EPAM reported strong third quarter 2025 results, with revenue reaching $1.394 billion, up 19.4% year-over-year on a reported basis and 7.1% on an organic constant currency basis, exceeding the high end of guidance. Non-GAAP EPS was $3.08, and the company generated record free cash flow of $286 million. Growth was broad-based across five of six verticals, driven primarily by AI-native services, data platforms, and cloud modernization. The company raised its full-year 2025 revenue guidance to $5.430 billion–$5.445 billion (15% growth at midpoint) and increased its organic constant currency growth outlook to 4.6% at the midpoint. New CEO Balazs Fejes emphasized the 'AI/RUN Transform' strategy and internal AI adoption ('Client 0') as key differentiators. While operating margins faced pressure from lapping prior-year credits and higher variable compensation, management expressed confidence in improving profitability through account margin enhancements and 'pyramid balancing' of its workforce.
| Metric | Value | Change |
|---|---|---|
| Q3 Revenue | $1.394 billion | +19.4% YoY (Reported) |
| Organic Constant Currency Growth | 7.1% | YoY |
| Non-GAAP EPS | $3.08 | -1.3% YoY |
| Free Cash Flow | $286 million | Record High |
| GAAP Gross Margin | 29.5% | -510 bps YoY |
| Headcount (Consultants/Engineers) | 56,100+ | +6.4% Organic YoY |
EPAM is aggressively positioning itself as an 'AI Native' leader through the launch of 'AI/RUN Transform,' a unified strategy harmonizing go-to-market efforts with consulting, frameworks, and talent. This strategy is supported by internal adoption, with over 90% of employees completing mandatory AI literacy training. The company is leveraging 'Client 0' status to develop proprietary tools like 'Agentic QA,' which claims 10x efficiency over manual testing. This signal indicates a strategic shift from general IT services to high-value, IP-driven AI engineering, aiming to capture share as clients move from proof of concept to production.
Management highlighted a significant competitive advantage in 'complexity management.' CEO Balazs Fejes noted that clients are redirecting work from partners who 'successfully sold advanced capability, but failed to deliver it.' EPAM believes that while AI coding tools lower the cost of software development, they increase complexity, thereby raising the demand for 'true engineering expertise' and risk management. This suggests a strategic moat built on deep engineering DNA rather than just labor arbitrage.
The company is actively pursuing 'Agentic Business Process Outsourcing (BPO)' as a new growth vector, distinct from previous RPA waves. By utilizing acquisitions like First Derivative and LYNXUS, EPAM is experimenting with automating complex processes (e.g., KYC, fin crime) using 'Agentic AI.' This represents a strategic evolution from traditional IT services into business process operations, potentially opening new revenue streams that combine engineering with process automation.
Financial discipline and capital allocation are sharpening. The company announced a new $1 billion share repurchase program, signaling confidence in cash generation (record FCF of $286M). Operationally, management is focused on 'pyramid balancing'—optimizing the mix of junior and senior engineers globally—to improve account margins. This focus on profitability, alongside raising the lower end of full-year guidance, indicates a strategy to balance growth investment with margin expansion.
Gross margin compression remains a concern, with GAAP gross margin falling to 29.5% from 34.6% in the prior year, and non-GAAP gross margin dropping to 31% from 34.3%. Management attributed this to lapping a Poland R&D credit, higher variable compensation, and lower profitability from recent acquisitions. While they expect improvement, the current compression suggests integration challenges and cost inflation that could persist if pricing power does not keep pace.
The outlook for Q4 implies a deceleration in organic constant currency growth to 4.4% at the midpoint, down from Q3's 7.1%. CFO Jason Peterson cited the absence of 'excess client budgets' typical of year-end spending, a contrast to Q4 2024. This indicates a more disciplined spending environment by clients and introduces risk that the growth ramp may be bumping against near-term demand ceilings.
The NEORIS acquisition is presenting headwinds, with a 'modestly negative impact on organic constant currency growth' due to a lead customer in Mexico being impacted by U.S. tariffs and political instability. This highlights the risks associated with geographic exposure and integration volatility in specific regions, potentially masking the core performance of the legacy business.
