Finn
EPAM IT Services · AI services · Digital engineering · Global delivery · Thesis updated August 11, 2026

AI scales while North American demand falls

01 Running thesis

A transition taking longer than expected

EPAM is undergoing a structural shift. The company is trying to replace older, manual software engineering tasks with high-value AI transformation projects. The Q2 2026 results showed that the AI strategy works, but the older business is shrinking too quickly.

The bull case relies on pure AI revenue compounding fast enough to lift the entire company. With Q2 AI revenue topping $160 million, EPAM has a clear path to its $600 million goal for 2026. This new work is also structurally more profitable, which could boost margins once growth resumes.

The bear case centers on execution and timing. Management admitted a capability gap in its North American sales motion. Clients are cutting traditional work faster than EPAM can sell them AI replacements. To make matters worse, the massive AI vendor consolidation deals the company hoped to close in late 2026 have been delayed to 2027.

Finn's view is cautious. The AI growth engine is real, but a turnaround in North America will take time. Until the core business stabilizes and large deals actually convert to revenue, near-term growth will remain sluggish.

Aug 2026Q2 2026 earnings showed pure AI revenue exceeding $160 million, but full-year organic growth guidance was lowered to 2% to 3% due to North American weakness and delayed large deals.
May 2026Q1 2026 earnings changed the debate. Guidance was cut to 4% to 6.5% revenue growth, but pure AI revenue exceeded $125 million and was described as more profitable than the company average.
May 2026The Q1 2026 Form 10-Q showed revenue of $1.400 billion, up 7.6%, and Americas segment operating margin improved to 16.6% from 14.8%. This eased the margin concern, even as growth risk increased.
Feb 2026The 2025 Form 10-K confirmed the AI-native strategy but added margin concerns. Cost of revenue rose as a share of revenue, and EPAM formally warned that AI tools could replace some services demand.
Feb 2026Q4 2025 earnings gave investors a clear AI target. Management said AI-native revenue should exceed $600 million in 2026, giving the growth case a measurable driver.
Nov 2025The Q3 2025 transcript was not available, so the published view did not change from the 10-Q analysis.
Nov 2025The Q3 2025 Form 10-Q showed organic growth accelerating, which strengthened the case that demand was improving. Margin pressure from acquisitions remained the main offset.
Aug 2025Q2 2025 earnings disclosed 5.3% organic constant currency growth. That helped prove the business was not growing only because of acquisitions.
02 Business model

Selling expert teams to big clients

EPAM makes money by helping companies design, build, and run complex technology systems. Clients hire EPAM for software engineering, data analytics, cloud work, digital customer tools, and now AI-native transformations. Contracts are mostly time-and-materials, where clients pay for work hours, and fixed-price projects, where EPAM agrees to deliver a defined outcome for a set price.

The company relies on its reputation for high-quality engineering. That edge lets EPAM compete for hard work where cheap labor alone is not enough. The risk is that clients can delay projects fast when budgets tighten, which is what management is seeing right now in North America.

EPAM is also changing where the work gets done. India became its largest delivery center in 2024. This helps the company meet client demands for lower costs while trying to protect its premium service quality. If that shift hurts quality or if wage costs rise faster than pricing, margins can suffer.

The model works best when utilization is high, which means engineers are busy on paid client work. It breaks when clients pause projects, fixed-price contracts go over budget, or EPAM hires ahead of demand.

03 Product portfolio

From engineering core to AI-native work

Growth engine

AI-native transformation

This is the center of the current bull case. Pure AI revenue exceeded $160 million in Q2 2026, making up over 11% of the total business.

Cash cow

Core engineering

EPAM's base business is custom software engineering for large clients. It funds the company and supports the AI push, but it is currently facing a sharp drop in demand.

Steady

Data and analytics

Data work helps clients organize, clean, and use information across their business. It is also a critical building block for AI projects.

Steady

Digital engagement

This includes customer-facing digital products, design, and experience work. Demand is cyclical and highly dependent on client budget cycles.

Option

EPAM DIAL and EliteA

These proprietary platforms help EPAM package AI work and speed delivery. They are key to winning the large vendor consolidation deals expected in 2027.

04 Business segments

Two reported regions, one clear gap

Americas57%flat
Europe43%growing fast

Segment mix is based on Q1 2026 reportable segment revenue in EPAM's Form 10-Q. Europe continues to significantly outpace the Americas.

05 Risk factors

What could break the thesis

North America slowdown lasts longer

High impact · High odds

Management cut 2026 guidance again because traditional services are dropping off fast in North America. The Americas segment grew only 0.5% in Q2 2026. If the commercial rebuild in this region fails, AI growth will not be enough to lift the whole company.

We watchAmericas segment revenue growth and management comments on business development improvements.

Large AI deals pushed out again

High impact · Medium odds

EPAM is relying on large AI-native vendor consolidation deals for future growth. These deals were already delayed from late 2026 to early 2027. Further procurement delays or scope reductions could push the growth recovery further away.

We watchSigned multi-year AI transformation deals and revenue contributions in early 2027.

Software & Hi-Tech vertical weakness

Medium impact · High odds

The Software and Hi-Tech vertical declined 1.3% in Q2 2026. SaaS clients may be permanently redirecting services spend toward AI infrastructure and tokens instead of hiring EPAM for projects.

We watchRevenue growth specifically within the Software & Hi-Tech client vertical.

AI replaces some services work

High impact · Medium odds

AI is both an opportunity and a threat. EPAM warns that clients could use AI tools as a replacement for some services or software EPAM builds. If that happens faster than EPAM grows its own AI-native work, overall demand could shrink.

We watchClient comments about insourcing work with AI tools and pressure on billable headcount.

Margin pressure returns

Medium impact · Medium odds

While operating margins have shown resilience, gross margin faces pressure from compensation costs and a shift in project types. If pricing stays weak or utilization falls further in North America, profitability could suffer.

We watchGross margin trends and utilization signals in earnings reports.
06 Quick answers

In one breath

What does EPAM Systems do?

EPAM helps companies build and improve software, data systems, cloud platforms, and AI tools. It is a services company, so most revenue comes from paid client projects rather than selling packaged software.

Why is AI important to EPAM?

AI is now the main growth story. Management said pure AI revenue exceeded $160 million in Q2 2026 and is more profitable than the company average.

Why did EPAM lower its 2026 outlook again?

Clients in North America are reducing traditional task-based services faster than new AI work can replace them. Management also said large AI vendor consolidation deals were delayed from late 2026 to early 2027.

Is EPAM mainly a U.S. business?

The Americas segment was 56.8% of Q1 2026 segment revenue, making it the largest region. However, Europe has been growing much faster recently.

Get started with Finn today