Venture capital investment in AI startups in India has nearly doubled in two years. From $438 million in 2023, it grew to $832 million in 2025, as per Tracxn data. Since then, it has gathered even greater momentum.
The capital flowed into early stage AI ventures in the first three months of 2026 would amount to three-fourths of the total of the entire previous year. At this rate, VC investment in AI startups in India would triple this year.
Beneath this sharp shift of capital towards AI lies an exasperating fight against obsolescence as increasingly powerful foundational models are threatening to reduce these startups’ innovations into easily commoditised features.
The unique selling propositions that made them attractive just three years ago are quickly becoming worthless. Foundational models become more potent, compute costs decline and research labs move up the value chain, forcing startups to pivot at a breathtaking speed to stay relevant.
Two-thirds of the Indian GenAI startups pivoted as of 2025, a Nasscom study found. Most have shifted focus towards applications, primarily SaaS products with AI built into them or industry-specific ‘vertical AI’ solutions. Sure enough, that brought them an immediate product-market fit and commercialisation opportunities.
Investment thesis shifts
Amid this perpetual flux and churn, the venture capital industry’s AI investment theme too has sharpened. If two years ago most startups and investments were centred around chatbots, small language models and vertical AI SaaS products, the newer ventures have assumed more complexity and built more defensible moats.
“In 2023, a lot of what was getting funded was essentially a wrapper around GPT-4 with a vertical domain label on it,” said Hemant Mohapatra, Partner, Lightspeed. Wrappers - adding a layer of specific use cases such as legal drafting or sales emails - sprouted up by the dozens. “Some of those companies have done fine, but many discovered they had a feature, not a company, once the foundation models caught up to what they were doing.”
Things have markedly evolved from there. Today, capital is concentrated in three areas - “AI infrastructure and tooling, India-specific foundation models, and applied AI in sectors like healthcare, legal, and financial services where regulatory complexity and proprietary data create real moats,” Mohapatra said.
Prayank Swaroop, partner at Accel, said that in 2024, most AI startups were still early-stage experiments, with a majority of ideas built as wrappers on top of foundational models across sectors such as marketing, hiring and edtech, and only about 10% offering truly differentiated propositions. “Now, that number has risen to nearly 40% as founders better understand how AI can be applied across sectors,” he said.
Pivot or perish
The rapid pace of change has forced many startups to go back to the drawing board. Vardhan Dharnidharka, Principal, Stellaris Venture Partners said that companies that started in the earlier era of GenAI are having to pivot as LLMs become increasingly capable of replacing them.
For instance, in the software development space, startups focused on different slices – testing, writing PRDs (product requirement document), code generation, etc.
“The space has now become close to a monopoly, with Claude Code making many vertical and early stage horizontal plays obsolete,” Dharnidharka explained. Under this existential worry pivoting becomes the way out.
“Several chatbot companies pivoted to become AI automation platforms. SaaS players embedded copilots into their workflows. Infrastructure startups repositioned around AI agents. Those that treated their original model as permanent, rather than as a starting point, found it harder to stay relevant,” said Ankur Mittal, Partner Physis, cofounder, Inflection Point Ventures.
As for vertical AI SaaS startups, life has become a battle to defend themselves against the disruption caused by Anthropic’s plugins, and OpenAI Frontier.
“If your vertical SaaS business is primarily a workflow orchestration layer on top of someone else’s data and someone else’s model, you are extremely exposed,” Mohapatra said, warning that features like Anthropic plugins can commoditize what took you 18 months to build. “We have seen this happen, and it will keep happening.”
However, startups with strong data moats are likely to benefit. “If you have spent three years ingesting hospital records or building a comprehensive dataset of Indian court filings, no plugin changes that,” he added.
Evidently, domain depth becomes one’s defense. “We believe that while the likes of Anthropic will get smarter in their models and capture a significant portion of the horizontal market, the companies that build moats in their vertical domains will be highly relevant,” said Dharnidharka of Stellaris. Successful vertical AI SaaS companies are leveraging AI as a wedge and building multi-agent, deep vertical-specific workflows, he added.
Outlook
With enterprise AI spending expected to reach $300 billion by 2027 globally, investors believe the opportunity for Indian startups remains significant. “India’s advantage lies in applied AI and enterprise solutions - and the capital markets are beginning to price that in,” Inflection’s Mittal said.
