Image: Caleb Luke Lin
In a cramped office in Kathmandu, the manager of a technology firm speaks of the internet as a “calculated risk”. His firm sells data annotation services (which add tags or notes to information so AI can understand it) to machine-learning developers in Silicon Valley and Europe, part of a digital-services export sector that has grown by an average of 12% a year over the previous two decades. Yet to conduct a video call with his clients eight times zones distant, he must pray for a stable connection. On average domestic firms endure six outages a month, each lasting around six hours.
This is the core contradiction in Nepal’s young digital economy: it can serve the world with ease, but still struggles to serve itself.
The country’s digital sector is diminutive, accounting for under 2% of nominal GDP. But it punches surprisingly above its weight in the services trade. Exports of digitally delivered services have hit over half of Nepal’s total services exports, a performance that outpaces most South Asian peers. Employment in information and communication technology (ICT), though a mere 0.35% of the labour force, is swelling. The government dreams of an IT-led export boom that could diversify an economy reliant on tourism and remittances. The private sector, responsible for 80% of the sector’s value added, is trying to oblige.
Yet this promise is systematically stifled by a hostile topography and an often-ineffectual state. The fallout is a digital economy advancing in pockets of excellence, mostly in the Kathmandu valley, but failing to overcome foundational gaps in connectivity and domestic adoption, as well as skills.
The big constraint is physical. Nepal is landlocked, which in the digital age means it has no direct access to the submarine fibre-optic cables that carry 99% of the world’s international internet traffic. The entire country must instead buy bandwidth from India and China.
This geopolitical dependence is costly. The annual bill for imported bandwidth is more than Rs4.7bn (about $33.1m). This fee is basically a bandwidth tax Nepal pays to its neighbours, raising costs for every digital business. The service that results is slow and unreliable—more of a risk, as the Kathmandu manager puts it, than a service.
Domestically, the connectivity gap is vast. The state-run Nepal Telecommunications Authority (NTA) is attempting to extend the fibre-optic network, a fiendishly expensive task in the mountainous terrain. The domestic fibre network is skeletal: nearly half the population lives more than 10km from a fibre-optic connection. A government plan to finance expansion through the Rural Telecommunication Development Fund (RTDF) is mired in regulatory and legal disputes, the classic fate of infrastructure plans that rely on state execution in Nepal.
Where fibre fails, wireless compensates. Mobile coverage is commendable: with 4G reaching 88% of the population, a higher share than even in India. This coverage exists mainly because companies have strong commercial incentives to serve the densest populations. However, the network is slow and unreliable. Download speeds lag regional peers, and for small firms that cannot afford expensive backup lines, each outage translates into lost income and missed deadlines.
Curiously the core data-centre infrastructure is relatively sophisticated. Nepal boasts six colocation data centres, which provide shared storage and processing power. Relative to population, this density is among the highest in South Asia. Furthermore, the country’s Internet Exchange Point (NIXP), where local providers swap traffic, helps keep domestic data local and prices lower. This advanced core is a legacy of early deregulation and private investment: a sign of what entrepreneurship can build when left unencumbered. However, it functions as an island of excellence in a sea of mediocre last-mile connectivity.
Outside the cities and export-oriented firms, the digital economy quickly weakens. This is where the real usage gap shows up. Not only because infrastructure is lacking but because many people see little need or benefit in going digital, as well.
Household adoption is low: only 38% of households have internet at home and just 15% own a computer. This gap persists in the commercial sector. A World Bank survey found that 37% of firms without a website simply saw no need for one. In 2023 only 9.6% of business payments were made electronically. Just 23% of sales were paid for online. The vast majority of commerce remains a cash-and-ledger affair.
Digital adoption is leaping, though only on the surface. Many firms build websites to look modern; yet without integrating its inventory, payments, logistics and so on the digital presence is just for show. Cash does not survive because people fear technology. It survives because it is the safer, simpler choice in a place where mobile data is costly and point-of-sale terminals are unreliable. Today paying digitally still takes more effort than handing over cash.
The government is attempting to bridge this gap through the development of Digital Public Infrastructure (DPI). The Nagarik app and the connectIPS payment system have seen respectable uptake. But the linchpin of any modern DPI, which is a national digital-identity system, is advancing at a glacial pace. A decade after the scheme was written into law, only 17% of Nepalis possess a digital ID. Without a basic digital ID system, smooth public and private digital services are only an idea.
To complicate matters, the recent devolution of power to local governments has run into the same old institutional limits. Many new local administrations lack the funds, or technical staff, or digital literacy to put public services online.
Should I stay or should I go?
The third, and perhaps most debilitating, constraint is human capital. The talent pool is meagre: fewer than 1% of Nepalis have a qualification in computer science or engineering. Technical colleges struggle to fill their seats and enrolment in computer-engineering courses runs below capacity, a sign that market signals are driving potential students elsewhere.
This is because the best and brightest tend to leave. The out-migration of skilled youth is staggering, a perennial feature of the economy. Each year over 250,000 young Nepalis receive permits to study. Tech firms in Kathmandu complain of investing heavily in training workers only to see them depart for higher salaries and more stable careers in Australia or the Gulf states. The economy is effectively training talent for export, subsidizing the labour needs of rich nations.
This skills shortage profoundly shapes the character of the digital sector. Most digital exports are currently concentrated in low-value, labour-intensive niches such as data entry, data annotation, call-centre operations and so on. Although companies like CloudFactory have proven that domestic workers can handle sophisticated work for artificial intelligence algorithms, such firms are rare. Without a better-retained talent pool, Nepal will struggle to climb the value chain from data annotator to software architect.
The sector also suffers from a brutal gender gap: women account for just 0.2% of the female labour force in ICT, compared with 0.5% for men. This represents a big untapped resource and shows entrenched societal or educational barriers that further constrain the available talent.
The government has plans for regulatory reform, expanded skills programmes, affordable devices and so on. But the distance between ambition and execution is big. The digital sector, like much of the economy, is left to find its way between a hostile topography and a hapless state.
In a Patan neighbourhood, two engineers are seen clambering onto a roof to attach a wireless transmitter to a bamboo pole. They are beaming internet to a school in a valley below, a place where fibre cannot reach. This is the pragmatism of necessity: a small, private-sector workaround to a vast, systemic problem.
Nepal has the essential components such as entrepreneurial vigour, an increasingly educated workforce and pockets of infrastructural excellence for a digital take-off. Its export success proves that the country is plugged into the global digital economy. But the ugly truth is that this external success is not an indicator of internal transformation. ■
