Global spend on vertical AI applications to hit $47 billion by 2030, Indian startups poised to gain: Upekkha Report

Many of these startups are already seeing enterprise traction.

VaultEdge handles thousands of mortgage documents daily for U.S. lenders, DeepTek helps hospitals cut radiology turnaround time, while Credgenics is used by Indian NBFCs to automate loan recovery. In manufacturing, firms like Frinks and UptimeAI are helping reduce unplanned downtime, which costs the industry over $50 billion annually.

"We haven’t written off horizontal AI. But in that space, large incumbents like Microsoft and Amazon have too much of an edge. It’s difficult for startups to compete unless they’re building to be acquired," said Prasanna Krishnamoorthy, Managing Partner at Upekkha said in a media briefing.

“From India, our strength lies in verticals, domain expertise, strong AI talent, and contextual understanding. That’s where the real opportunity is," he added.

Fueling this momentum is a dramatic drop in AI infrastructure costs. Foundational model training, which cost $100,000 just a year ago, can now be done for $10,000, making vertical AI more affordable and scalable.

India, the report notes, has a structural advantage: over 100,000 engineers with experience in both U.S. enterprise workflows and Indian operational constraints, allowing them to build tailored solutions at a fraction of Silicon Valley costs. Several startups are already following a “built in India, sold globally” model.

In the next two years, Upekkha expects three shifts to accelerate adoption—cloud providers subsidising AI compute, local language models opening new markets, and the emergence of agent framework standards.

With enterprises moving from experimentation to execution, and CFOs demanding ROI from AI investments, vertical AI is emerging as the next big opportunity, and Indian founders are well-positioned to lead the charge.