IMAGE VIA CAMPION COLLEGE
NEPAL HAS pulled off an economic miracle. Wedged between two Asian giants, this nation of 30m people has built a billion-dollar digital-services industry from almost nothing. IT exports surged by 64.2% in 2022 alone, and continued growing to an estimated $1bn by 2025. Freelancers on Upwork earn salaries that dwarf those of local graduates. The sector now accounts for perhaps 1.7% of GDP (a conservative estimate; official figures likely undercount informal digital work). For a country long dependent on selling labour as migrant workers, this looks like the future.
Yet ask any tech boss in Kathmandu about the country’s 17,000 annual IT graduates, and a different story emerges. Around 85% of them need six to nine months of retraining before they can do anything useful. Over 60% struggle to find any job at all in the six months after graduation. A staggering majority of firms report they cannot find qualified people for senior roles: a shortage so acute that it has become the sector’s biggest growth constraint.
So here is the paradox: a booming sector, a flood of graduates and a desperate shortage of people who can actually work. What is going on?
The short answer is that Nepal’s universities have become machines for producing credentials and not competence. Their curricula are stuck in a time when the civil service was the main employer and memorising definitions mattered more than building things. Students spend years on theory; sit semester-end written exams; and graduate without ever having worked in an Agile team or shipped a piece of software. The system confuses passing with knowing.
One recent farce captured it nicely: Tribhuvan University placed its flagship IT degree (BCA) under the Faculty of Humanities and Social Sciences, rendering its graduates ineligible for the very bank jobs the degree was meant to unlock (an error later reversed).
Industry has responded by building a parallel universe. Coding bootcamps that cost under $500 can turn out a junior developer in four to six months, with job-placement rates above 85% (self-reported, but widely echoed by employers). Self-taught kids on YouTube and freeCodeCamp often emerge more employable than degree-holders. This two-tier system entrenches inequality. Wealthy, urban, English-speaking students can afford to top up a weak degree with real skills. Their poorer, rural peers get the same piece of paper and nothing else—and find the door slammed shut.
If that were the only problem, it would be bad enough. But two bigger storms are gathering. The first is the experience paradox. Entry-level job postings commonly demand two years of experience. Firms, operating on slim margins, say they cannot afford to train fresh graduates. Yet if every firm refuses, no one builds the pipeline. This is a collective-action failure.
The second storm is AI. Entry-level coding jobs—the roles Nepal’s system churns out graduates for—are being automated away. A Harvard study found that after GPT-4’s release, employment of 22- to 25-year-olds in software development fell by 13%. A peer-reviewed paper by two Microsoft executives warns of a “structural crisis”: AI supercharges senior engineers but imposes an “AI drag” on juniors who lack the judgment to fix machine-generated code. Entry-level tech hiring has already dropped by 20–25% across several markets. In Bangladesh, some outsourcing firms have cut junior staff by 80% after adopting AI—a documented case that Nepal ignores at its peril.
Nepal’s entire strategy—produce cheap graduates in bulk, sell their labour to foreign clients—relies on a premise that is dissolving. Why hire ten juniors in Kathmandu when two AI-augmented seniors in Melbourne can do the same work? The cost advantage that built the boom is eroding in real time.
The government talks big. An “IT Decade” targets $22bn in exports and 500,000 jobs. Tax breaks are generous. But the infrastructure remains third-world: over $900 annual cap on international payments blocks access to basic tools like cloud APIs and software subscriptions. The national digital-ID programme has millions of printed cards stuck in limbo (7.5m printed, only 3.9m delivered). Only 85 of 380 local governments use a basic firewall to secure their systems.
The bigger failure is imagination. Nepal treats its tech sector as a labour-export business, not as a foundation for building institutional knowledge. Its brightest developers increasingly work directly for foreign clients, contributing nothing to local training or entrepreneurship. They are digital migrants doing the same economic activity as a construction worker in the Gulf, just with a laptop.
What must change? First, universities need a shock. Accreditation should depend on graduate employment outcomes. Curricula must become project-based with industry practitioners in the classroom. Pokhara University’s new AI degree, which balances theory with hands-on skills, offers a glimpse of what is possible. Second, the government must fix the basics: internet reliability, payment caps, regulatory chaos. Third, firms must accept that training juniors is a collective good, and not a cost to be avoided. No one benefits from a system where everyone poaches and no one teaches.
Nepal’s tech boom is real. But it is a boom built on sand. The country has a choice: keep producing graduates who cannot work, watch AI eat the jobs they were trained for, and settle for being a low-cost back office forever. Or fix the institutions that turn youthful energy into real capability. The credentials are piling up. The competence is not. That gap will not close by itself. ■
