Build the infrastructure
Compute, AI products, and government platforms. The physical and digital layer that global AI companies and sovereign institutions need, in a geography where almost no one else is building it.
Impact
What is already on the ground.
HSV is the AI and technology arm of The Promised Group. We hold government contracts, partner with multilateral institutions, and operate the compute, products and workforce that the next several years of regional AI work will run through. Nepal is where the advantage is; the customers sit elsewhere.
What is live
The work that has shipped, or is shipping.
December 12
Summit on AI and Consciousness
Convened at the World Peace Program in Lumbini, the same gathering at which the company was incorporated. Researchers, technologists and Buddhist scholars in the same room for a day, working through what AI ought to learn from humans before it teaches anything back.
Kishan AI
Open-source AI for farmers, in 40 governments
An open-source agriculture model, free at the point of use, subsidised by HSV and The Promised Land Foundation. Around forty governments across Nepal and India are deploying it at the village level. Soil, crop guidance and weather information, in the farmer's own language.
Carbon
The National Carbon Platform
Built with The Promised Group, working directly with the Ministry of Environment on Nepal's National Registry Program. Verified credits. Revenue routed back to communities. A measurable place for the country in the global climate stack.
Lumbini
Buddha AI and the Lumbini Platform
Contracted by the Lumbini government to build and operate the digital infrastructure of the city: AI products and a unified municipal platform serving pilgrims, residents and the institutions that work around them.
Workforce
A trained operator base
Annotators, reviewers, coordinators and project leads working inside HSV's quality pipeline and delivering to partners on four continents.
X001-0001
First pilot delivered
Twenty structured AI-companion interviews across four participant segments. Annotated, reviewed and shipped through the full QA pipeline. The schema held; the operation scales from here.
Each of the above is a signed contract, a deployed system, or a delivered pilot. None of it is roadmap.
Three vectors
The same engine runs every workstream.
Operate the pipeline
Every major AI model is trained on human-labelled data. We run a pipeline that ships clean output: schemas designed with the partner, operators trained against them, every batch reviewed before it leaves the building.
Build with constraints
The Buddhist tradition is treated here as an engineering constraint rather than branding. It shows up in pay structures, partner selection and product decisions. The same rigour applies to QA review and to wages.
The number
Ten thousand people in well-paid digital work by 2027. The metric is precise on purpose. It is the natural output of running this kind of company well, not a substitute for one.
What is next
The pipeline that is in build.
Compute
The first institutional GPU cluster in Nepal
H100, A100, L40S, RTX 6000 capacity for training, inference and research. The hardware floor that the rest of the company runs off.
Academia
University partnerships
Active collaborations with Lumbini Buddhist University and Kathmandu University to put graduating engineers into AI work that travels.
Global AI
An operating address for foreign teams
Bringing AI teams from the United States, Europe, China and India to train, build and run pilots from Nepal, then routing the data work back to them.
Farmers
Kishan AI at the next scale
Pushing the platform across South Asia, with the cost of use kept at zero so the farmer in the most remote village has access to the same intelligence as a global enterprise.
Climate
The National Carbon Registry, live
The registry goes live with the Ministry of Environment. Verified credits, community revenue, and Nepal's forests formally recognised in the global carbon market.
2027
Ten thousand operators, one deadline
Compute, products, annotation and infrastructure all converge on a single metric: ten thousand people in well-paid digital work by 2027.
An AI company that happens to be based where the question of how to live well is already old. The two halves are the same job.
Come in early