ILLUSTRATION: KATMANDU JOURNAL
IN THE past decade or so Nepal has acquired two reputations. One is for exporting people. The other is for exporting software. From a standing start, the country’s IT service exports have reportedly surged past $1bn a year, driven by a freelance-to-agency pipeline that turns bedroom coders into respectable firms. Kathmandu’s coffee shops buzz with founders pitching everything from digital wallets to AI tutoring. The government has noticed. It now offers tax holidays, subsidised loans and a formal “startup policy”. Finally, it seems a poor nation has found its 21st century niche. Do not believe the hype.
The trouble is not a lack of activity. More than 1,500 formally registered startups exist, perhaps ten times that if you count tiny outfits. Private equity and venture capital deployment hit $64m in 2024, the highest on record. A handful of companies, such as the digital wallet eSewa and the coding-tutorial platform Programiz, have built real businesses.
Yet the boom rests on three faults. First, the vast majority of “startups” are not innovative product companies but IT service shops: glorified outsourcers living on labour arbitrage. Second, the people who make those shops work keep leaving. Third, the state’s enthusiasm is no substitute for competence.
Call it the service trap. Nepali tech firms have discovered that selling cheap developers to American and Australian clients is a reliable way to generate cash. The 2025 budget reinforced that logic by handing a 75% income-tax exemption to IT service exports. That is a fine way to boost revenue. It is a terrible way to build durable firms. Service work creates little intellectual property, generates no network effects as well as evaporates the moment a cheaper competitor appears. Founders who dream of pivoting to products find themselves trapped: they need the services income to survive, but that income consumes the very time and talent needed to build something of their own.
The talent problem is even more acute. Nepal produces roughly 17,000 tech graduates a year. An estimated 85% of them need six to nine months of remedial training before they are useful. Then, just as they become productive, many leave. Over 60% of technical graduates plan to settle abroad permanently. A qualified engineer in Kathmandu earns less than a waiter in Dubai. That is not a labour market, meaning it is a revolving door. The 2025 IT ordinance, which allows Nepali employees of foreign firms to receive equity, is a welcome attempt to slow the exodus. But equity in a startup does not pay rent. Until local wages rise the brightest will keep voting with their feet.
Then there is the government. Its enthusiasm for “Digital Nepal” is real; its execution is not. The startup loan programme received over 10,000 applications in 2025. It approved perhaps 3-5% of them. The maximum loan amount was recently cut from $18,500 to $15,000. A study of budget implementation found that in the first quarter of this fiscal year, only 2.4% of announced economic programmes were fully implemented. More than half recorded no progress at all. Nepal’s tax regime requires 46 separate payments a year, consuming 377 hours of a firm’s time. That ranks 175th out of 190 countries. The government talks like Silicon Valley but governs like the 18th century.
What then is to be done? The temptation is to call for more subsidies and more training programmes and more slogans. Resist it. Nepal’s startup ecosystem does not need more government money. It rather needs less government interference. Start by scrapping the 46-tax-payment circus and replacing it with perhaps a single quarterly filing. Next, ease capital controls to allow foreign investors to put money in and take money out without begging for permission. Then fix the payment gateway mess so that a Nepali startup selling to a customer in London does not require a special waiver from the central bank.
The bigger lesson is bitter. For a decade Nepal has hoped that cheap coders would be its escape route from poverty and emigration. That hope is not wrong but it is incomplete. Cheap labour is a commodity; commodities are always undercut. The countries that succeed in tech are those that move from selling hours to selling solutions, from arbitrage to invention. Nepal is not there yet. Its founders are brave. Its diaspora is engaged and its recent growth is real. But sand does not become rock just because you shout “Digital Nepal” into the wind. ■
