered structure by 2028; say-on-pay drew 94% support; and there are no related-party dealings around the founder.
The qualifier. Insiders own just 3.8% — founder Arkadiy Dobkin 2.9% — so alignment leans on at-risk equity, not a large stake; and after the 2025 succession the quarterly calls are led by the new CEO, not the founder.

Margin of safety

Priced below what its own cash flow is worth standing still

No-growth value per share (owner FCF) vs price

At 9% return

$117

At 10% return

$107

At 12% return

$94

At 13% ≈ spot

$88

Owner FCF of ~$436M capitalized flat, plus net cash, over ~52.4M shares.

At $86 EPAM is priced below the value of its own cash flow standing still — owner free cash flow of ~$436m (reported FCF less a full $176.8m stock-comp charge) capitalised flat at 10% is worth ~$107 a share — so the price embeds a permanent ~3%/yr decline in owner cash flow even though FY2025 organic revenue still grew 4.9% in constant currency and management guides FY2026 organic growth to 3-6%.
A thin, real cushion. Run owner cash flow about 20% lower — stock comp rising to $202M and the accrued-liability tailwind reversing — and the no-growth value falls to about $91, essentially today's price; it takes a ~13% required return to justify $86 outright.
The counter, in the same breath. In a genuine structural decline earnings and the multiple compress together with no stable normal, and the net cash is only a partial floor — first-quarter 2026 buybacks were struck at $143.84 against about $86 now.

Three ways it resolves

Modelled downside near 27%, against 74–164% if the stall proves cyclical

Illustrative 2028 scenarios

Scenario	2028E EPS	Exit P/E	Implied price
Bear — structural	$10.50	6x	$63
Base — peer-parity trough	$15.00	10x	$150
Bull — re-rate + reaccelerate	$17.50	13x	$228

Non-GAAP EPS three years out times an exit multiple. Spot $86.

The asymmetry. The bear case models EPS eroding to about $10.50 at a distressed 6x — roughly $63, or 27% below spot. Base and bull run to $150 and $228; the downside is milder because the multiple already sits near a floor.
Most sensitive to the multiple. Value moves most with the exit multiple, itself a proxy for how the cyclical-versus-structural question resolves: at $15 of EPS, each turn of multiple is about $15 of price.
The honest exception. A true AI reset, where earnings and the multiple fall together and there is no normalized level to anchor to — the pattern the pure-play cohort already shows — would break the asymmetry.

The street

The sell-side underwrites a re-rating on cash flow that barely grows

Analyst price targets vs current price

19 analysts, mid-July 2026; eleven buys, eight holds, no sells.

Close to the base case. Nineteen analysts carry a mean target of $139 — about 61% above spot — with no sell ratings, and that mean lands almost on top of EPAM's reported free cash flow capitalized flat at 10%.
A modest ask. The street is underwriting not a heroic growth recovery but a re-rating toward fair value on cash flow that barely grows — the same thing the base case needs: that the stall proves cyclical, not terminal.
Where the safety lives. The margin of safety is genuine at the level of the balance sheet and today's cash flow; it is contingent at the level of the multiple, which will move with a debate the evidence cannot yet settle.

What to watch

A cash-generative, debt-free franchise priced for decline — a cyclical trough set to re-rate, or an AI-driven fade.

01Software & Hi-Tech and the other discretionary verticals resuming organic growth into the second half of 2026 on their own clients — not on acquired revenue or a weaker dollar.
02Revenue per delivery professional holding, versus two or more quarters of organic, mix-adjusted decline as the India mix rises.
03Non-GAAP operating margin holding the 15% guide floor, versus breaking below it.
04Buybacks struck near today's ~$86 rather than the $143.84 average paid in the first quarter of 2026.

Read the full report →

This distills a guided study of EPAM built chapter by chapter — the statements, the demand mix, the delivery rebuild, earnings quality, governance, and what the price implies.

Compiled from the full report · 2026-07-18 · For information, not investment advice.