EPAM Systems: AI Is Eating The Old Business, But 8.7x Earnings Is Too Cheap

EPAM warns of near-term revenue headwinds as AI cannibalizes legacy services, but AI-native revenue now tops 11% and is growing double-digit sequentially. Even with flat topline growth, margins and adjusted EPS are improving and the stock trades at 8.7x forward non-GAAP earnings. Activist Engine Capital is pushing a $750M buyback, which could materially reduce the share count at current valuations.

Analysis

EPAM looks like a classic "good slowdown" setup: if AI is pulling low-value legacy work out of the P&L, reported topline can stay soft while mix, utilization, and buyback-driven EPS all improve. At sub-9x forward earnings, the market is pricing a structurally impaired services vendor; that leaves room for a re-rate if management proves AI work is accretive rather than merely substitutive. The activist angle matters because a large repurchase can mechanically lift EPS and free cash flow per share faster than the street is likely modeling.

The main risk is that the AI mix shift is still early and may be more pilot-heavy than recurring, so the sequential growth signal could fade before it becomes a durable backlog story. In that case, the stock may remain a value trap despite a cheaper multiple, and the read-through turns negative for other engineering-heavy services names with less capital return support. Over 1-3 months, the catalyst is the next earnings/guidance cycle and any formal buyback authorization; over 6-18 months, the question is whether EPAM becomes a higher-ROIC, lower-growth compounder or simply a shrinking legacy book.

Consensus may be over-focusing on revenue cannibalization and underweighting the operating leverage from labor productivity plus share shrink. The missing data is the quality of AI-native bookings: gross margin, contract duration, and whether those wins are net-new or just replacing low-margin maintenance. If management does not show a step-up in backlog or guidance, the multiple likely stays capped; if it does, the stock can rerate even without topline acceleration.

Key Decisions for Investors

• Initiate a tactical long EPAM at or below ~9x forward earnings, sized modestly, targeting a re-rate toward 10-11x on evidence of buyback execution and margin expansion over the next 1-2 quarters; thesis fails if legacy revenue decelerates faster than AI-native can offset it.
• Use EPAM as a pair trade versus a richer-valued IT-services peer basket (ACN/GLOB) if the market starts pricing AI disruption across the group; EPAM has a cleaner capital-return catalyst and more immediate EPS leverage.
• Set a watch item for the next earnings release: AI-native backlog quality, gross margin, and management's buyback authorization/timing; if those are not improving, treat the move as value trap risk rather than a durable thesis.
• If the stock rallies sharply on activist headlines before the buyback is formally approved, take profits on part of the position and keep the remainder only for the execution catalyst.