Initiated 2026-10-20 · Price $129.68 (intraday 2026-10-20 10:00am ET, stockanalysis.com) · Mkt cap $6.78bn (52.27mm diluted shares) · Information technology / IT services · Model: verified

Rating: Hold — Conviction: High

EPAM is cheap on every lens: ~9.3x model-basis EBITDA, ~9.9x the FY2026 non-GAAP EPS guide and a 7.3% FCF yield, with net cash. But the price already sits close to a Base case of low-to-mid-single-digit growth. Organic growth is 2–3%, and the debate over whether AI shrinks billable hours faster than it creates new work is unresolved. Neither Globant's stall nor Accenture's steadier prints settle it. The market is paying for that Base, not for a recovery.

Conviction tests (3a-v-c): T1 pass (Base −0.7% at +1pp, +13.3% at −1pp; Hold at both) · T2 pass (Gordon alone −0.4%, exit-multiple alone +11.7%; Hold on both) · T3 pass (every FY0 fact and historical series from EDGAR XBRL, the 10-K balance sheets or the Q2 2026 release; no est. inputs) · T4 pass (Q3 print on 2026-11-05 is a guidance event inside the cases; no pending deal, ruling or contract binary)

Business overview

EPAM sells software engineering, digital product, data and AI services, billed mostly by time and materials. Its customers are large enterprises in financial services, software and hi-tech, media, life sciences, retail and travel. FY2025 revenue was $5.46bn, up 15.4%, partly from acquisitions made in 2024. Financial services was ~24% of revenue in Q3 2025 (Zacks). The Americas were ~59% and EMEA ~39% in Q2 2025 (Fintool; both aggregator figures, not filing-checked). At June 2026 headcount was ~62,850, including ~56,650 delivery professionals, up just 0.3% in the quarter. Three variables drive earnings:

• billable headcount and utilization;
• the price of a unit of work, as AI tooling changes how many hours a project needs;
• wage inflation in the Central and Eastern European, Indian and Latin American delivery centers against client bill rates.

Competition

Three competitors matter for the growth driver. Accenture grew FY2026 (August year-end) revenue 5% in local currency, Q4 +7%, and guided FY2027 to +3–6% (release, 2026-10-01). Globant's Q2 2026 revenue was $614.4mm, roughly flat on the year; it cut FY2026 guidance to $2.428–2.462bn, guided Q3 to −1.6% to −0.3%, and is moving to output-based pricing (call, 2026-08-14). Cognizant competes for the same enterprise budgets; its source page shows only a 5.0% three-year consensus revenue forecast. EPAM's own organic constant-currency growth of +3.4% in Q2 and the +2–3% FY2026 guide sit between Globant and Accenture. Pressure shows up first in pricing per deliverable and in flat delivery headcount, not in lost logos. Cognizant, the closest listed comparable by multiple, trades at 7.7x aggregator EV/EBITDA and 11.2x forward P/E, against EPAM's 7.9x and 9.5x (stockanalysis.com, 2026-10-20).

Bull case

1. AI-native work re-accelerates growth. Projects rebuilding client systems around AI add more revenue than AI-assisted coding removes, and growth reaches 7–8% by FY2027–29. The cue for this was Q2's beat, which came "with continued AI-native momentum" (Q2 release). Plays out if organic cc growth beats the top of the 2–3% FY2026 guide and FY2027 guides higher. Model: rev_growth
2. Profitability program keeps lifting margin. Q2 GAAP operating margin was 10.8%, up from 9.3%. Higher utilization and AI-assisted delivery take model-basis EBITDA margin from ~13% to ~15% by FY2030. Plays out if GAAP operating margin holds at or above 11% through FY2027. Model: ebitda_margin
3. Re-rating toward Accenture. A growth recovery earns 11x exit EBITDA, Accenture's 10.5x aggregator multiple. Model: exit_ev_ebitda

Bear case

1. AI deflates the billable-hours model. Clients get the same output from fewer engineer-hours, and new AI work doesn't fill the gap. Revenue stalls in FY2027 and falls 3% in FY2028, which is Globant's 2026 path one year later. Plays out if the FY2027 guide is at or below 0% organic or delivery headcount starts falling. Model: rev_growth
2. Price concessions compress margin. Output-based pricing and reinvestment take margin to 11.5%, below the FY2025 trough of 11.8%. Plays out if GAAP operating margin falls back below 10%. Model: ebitda_margin
3. Structural de-rating. The market prices EPAM like a shrinking labor business: a 6.5x exit (between Globant's 5.3x and Cognizant's 7.7x) and a 1pp higher discount rate. Model: exit_ev_ebitda, wacc

Valuation & balance sheet

Model-implied value range (from model-summary.json; generic module, Gordon-growth DCF and exit-EV/EBITDA DCF, midpoints): Bear $86.41 · Base $137.03 · Bull $185.90 per share, i.e. implied returns of −33.4% / +5.7% / +43.4% vs $129.68. These ranges show how the bull and bear drivers translate into value; they are not price targets.

• What's priced in. The price sits ~85% of the way from Bear to Base, so the market already pays for roughly the Base path and little of the bull. A single-digit Base return is why this is a Hold and not a Buy, cheap multiples notwithstanding.
• Re-rating. Base's 9.0x exit is a ~3.5% de-rating from today's 9.3x model-basis multiple. The methods differ by 12% (Gordon $129.17, exit $144.89). At a 12.1% discount rate and 3% terminal growth, the Gordon method embeds a lower terminal multiple than 9x, so the exit method is the more generous read. No tail run: no client or product is ≥50% of revenue, and Bear is a decline, not a plateau.