Management acknowledged that while AI projects are scaling, the 'Agentic' delivery lifecycle is 'complex or even more complex than compared to SDLC.' This complexity could slow down deployment times and increase the cost of delivery, potentially offsetting the efficiency gains promised by AI automation. If clients struggle to realize ROI due to this complexity, demand could slow.
Overall: Management conveyed a high level of confidence and energy, particularly from new CEO Balazs Fejes, who opened with a metaphor about the quarter arriving faster than a 'double espresso shot.' The tone was assertive regarding EPAM's competitive differentiation in engineering quality versus competitors who 'failed to deliver.' During Q&A, the tone remained pragmatic but optimistic, acknowledging the complexity of AI implementation while emphasizing the strength of the pipeline and pricing integrity.
Confidence: HIGH - Management raised guidance for the full year, cited record free cash flow, and provided specific metrics on AI project expansion (60-70% moving from PoC to larger programs). The specificity regarding the 'AI/RUN' strategy and the declaration that 2026 organic growth will be higher than 2025 underscores their conviction in the trajectory.
$5.430 billion - $5.445 billion (15% growth at midpoint)
4.6% at midpoint (Raised from 4%)
$11.36 - $11.44
$1.380 billion - $1.395 billion
4.4% at midpoint
Hedging & Uncertainty: Management used specific qualifiers to temper expectations regarding the speed of AI adoption, frequently using phrases like 'early days' and 'we don't know yet where the market will go' when discussing Agentic BPO. However, they were less hedged on core performance, using definitive language like 'we are confident' and 'upward trajectory' regarding the overall business. The use of 'believe' was prominent when discussing long-term structural shifts (e.g., 'We believe that as complexity rises so does enterprise risk'), indicating a strategic thesis rather than immediate realized revenue.
AI-native revenues continues to grow nicely with a third consecutive quarter of double-digit sequential growth. - Balazs Fejes, CEO
We are seeing our clients building more solutions... the equation of or the percentages will start skewing towards build versus the buy side. - Balazs Fejes, CEO
We are not expecting to see a significant release of excess client budgets... typical seasonality will also have an impact. - Jason Peterson, CFO
The absolute need for true engineering expertise... are overlooked and underestimated. - Balazs Fejes, CEO
We expect year-over-year revenue growth on an organic constant currency basis to now be 4.6% at the midpoint. - Jason Peterson, CFO
Analyst Sentiment: Analysts were generally positive and inquisitive, focusing heavily on the mechanics of the new AI strategy ('Agentic BPO', 'AI/RUN') and the sustainability of the margin expansion. There was skepticism regarding the 'build vs buy' dynamics in software and the specific impact of the NEORIS acquisition.
Management Responses: Management responses were detailed and technical, particularly CEO Balazs Fejes, who leaned into engineering complexity to justify EPAM's value proposition. CFO Jason Peterson was direct about margin headwinds and the lack of 'excess budget' release in Q4, managing expectations effectively.
Discussion on 'Agentic BPO' and how it differs from RPA, with management noting it is 'early days' and requires high-level engineering.
Inquiry into the sources of Q4 growth deceleration, attributed to seasonality and lack of budget release rather than demand weakness.
Questions on competitive positioning, with management asserting they are winning work from failed competitors.
Detailed explanation of 'pyramid balancing' to drive cost efficiency and margin improvement.
EPAM is successfully navigating the AI transition, leveraging its deep engineering heritage to capture share in 'AI-native' services while competitors struggle with execution. The Q3 beat and raise, combined with record free cash flow, demonstrate operational resilience. While margins are currently pressured by lapping credits and investments, the strategic focus on 'AI/RUN' and internal adoption ('Client 0') positions the company for sustainable growth. The shift towards higher-value, complex AI projects and the expansion into Agentic services provide a strong multi-year growth narrative. The valuation is supported by a new $1B buyback authorization, signaling management's confidence in the intrinsic value of the business.
AI is driving a 'structural shift' in demand, moving from proof of concepts to production. Clients are prioritizing data and cloud foundations, and there is a noticeable trend of 'build vs buy' skewing towards building custom solutions.
Clients are not releasing 'excess budgets' at year-end as seen in previous years, indicating a more disciplined spending environment. However, investment in AI remains a priority.
Political instability and tariffs in Mexico are impacting specific clients (referenced regarding NEORIS), highlighting geopolitical risks in nearshore delivery locations.