AI deeptech startup Zehn has announced the successful close of its $300,000 pre-seed funding round, valuing the company at a $15 million post-money, during the official launch of Area AI 1.0 in Dushanbe.
The round attracted backing from a mix of international and regional investors, including Battery Road Digital Holdings (US), FIFTYFIVE GROUP (Tajikistan), and Accelerate Prosperity (Tajikistan)—highlighting growing confidence in Zehn’s approach to solving one of artificial intelligence’s most persistent challenges.
Building AI for the edge cases that matter
Zehn focuses on outlier-centric AI modeling, targeting scenarios that traditional datasets often overlook. Working across large language models (LLMs) and computer vision, the company generates high-fidelity synthetic data to train AI systems for rare but critical real-world conditions—ranging from extreme weather events affecting autonomous systems to culturally specific dialects and linguistic nuances for language models.
This approach addresses a major reliability gap in modern AI, particularly in emerging and frontier markets, where global datasets frequently fail to capture local realities.
Local intelligence, global relevance
Zehn’s technology has already been applied in the development of SoroLLM, Tajikistan’s national language model, helping ensure accurate representation of local language, culture, and context. As demand for localized, culturally aware AI accelerates, Zehn positions itself at the intersection of deeptech innovation and regional relevance.
Recognition and ecosystem momentum
The company’s progress has been reinforced by a growing list of achievements, including a Top-15 finish at Constructor Start in Germany (from over 3,500 applicants), participation in the Misk Accelerator Program in Saudi Arabia, and winning AICA 2025 – Best AI Development. Zehn is also a licensed AI company under DIFC and a founding member of the Area AI Consortium, which aims to build the world’s first AI freezone in Tajikistan.
What’s next
The newly raised capital will be used to accelerate R&D and regional expansion, as enterprises and governments increasingly seek AI systems that perform reliably beyond the “average” case.
As AI adoption deepens globally, Zehn is betting that the future belongs to models trained not just on what is common—but on what truly matters when stakes are highest.
