Chapter 2: The IT Decade: Policy, Ambition, and the Gap Between

A Country Declares a Decade

Nepal has a habit of declaring decades. There was the Agriculture Development Decade. The Tourism Decade. Each came with ambitious targets, ministerial speeches, and budget allocations that looked impressive on paper. Most faded from public memory before they reached their midpoint.

In 2024, the government declared a new one: the IT Decade, spanning 2024 to 2034. The stated targets are staggering: NPR 3 trillion (roughly $22.5 billion) in cumulative ICT exports and 1.5 million new technology jobs within ten years.

The potential upside justifies the ambition. Global estimates suggest that successful digital transformation can yield 14 percent cumulative additional GDP growth by 2030. For Nepal, capturing even a fraction of that growth would structurally reshape the national economy.

To put the government's export targets in perspective: Nepal's total IT and ICT exports in 2022 were estimated at $515 million. By early 2026, that figure has grown to roughly $1 billion. Reaching $22.5 billion by 2034 would require the sector to grow at approximately 35 to 40 percent annually, sustained for a decade. That growth rate is not impossible in technology (India's IT sector grew at similar rates during its breakout years in the late 1990s and early 2000s), but it demands a level of institutional execution that Nepal has not historically demonstrated.

So the question is not whether the ambition is admirable. It is whether the machinery exists to deliver on it.

The Tax Architecture

The government's fiscal approach to the IT Decade has been, by Nepali standards, genuinely aggressive. The 2025/26 budget classified information technology as a "special industry," a designation that carries real financial weight.

IT companies exporting services now receive a 75 percent exemption on export income tax. Individual freelancers and contractors exporting IT services abroad face a final tax rate capped at 5 percent. A 2 percent Digital Service Tax has been introduced for foreign digital service providers whose annual gross transactions in Nepal exceed NPR 3 million, signaling that the government is at least attempting to modernize its tax code for a borderless digital economy.

These are meaningful incentives. A Nepali software engineer earning $3,000 per month from a remote contract with a US company (a realistic figure for mid-level talent) would pay roughly $150 per month in taxes under the new regime, compared to substantially more under standard income tax brackets. For a freelancer weighing whether to stay in Kathmandu or move to Dubai for a tax-free salary, those numbers matter.

But tax incentives alone have never built a tech sector anywhere. India's IT boom was powered by tax holidays in Software Technology Parks, but it was sustained by IITs producing tens of thousands of engineers annually. Ireland attracted pharmaceutical and tech companies with a 12.5 percent corporate tax rate, but it succeeded because it already had an English-speaking, university-educated workforce. Tax policy creates the conditions for growth. It does not create the growth itself.

The National AI Policy: What It Says and What It Doesn't

In August 2025, the government approved the National Artificial Intelligence Policy 2082, a document that represents Nepal's first formal attempt to articulate a national AI strategy.

The policy's stated objectives are specific and, in several cases, commendably concrete:

• Increase the ICT sector's contribution to GDP by 1 percent through AI integration alone.
• Train 5,000 specialized AI professionals within five years.
• Establish AI Excellence Centres in all seven provinces.
• Integrate AI literacy into educational curricula from the primary level upward.
• Create an AI Governance Council for policy direction and a National AI Centre for implementation.

The provincial distribution mandate deserves particular attention. By requiring AI excellence centers across all seven provinces, rather than just in Kathmandu, the policy acknowledges a political reality that most technology strategies ignore: if the benefits of AI development are concentrated in the capital, the IT Decade will face the same geographic inequality that has undermined every previous development initiative.

What the policy does not address is equally important. It does not specify funding levels for any of its mandates. It does not identify which existing institutions will house the AI Excellence Centres or who will staff them. It does not confront the fact that Nepal's higher education system currently produces far too few technical graduates to fill a pipeline of 5,000 AI specialists, let alone the broader IT workforce the decade requires.

Policy without implementation architecture is aspiration. Nepal has no shortage of aspiration.

The Human Capital Crisis

This is where the IT Decade's ambitions collide most violently with reality.

Nepal's higher education system produces roughly 80,000 to 120,000 graduates annually. On paper, that sounds like a deep talent pool. In practice, the composition of that pool is badly misaligned with the demands of a digital economy.

Between 70 and 80 percent of these graduates hold degrees in management, education, humanities, and social sciences. STEM disciplines (science, technology, engineering, and mathematics) account for only 10 to 20 percent of the total. Compare this to the benchmark economies that Nepal's IT sector competes against: China produces 35 to 45 percent STEM graduates. India produces 25 to 35 percent. Even the United States, which is frequently criticized for its own STEM shortages, produces 20 to 25 percent.

The roots of this imbalance go deeper than university admissions. At the secondary school level, the entire national public school system employs just 2,670 mathematics teachers and 3,203 science and technology teachers for grades 9 and 10.12 That is fewer than 6,000 technical instructors responsible for the foundational math and science education of millions of students. The inevitable result: students arrive at university with weak quantitative skills, and relatively few pursue technology degrees.

The numbers downstream confirm this. According to the National Census, out of 3.97 million individuals who pursued education beyond grade 11, only 1.2 percent (roughly 45,880 people) specialized in computer and information technology.