Balance sheet: not meaningful. Net cash $764mm (~$14.60/sh, −1.19x FY2025 EBITDA); $25mm of debt; operating leases $127mm, excluded from EV. Cash fell from $1.30bn at December 2025 to $0.79bn at June 2026. Over that half, $409mm of buybacks ran against −$38.8mm of operating cash flow. Ratings: n/a (unverified).

Model note: tier full, status built, verification verified (3,207 formula cells matched in LibreOffice). No unverified inputs and no assumptions without a basis. Scenario consistency is OK on all 14 rows.

• Sourcing. Revenue, operating income, D&A and capex come from EDGAR XBRL 10-K annual values. Receivables-and-contract-assets and payables come from the FY2023 and FY2025 10-K balance sheets: two EDGAR concept fetches for receivables returned 404, which is why history starts at FY2022. Cash, debt and shares come from the Q2 2026 release.
• EBITDA basis. GAAP, so stock-based compensation and acquisition amortization stay in as costs. That is well below company non-GAAP: Q2 operating margin 16.4% non-GAAP vs 10.8% GAAP.
• EPS vs guidance. Base Y1 EPS $8.43 is 1.6% above the $8.30 GAAP guide midpoint and far below the $13.16 non-GAAP midpoint, which excludes those costs.
• WACC. 12.1% = CAPM (10Y 5.15% + beta 1.39 × 5% ERP). The company is net cash, so WACC equals the cost of equity. Bull −1pp, bear +1pp.
• Placeholders. The revolver's coupon and maturity are placeholders on an immaterial $25mm balance.

Scenario stress test

Reasoned from the bull/bear drivers above. The model column comes from the scenario overlays (Base case + shock).

Currently active/on watch per the playbook: S3 partially active (price-level leg only); S8, S10 and S11 on watch.

What would change the call

Upgrades if: the Q3 (2026-11-05) or Q4 print shows organic cc growth above the 2–3% FY2026 guide and delivery headcount growing again, and the FY2027 guide calls for mid-single-digit or better organic growth; or the stock de-rates toward the Bear case with no deterioration in the data. Downgrades if: the FY2027 guide is at or below 0% organic, delivery headcount falls for two quarters, GAAP operating margin drops back below 10%, or buybacks keep running ahead of operating cash flow.

Watch items

• W1: Q3 organic constant-currency growth — at or above the ~1.8% guide midpoint, with the FY2026 +2–3% organic guide held — Q3 2026 release — 2026-11-05 — Model: rev_growth
• W2: Delivery professionals — growing vs ~56,650 at June 2026 (Base) or falling (Bear) — Q3 2026 release — 2026-11-05 — Model: rev_growth
• W3: GAAP operating margin — FY2026 within 10.5–11.0% guide — Q4 2026 release — ~Feb 2027 — Model: ebitda_margin
• W4: FY2027 guidance — organic growth ≥3% (Base) vs ≤0% (Bear) — Q4 2026 release — ~Feb 2027 — Model: rev_growth, ebitda_margin

Sources

• EPAM Systems (EPAM) statistics — https://stockanalysis.com/stocks/epam/statistics/ — accessed 2026-10-20
• Accenture, Cognizant, Globant statistics — https://stockanalysis.com/stocks/acn/statistics/ · /ctsh/ · /glob/ — accessed 2026-10-20
• EDGAR XBRL companyconcept (RevenueFromContractWithCustomerExcludingAssessedTax, OperatingIncomeLoss, DepreciationAndAmortization, PaymentsToAcquirePropertyPlantAndEquipment), CIK 0001352010 — https://data.sec.gov/api/xbrl/companyconcept/CIK0001352010/us-gaap/OperatingIncomeLoss.json — accessed 2026-10-20
• EPAM FY2025 Form 10-K, consolidated balance sheets — https://www.sec.gov/Archives/edgar/data/1352010/000135201026000015/R3.htm — accessed 2026-10-20
• EPAM FY2023 Form 10-K, consolidated balance sheets — https://www.sec.gov/Archives/edgar/data/1352010/000135201024000008/R3.htm — accessed 2026-10-20
• EPAM Q2 2026 results release — https://www.sec.gov/Archives/edgar/data/0001352010/000135201026000043/exhibit99_q2x2026.htm — accessed 2026-10-20
• EPAM Q3 2026 earnings date — https://www.mychesco.com/a/news/regional/epam-systems-sets-november-5-earnings-call-for-third-quarter/ — accessed 2026-10-20
• Accenture Q4 FY2026 results — https://www.sec.gov/Archives/edgar/data/0001467373/000146737326000037/q4fy26earnings8-kexhibit.htm — accessed 2026-10-20
• Globant Q2 2026 call highlights — https://www.marketbeat.com/instant-alerts/globant-q2-earnings-call-highlights-2026-08-14/ — accessed 2026-10-20
• EPAM vertical mix (Q3 2025, Zacks) — https://www.zacks.com/stock/news/2787574/epam-stock-gains-4-as-q3-earnings-and-revenues-beat-estimates — accessed 2026-10-20
• EPAM regional mix (Q2 2025, Fintool) — https://fintool.com/app/research/companies/EPAM/earnings/Q2%202025 — accessed 2026-10-20
• Macro log, 10Y 5.15% and fed funds on 2026-10-19 — logs/macro-2026-10.md — accessed 2026-10-20