Outsourcing IT to Nepal in 2026: A Buyer's Guide to the Kathmandu Market

The billion-dollar number, and what's behind it

NAS-IT, Nepal's software and IT services association, put the country's annual IT service exports at roughly Rs 145 billion in early 2026. Call it a billion US dollars. For a country of about 30 million whose total goods exports were roughly Rs 277 billion in 2024/25 (more than a third of that refined soybean oil sold to India), that's a large number, and it arrived fast: the comparable figure for 2022, from an Institute for Integrated Development Studies (IIDS) study, was somewhere near USD 515 million. Treat the new number as an industry estimate, not an official count. Nepal Rastra Bank books a lot of these payments as remittances, so nobody can measure them cleanly.

Around 100,000 people work in the sector. Industry estimates put the number of outsourcing firms above 6,000, though that count includes a very long tail of two-person operations. Roughly 90 percent of IT employment sits in Kathmandu Valley, which is both the industry's strength and its single biggest structural risk.

What are they actually doing? Three clusters dominate. Software product engineering for US and European clients. Healthcare data work for American firms, claims processing, coding, analytics. And home-loan and financial data processing for Australian lenders, a niche where more than a dozen firms employ thousands of people and almost nobody outside it has heard of. AI data work, annotation and evaluation, has grown on top of all three.

Related reading: India-Plus-One in 2026: A South Asia Delivery Strategy Beyond Bengaluru · Bangladesh's LDC Graduation 2026: What IT Services Buyers Need to Know · Setting Up a Global Capability Centre in Pakistan in 2026: An Employer's Guide.

Who's actually delivering there

Nepal's outsourcing story has a handful of firms doing most of the reputational work, and knowing their names is useful because they set the ceiling for what you can expect.

CloudFactory is the one most Western buyers have encountered without realising it started in Nepal. Founded in Kathmandu in 2010 on an impact-sourcing model (it's headquartered in the UK these days), it built a managed workforce for data annotation and now serves machine learning teams globally. Fusemachines, started by Sameer Maskey, runs AI engineering and AI education out of Nepal with a New York front end. Leapfrog Technology has built product engineering teams for US healthcare and fintech clients for over a decade. Deerwalk grew out of US health-data analytics work and was bought by Cedar Gate Technologies in 2020. F1Soft sits on the domestic side and matters because it proves Nepali engineers have shipped payment infrastructure at national scale.

That list is short, and that's the point. Below the top tier, quality falls off steeply and quickly. Nepal does not yet have the depth of credible mid-market firms that Sri Lanka or Vietnam has. If your shortlist includes a company you cannot find three referenceable Western clients for, you are the experiment.

There's a second-order effect worth understanding. Because the credible firms are few, most of Kathmandu's senior engineers have passed through one of them, which means the informal network is dense and reference-checking a candidate or a subcontractor is unusually easy. Ask a CTO at Leapfrog about a firm you're evaluating and you will get a straight answer, because they've almost certainly worked with or hired from them. That kind of market transparency disappears the moment an industry gets big. Use it while it lasts.

Rates, and how they compare

Here's where Nepal is interesting. Blended rates usually sit below India for equivalent seniority, sometimes well below. The catch is that published rate guides disagree with each other wildly, so the bands below are rough indicative ranges, not survey data.

Treat these as negotiating anchors, not price lists. They move with contract length, team size, and how much bench the vendor has to carry. A 20-person dedicated team on a two-year commitment should land near the bottom of each band. A four-person squad on a rolling three-month contract will land near the top, and you'll deserve it.

The time zone is the other quiet advantage, and it's properly odd: Nepal runs at UTC+5:45, the only national time zone on a 45-minute offset (New Zealand's Chatham Islands, at UTC+12:45, are the other official one). For a London team that means roughly three to four hours of overlap, your morning and their afternoon, depending on whether the UK is on summer time. For US East Coast it means a real handoff model or nothing.

What Nepal is not good for

Vendor guides rarely include this section, which is why buyers keep discovering it themselves at cost.

Do not go to Nepal for scale-out contact centre work. The English is good but the accent-neutralisation infrastructure, the training factories, and the sheer volume of people willing to do night shifts simply aren't there the way they are in Manila or Hyderabad. A few firms have tried. None has come anywhere near Manila scale.

