Is AI Eating App Development Outsourcing? What Ukraine’s IT Export Numbers Say

Global spending on technology services just had its biggest quarter on record, and very little of it went to the people who write software for hire. In Q3 2026, the ISG Index recorded $52.5 billion in large contract value, up 63% year over year. Traditional IT outsourcing barely moved. Application development and maintenance, the core of the custom software business, went backwards.

Yet Ukraine, one of the world’s largest software outsourcing hubs, keeps growing its IT exports through a war and a global slowdown. This piece looks at what the numbers say, why the two trends diverge, and what it means for companies buying development work in 2027.

What the ISG numbers show

ISG tracks commercial outsourcing contracts worth $5 million or more a year. Its Q3 2026 data paints a split market:

• Infrastructure as a service more than doubled to $35.5 billion, driven by AI compute and cloud capacity.
• Cloud now makes up 78.5% of the combined market, its highest share ever.
• IT outsourcing was essentially flat, up 0.3% to $8.4 billion.
• Application development and maintenance (ADM), its largest service area, fell 8%.

Q3 was not a one-off. Earlier in the year, ISG estimated that the global IT outsourcing market fell 7% year on year in Q1 2026 to $7.9 billion, with ADM excluding infrastructure down 20%.

ISG’s own explanation is direct. Its leadership says incremental investment is flowing into infrastructure, platforms and software rather than outsourcing contracts, and that application development work is waiting behind AI-driven process redesign. Companies are buying the GPUs and platforms first and deciding later what software to build on top of them.

The pricing problem nobody can ignore

There is a second force at work. In its Q3 recap, ISG notes that service providers are using AI mainly to improve delivery, and that this is having a significant deflationary effect on pricing.

That is the uncomfortable math for any firm that bills by the hour. If an engineer using AI coding agents ships the same feature in a third of the time, a time-and-materials contract shrinks by two thirds. The client is happier, the vendor earns less, and the market statistics record a decline even when the same amount of software gets built.

The question has become a live debate inside the Ukrainian industry itself. In April, dev.ua’s AI show hosted EPAM Ukraine’s IT director to discuss whether AI will destroy Ukrainian outsourcing, including whether clients now demand lower prices because of AI.

Ukraine’s counter-trend

Against that backdrop, Ukraine’s IT export figures look surprisingly steady:

• In H1 2026, computer services exports reached $3.343 billion, up 4.1% on the same period of 2025, according to National Bank of Ukraine data.
• Q3 2026 brought $1.67 billion in IT services exports, 1.3% above Q3 2025, though 1.5% below Q2.
• Over January to August, IT accounted for 41% of Ukraine’s services exports and about one in every eight dollars of total export revenue.

At the company level, five of the ten largest IT service firms in Ukraine, including Grid Dynamics, Capgemini Engineering, Ciklum and EPAM, grew local revenue in Q1 2026 while the global outsourcing market was shrinking.

The picture is not all rosy. Exports have recovered to 2023 levels but remain about 9% below the 2022 peak, and growth slowed through Q3. The trend in client geography is also shifting, with the US gaining share while Germany slips. Still, an industry growing at all while its global segment contracts deserves a closer look.

Why Ukraine is holding up

Several factors help explain the gap.

The mix is moving toward AI work

Ukrainian firms have spent the last two years repositioning from staff augmentation toward AI integration, data engineering and agent development. That work sits closer to where ISG says the money is going: platforms and AI-enabled systems, not maintenance of legacy applications.

Cost pressure favors efficient delivery hubs

When AI compresses project budgets, buyers look harder at cost per outcome. Teams that combine senior engineering talent with lower regional rates can absorb price deflation better than high-cost onshore vendors.

Wartime resilience became a selling point

After nearly four years of operating through blackouts and air raids, Ukrainian vendors have built redundant power, distributed teams and continuity plans that many Western buyers now treat as proof of reliability rather than a risk.

The ADM decline hits maintenance hardest

Routine maintenance and simple feature work are the easiest tasks for AI coding agents to absorb. Vendors whose revenue depends on complex builds, integration and domain-heavy engineering are more insulated.

Is outsourcing just exporting labor?

Not everyone in Ukraine sees the current model as sustainable. Economist Viktor Galasyuk has argued that the country is mechanically selling the labor of its IT workers abroad, comparing outsourcing to raw material exports in agriculture. AI sharpens that critique. If hours become cheaper to produce, a business built purely on selling hours loses pricing power.

The firms likely to do best are the ones that stop selling hours and start selling results: fixed-scope AI builds, outcome-based pricing, reusable accelerators and intellectual property layered on top of engineering talent.

What buyers should ask an outsourcing partner now

If you are hiring a development partner in 2027, the AI shift changes the questions worth asking:

• How do you use AI in delivery, and who benefits? Ask whether productivity gains show up in your price, your timeline or the vendor’s margin.
• Do you offer fixed-scope or outcome-based pricing? Pure time and materials makes less sense when AI changes the time.
• What is your review and security process for AI-generated code? Faster output only helps if quality and IP controls keep up.
• Can you build AI systems, not just use AI tools? Integration, agents and data pipelines are where demand is growing.
• What is your continuity plan? For Ukrainian and other nearshore vendors, ask for specifics on power, connectivity and team distribution.

Conclusion

AI is not eating software outsourcing, but it is eating the version of it that sells hours to maintain applications. The global numbers show ADM spending stalled while AI infrastructure booms. Ukraine’s steady export growth suggests that vendors who reposition toward AI engineering and outcome-based delivery can still grow in a shrinking segment. For buyers, the opportunity is to demand that AI productivity shows up in what they pay and what they get.