About Himalayan Silicon Valley
An AI infrastructure company, operating from Lumbini.
HSV is the AI and technology arm of The Promised Group, a thirty-year institutional group across South Asia. We build compute, products, and a trained workforce for the AI economy, and sell into the markets that buy them. The company was incorporated on December 12 at a World Peace Program in Lumbini.
Foundation
The competitive advantage is geographic, philosophical, and operational at once.
Three things put a company in this part of the world. The talent: a generation of Nepali engineers and operators who, given training and a clean pipeline, deliver to the same standard as any team in San Francisco or Singapore. The energy: hydropower running off snowmelt, sun on the foothills, and a grid we can plug compute directly into. And the philosophical inheritance: a 2,500-year tradition of asking what right action looks like, which is not a marketing line so much as an operating constraint we apply to what we build, how we pay, and who we work with.
"The most useful AI infrastructure in this part of the world will not be built in the obvious places."
What we are building
The pieces of the stack, with the work that is already live.
Compute
An institutional-grade GPU cluster
H100, A100, L40S, RTX 6000. Built in Nepal, open to partners outside it. Training, fine-tuning, inference and research, with engineers on the floor rather than a hands-off data centre.
Mandate
Buddha AI and the Lumbini Platform
Contracted by the Lumbini government to build and run the digital infrastructure of the city: AI products and platforms that serve pilgrims, residents and the institutions that operate around them.
Academia
Lumbini Buddhist University, Kathmandu University
Active collaborations with two of the country's leading universities to put graduating engineers into AI work that travels.
Capital
Five financial centres
Capital deployed and committed from the United States, Hong Kong, China, Singapore and London. Long horizons, institutional partners, a thesis that doesn't fit on a single slide.
Workforce
Twenty thousand trained operators by 2027
A specific number with a date attached: twenty thousand annotators, reviewers and project leads working on AI for global partners, paid well, without leaving the country.
Bridge
A working address for global AI
We bring foreign AI teams here to train, build and run pilots, and we run their data work back out. Nepal is the operating address; the customers sit elsewhere.
Leadership
The people who built the system and ran the first pilot through it.

Chairperson
Nikesh Adhikari
Sets strategy, capital structure and institutional architecture. Comes out of The Promised Group and runs HSV as its AI and technology arm.

Managing Director
Samiksha Baral
Owns delivery. Workstreams, partner relationships and output standards run through her. The reason the work ships.
Operations · Ventures · Partnerships
The people behind the pipeline.
Prakash Jnawali
Chief Accountant
Prajwol Karki
Operation Associate
Aananda P. Giri
Automation Engineer
Nushant Ghimire
Operation Associate
Prashamsa Poudel
Operation Associate
Prasanna Poudel
Operation Associate
Anisha Regmi
Brand Manager
Usha Niraula
Sales Lead
Subash Khatiwada
Field Sales Representative
Sumitee Budhathoki
Field Sales Representative
Sangam Shrestha
Field Sales Representative
Samana Subedi
Operations Manager
Sulochana Khatiwada
Government & Compliance Lead
Bishal Khadka
Operational Associate
Krishna Gyawali
Consultant
Sushil Khadka
Intern
Jasmine Dahal
HR & Admin Officer
A career ladder, not a labour pool.
Six tiers, one track.
Progress is not self-reported and it is not decided in a meeting. An automated system tracks performance in real time and handles promotion, task assignment and quality checks against a fixed set of criteria, the same rules for everyone, applied continuously.
Auto-tracked KPIs
Every submission, review and hour worked updates your record automatically. No self-reporting.
Quality-gated advancement
Promotion requires clearing a minimum approval-rate bar, not just time served.
Smart task assignment
Work is matched to your tier and skill profile automatically as it becomes available.
No manual paperwork
The moment your numbers clear the threshold, you move up: no forms, no waiting on a manager.
Tier 1
Auto-trackedAnnotators
Image, video, dialogue, RLHF, and classification tasks.
Tracked from day one: output volume, approval rate.
Tier 2
Auto-trackedContent Operators
Subtitle, dialogue, scene tagging, and localisation.
Auto-promoted once volume and accuracy clear the bar.
Tier 3
Auto-trackedQA Reviewers
Output accuracy, consistency, and client satisfaction.
Quality-gated: requires a sustained minimum approval rate.
Tier 4
Auto-trackedCoordinators
Task tracking, team communication, and reporting.
Unlocked by tenure, badges, and sustained quality.
Tier 5
Auto-trackedProject Leads
Client management, delivery strategy, and team scaling.
Requires a proven delivery record across projects.
Tier 6
Auto-trackedInfrastructure
Compute, APIs, automation, and technical systems.
Reserved for the highest tenure and contribution scores.
Institutional network
The names behind the work.
HSV's counterparty list runs through The Promised Group's thirty-year institutional network: multilateral agencies, sovereign development banks, global technology companies and major corporates. We do not show up alone.
Referenced under public-relationship status
Three lines of work
Compute, products, people.
Compute
An institutional GPU cluster in Nepal, with hydropower behind it and engineers on the floor. Open to partners abroad for training, inference and research.
Products
Buddha AI, the Lumbini Platform, Lucy, Kishan AI. Products built for governments, farmers and the people who use them daily, not only enterprise buyers.
Workforce
A trained Nepali team running annotation, data operations and AI services for partners abroad. Verified output, structured careers, no migration required.
How it fits. The three lines reinforce each other. Compute attracts customers. Customers create work for the operators. The operators feed back into the products. Each one would be a thin business on its own. Together they look like a company.
Where the customers are
Built in Nepal, sold elsewhere.
United States
Data and research partnerships
Annotation, data operations and workforce services delivered into the largest AI market in the world.
Europe
Aligned on conditions
Working with European institutions for whom how the work is done, and by whom, is part of the specification.
China and India
A working middle ground
Nepal sits between the two largest populations on the planet, with access to both and obligations to neither. We work that geography rather than apologise for it.
Lumbini
The government mandate at home
Contracted by the Lumbini Development Trust to build and operate Buddha AI and the Lumbini Platform: the digital layer of one of the most-visited religious sites in Asia.
The thesis, in plain terms
A useful AI company doesn't have to come from the obvious places.
The geography is the moat. Cheap clean power, talent that hasn't yet been priced into the global AI market, and a position between China and India that makes Nepal an unusually workable place to operate from. Companies that figure this out first generally keep the advantage for a decade.
The philosophy is operational, not decorative. Most of the firms shaping AI are running without much of a framework for what they are doing. We use the one that has been here for a couple of millennia. It shows up in pay structures, partner selection and product decisions, not in slogans.
The plan is straightforward enough: build well, ship clean work, treat the people doing it the way they should be treated, and let the rest follow.
Build with HSV
We're early in the build. Come in now.
Buy compute, run a pilot through the data operation, or co-develop infrastructure with us. The room is open and the people who walk in early generally end up shaping what comes next.

