Do not go to Nepal if your programme requires a wholly-owned entity stood up in under six months. A foreign-owned IT services entity needs Department of Industry approval before the Office of Company Registrar will incorporate it, and Nepal Rastra Bank has to sign off on money going back out. The US State Department's 2025 investment climate statement still describes the one-window service promised under the 2019 FITTA law as improving only slowly. If you need speed, use an established local partner or an employer of record and revisit the entity question in year two.

Do not go to Nepal expecting deep enterprise-package skills. SAP, Oracle Financials, Workday, the big ERP implementation practices: that ecosystem is thin. You will find individuals, not benches.

And be careful with anything requiring heavyweight regulatory certification. Nepal has firms with SOC 2 and ISO 27001, and firms with HIPAA-aligned processes for the US healthcare work, but the pool of properly audited operators is small enough that you should verify the certificate and its scope rather than accepting the logo on a slide. Scope is where these things go wrong. A certification covering a single facility and one delivery line is not a certification covering your engagement.

The friction nobody puts in the pitch deck

Three things will cause you more grief than rate negotiation, and none of them appear in a capability deck.

Payments and foreign exchange. Nepal Rastra Bank keeps the rupee pegged to the Indian rupee at 1.6 and has to approve repatriation of funds, a process the US State Department calls difficult and time-consuming. FATF also put Nepal on its grey list in February 2025, which tends to mean extra bank checks on cross-border payments. For you as a buyer this mostly shows up as your vendor being fussy about invoicing structure and slow to set up new payment rails. It becomes your problem if you try to establish your own entity, because that's where the repatriation paperwork lands on your desk.

Contract enforcement. Nepal has no dedicated commercial courts (ordinary civil courts hear business disputes), and the State Department's verdict is blunt: enforcement of contracts is weak. In practice, disputes get settled commercially or not at all. Structure accordingly: milestone payments, source code in your own repository from day one, and no situation where the vendor holds something you cannot recreate.

Concentration. Ninety percent of the industry in one valley means one earthquake, one prolonged political disruption, or one serious connectivity failure hits your whole supplier base at once. The April 2015 Gorkha earthquake, which killed around 9,000 people, is still inside living memory for everyone in the industry, and the better firms have answers about continuity. Ask for the plan, in writing, and ask when it was last tested.

Is the talent pool deep enough for what you're buying?

For teams under about 30 people, yes, comfortably. Nepal's engineers are well regarded on fundamentals and, by most vendors' account, attrition runs lower than in Indian tier-1 cities, partly because there are simply fewer places to jump to. A team you build in Kathmandu tends to stay built, which over a three-year horizon is worth real money.

For teams over 100, the picture changes. The senior pool is thin and heavily competed over by the same five or six firms. If you want 15 senior engineers in Kathmandu inside a quarter, you are going to be hiring them away from CloudFactory and Leapfrog, and they know it. Price accordingly or plan a longer ramp.

There is also a steady drain of experienced people toward Australia, Japan, and the Gulf. Nepal exports labour at scale; the tech sector is not immune. The firms that retain well do it with equity, international exposure, and travel, not with salary alone.

How to structure a first engagement

Start with a bounded, non-critical workstream and a fixed three-month window. A migration, an internal tool, a test automation suite. Something where a bad outcome costs you a quarter and not a customer.

Insist on named individuals in the contract and a no-substitution clause with teeth. Body-shop substitution is one of the commonest complaints Western buyers raise about the Nepali mid-market, and it's entirely preventable at contracting time.

Then go. Kathmandu is a long flight from most places and almost every buyer who has made the trip says the same thing afterward: the offices are better than they expected and the engineers are more senior than the CVs suggested. Nepali firms undersell themselves on paper, which is unusual in this industry and, once you've sat through enough Indian and Eastern European pitches, oddly refreshing.

One practical note on the visit. Go in October or November, after the monsoon and before the winter inversion settles over the valley, but check the Dashain and Tihar festival dates first, because offices empty out for both. And build in two extra days. Nepali business culture runs on relationships more than most South Asian markets, and the meeting that decides whether you get the vendor's best team rather than their available team is usually the informal one on day three. Buyers who fly in for a single afternoon of office tours get treated exactly like buyers who flew in for a single afternoon.

The industry there is at an interesting point: large enough to be credible, small enough that a mid-sized Western client is a meaningful account rather than a line item. That asymmetry is worth something, and it won't last past the next few years of growth.