India’s 1.4 billion people, 22 official languages and the massive informal-sector economies in fintech, healthcare, agriculture, are its tailwind, Lightspeed’s Mohapatra believes.
The capital flowed into early stage AI ventures in the first three months of 2026 would amount to three-fourths of the total of the entire previous year. At this rate, VC investment in AI startups in India would triple this year.
Beneath this sharp shift of capital towards AI lies an exasperating fight against obsolescence as increasingly powerful foundational models are threatening to reduce these startups’ innovations into easily commoditised features.
The unique selling propositions that made them attractive just three years ago are quickly becoming worthless. Foundational models become more potent, compute costs decline and research labs move up the value chain, forcing startups to pivot at a breathtaking speed to stay relevant.
Two-thirds of the Indian GenAI startups pivoted as of 2025, a Nasscom study found. Most have shifted focus towards applications, primarily SaaS products with AI built into them or industry-specific ‘vertical AI’ solutions. Sure enough, that brought them an immediate product-market fit and commercialisation opportunities.
Investment thesis shifts
Amid this perpetual flux and churn, the venture capital industry’s AI investment theme too has sharpened. If two years ago most startups and investments were centred around chatbots, small language models and vertical AI SaaS products, the newer ventures have assumed more complexity and built more defensible moats.
“In 2023, a lot of what was getting funded was essentially a wrapper around GPT-4 with a vertical domain label on it,” said Hemant Mohapatra, Partner, Lightspeed. Wrappers - adding a layer of specific use cases such as legal drafting or sales emails - sprouted up by the dozens. “Some of those companies have done fine, but many discovered they had a feature, not a company, once the foundation models caught up to what they were doing.”
Things have markedly evolved from there. Today, capital is concentrated in three areas - “AI infrastructure and tooling, India-specific foundation models, and applied AI in sectors like healthcare, legal, and financial services where regulatory complexity and proprietary data create real moats,” Mohapatra said.
Prayank Swaroop, partner at Accel, said that in 2024, most AI startups were still early-stage experiments, with a majority of ideas built as wrappers on top of foundational models across sectors such as marketing, hiring and edtech, and only about 10% offering truly differentiated propositions. “Now, that number has risen to nearly 40% as founders better understand how AI can be applied across sectors,” he said.
Pivot or perish
The rapid pace of change has forced many startups to go back to the drawing board. Vardhan Dharnidharka, Principal, Stellaris Venture Partners said that companies that started in the earlier era of GenAI are having to pivot as LLMs become increasingly capable of replacing them.
For instance, in the software development space, startups focused on different slices – testing, writing PRDs (product requirement document), code generation, etc.
“The space has now become close to a monopoly, with Claude Code making many vertical and early stage horizontal plays obsolete,” Dharnidharka explained. Under this existential worry pivoting becomes the way out.
“Several chatbot companies pivoted to become AI automation platforms. SaaS players embedded copilots into their workflows. Infrastructure startups repositioned around AI agents. Those that treated their original model as permanent, rather than as a starting point, found it harder to stay relevant,” said Ankur Mittal, Partner Physis, cofounder, Inflection Point Ventures.
As for vertical AI SaaS startups, life has become a battle to defend themselves against the disruption caused by Anthropic’s plugins, and OpenAI Frontier.
“If your vertical SaaS business is primarily a workflow orchestration layer on top of someone else’s data and someone else’s model, you are extremely exposed,” Mohapatra said, warning that features like Anthropic plugins can commoditize what took you 18 months to build. “We have seen this happen, and it will keep happening.”
However, startups with strong data moats are likely to benefit. “If you have spent three years ingesting hospital records or building a comprehensive dataset of Indian court filings, no plugin changes that,” he added.
Evidently, domain depth becomes one’s defense. “We believe that while the likes of Anthropic will get smarter in their models and capture a significant portion of the horizontal market, the companies that build moats in their vertical domains will be highly relevant,” said Dharnidharka of Stellaris. Successful vertical AI SaaS companies are leveraging AI as a wedge and building multi-agent, deep vertical-specific workflows, he added.
Outlook
With enterprise AI spending expected to reach $300 billion by 2027 globally, investors believe the opportunity for Indian startups remains significant. “India’s advantage lies in applied AI and enterprise solutions - and the capital markets are beginning to price that in,” Inflection’s Mittal said.
India’s 1.4 billion people, 22 official languages and the massive informal-sector economies in fintech, healthcare, agriculture, are its tailwind, Lightspeed’s Mohapatra believes.