This creates a paradox that anyone in Nepal's tech industry recognizes instantly. The graduate unemployment rate hovers between 20 and 25 percent. Over 60 percent of technology employers report difficulty finding candidates with the right skills. Young people cannot find jobs. Employers cannot find workers. The problem is not the quantity of graduates, but the match between what they studied and what the economy needs.

The transition from graduation to employment now exceeds 18 months on average, and nearly 45 percent of employed graduates end up in jobs unrelated to their field of study. Faced with these odds, many of the most capable students leave. Every year, over 100,000 Nepali students apply for No Objection Certificates (NOC) to study abroad, representing a stark reminder that the brain drain is accelerating as young talent seeks viability elsewhere, and that Nepal must build a competitive domestic tech sector to retain this crucial workforce.

The Realistic Path to 5,000 AI Professionals

The AI Policy's target of 5,000 trained AI professionals by 2030 is achievable, but only if the approach changes fundamentally.

The traditional path of waiting for universities to update curricula, hire faculty, and produce four-year graduates is too slow. By the time a curriculum committee convenes, drafts a syllabus, hires instructors, and graduates a class, five years have passed.

What has worked in comparable economies is a layered approach:

Accelerated bootcamps and certification programs. Rather than requiring four-year degrees, the government and private sector should co-fund intensive six-to-twelve-month AI and machine learning programs targeted at existing software developers, data analysts, and STEM graduates. The raw talent exists; what is missing is specialized training in neural networks, computer vision, NLP, and the frameworks (PyTorch, TensorFlow) that the industry actually uses.

University-industry partnerships. Progressive colleges in Nepal are already moving in this direction, affiliating with UK universities to provide degrees that meet international standards and integrating internship mandates into their programs. This needs to become the norm, not the exception.

Diaspora engagement. The Nepali tech diaspora, consisting of engineers working at Google, Meta, Amazon, and startups across North America and Europe, represents an underutilized resource. Structured mentorship programs, visiting faculty arrangements, and remote instruction partnerships could channel expertise back into the domestic pipeline without requiring people to physically return.

AI as a training tool. There is a recursive opportunity here that is easy to overlook: the same AI systems that Nepal needs to train professionals to build can be used as training tools themselves. An engineering student in Pokhara with access to Claude or GPT-4 has a more responsive tutor for machine learning concepts than most universities can provide.

AI as an Education Force Multiplier

There is a recursive opportunity here that deserves its own section because the stakes are so high.

Nepal has 2,670 math teachers for grades 9 and 10 across the entire public school system. That is one teacher for every several hundred students in a subject where foundational skills determine whether a student can ever pursue a technology career. Building more teachers takes years. AI tutors work today.

Platforms like Khan Academy's Khanmigo, Duolingo's AI language instruction, and custom-built LLM-based tutoring systems have demonstrated that AI can deliver personalized instruction at scale. A student in a remote village with a smartphone and a basic internet connection can have a one-on-one tutor that adapts to their pace, explains concepts in Nepali, runs practice problems, and never gets tired.

The strategic implication is clear: AI cannot replace teachers, but it can multiply their reach. Instead of one teacher delivering the same lecture to fifty students, an AI-augmented classroom lets the teacher focus on individualized support while the AI handles universal instruction, practice drills, and assessment.

For Nepal's IT Decade ambitions, this is not a peripheral social program; it is a pipeline strategy. Every student who graduates with strong math and science foundations is a potential candidate for the 5,000 AI professionals the National AI Policy demands, and beyond them, the 1.5 million technology workers the IT Decade targets. Investing in AI-powered education infrastructure today creates the workforce of 2035.

The National AI Policy mandates AI literacy from the primary level. This is correct. But implementation requires more than curriculum updates. It requires devices, connectivity, platform localization, teacher training on AI tools, and, critically, the compute infrastructure to run these services domestically rather than routing student data through foreign cloud platforms.

The IT Decade Targets: An Annual Reality Check

It is worth disaggregating the decade-long targets into annual milestones to understand what "success" actually requires.

These projections assume compounding growth rates of roughly 25 to 30 percent annually, which is ambitious but not unprecedented for a technology sector in breakout phase. The critical inflection point is around 2028 to 2029: if export growth has not accelerated beyond the current 18 to 20 percent baseline by then, the 2034 targets become mathematically unreachable.

This is why the infrastructure chapters that follow, which focus on hydropower, data centers, and compute capacity, are not tangential to the economic story. They are the economic story. Tax incentives and policy frameworks create the conditions for growth. But the physical infrastructure to host, power, and cool the compute capacity that the global AI industry needs; that is what will determine whether Nepal's IT Decade is different from all the decades that came before it.

Key Takeaways

• The IT Decade's NPR 3 trillion export target requires 35-40 percent annual growth, which is ambitious but achievable if infrastructure and workforce constraints are addressed.
• Tax incentives alone cannot build a tech sector; the binding constraint is the human capital pipeline, not the fiscal regime.
• Only 10-20 percent of Nepal's graduates are in STEM fields vs. 25-35 percent in India and 35-45 percent in China.
• The National AI Policy mandating AI excellence centers across all seven provinces is politically wise, but the policy lacks funding commitments and implementation architecture.
• AI-powered education is the only scalable solution to the teacher shortage (2,670 math teachers for the entire secondary system) and is essential for building the IT workforce of 2035.