Google’s new programme to support Indian AI startups has prompted analysts to caution that it may only increase reliance on Big Tech, even as a top executive at the technology behemoth said that the initiative will not involve direct capital infusion or equity stakes.
The market-access programme, announced last week, is aimed at helping 30 Indian AI startups per cohort beyond the seed stage, and it’s the “first of its kind that we’re launching in India,” Seema Rao, managing director for top partners India and corporate development at Google, told Mint.
“It is structured around three pillars—building an enterprise-ready playbook, enabling direct access to customers, and global immersion,” said Rao. “The idea is to help startups develop a global sales muscle, engage with enterprise buyers, convert conversations into pilots and contracts, and gain exposure to markets like India and the US.”
India’s AI opportunity is estimated to reach $17 billion by 2027, according to a report by Boston Consulting Group (BCG). And the government expects it to add around $1.7 trillion to India’s economy by 2035. Several global and Indian behemoths have already promised to expand data centre capacity, anticipating an explosion in compute demand.
The government launched a $1.2-billion AI Mission for early-stage startups last year to improve visibility and ease procurement hurdles. However, it does not address later-stage challenges such as global enterprise sales, long procurement cycles or scaling pilots into commercial contracts. It is this gap that Google aims to target.
“Startups don’t fail because their models are weak; they fail because of issues around pricing, packaging, procurement readiness and narrative,” said Jibu Elias, an AI policy expert. “With enterprises, the challenge isn’t finding a vendor, it’s building trust at scale, especially around data, liability, security and integration.”
Startups also face hurdles including weak data pipelines and misaligned expectations fuelled by hype around large language models.
“AI products/services still remain misunderstood by B2B and B2C market segments, which downgrades perceived expectations of AI solutions,” said Abhivardhan, president, Indian Society of Artificial Intelligence and Law, an industry forum.
A ‘customer acquisition’ strategy?
Google’s offer to help startups break through this phase has prompted analysts to suggest caution. While this may help startups break out of pilot-stage stagnation, it embeds startups into Google’s cloud and AI stack, shifts bargaining power toward the tech giant, and gives the company early visibility into promising competitors, they said.
“This is ultimately a customer acquisition strategy,” said Isha Suri, independent Research and Global AI and Market Power Fellow at European AI Society Fund. “Even where there is capital involved, money often doesn’t really change hands—it is typically wrapped into cloud credits and platform access.”
Once startups are onboard a cloud service provider or start building on a particular model stack, switching becomes extremely difficult.
“This creates stickiness, and over time that lock-in gives the platform far greater negotiating power, including the ability to change pricing, as moving out later from the cloud ecosystem is nearly impossible,” she said.
There is also a competitive logic at play—keeping an eye on potential rivals early, when they are still small. “At that stage, large platforms either end up acquiring them or outcompeting them,” said Elias.
Data sovereignty caution
Google is alsomaking a major infrastructure bet in India, planning a $15-billion artificial intelligence hub and gigawatt-scale data centre campus in Visakhapatnam, Andhra Pradesh—its largest investment in the country—backed by state government incentives including land, tax concessions and subsidised electricity.
As startups are trained on Google’s preferred architectures, security practices, deployment models and compliance norms, these patterns can quietly become default expectations for enterprise AI adoption, making Google’s stack the baseline rather than one option among many.
“If Indian startups that become successful global AI companies are trained, deployed and scaled using Google’s infrastructure and tools, Google embeds itself into the future AI supply chain,” said Sohom Banerjee, senior research associate at CUTS International. “Whoever owns the developer and startup ecosystem then has disproportionate influence over future enterprise adoption.”
In response to concerns around data sovereignty, Rao said Google’s cloud offerings are designed to comply with local regulations.
“We offer a range of solutions that meet local regulatory requirements, including a sovereign stack and private data cloud,” she said. “Across our cloud offerings, we have architectures that address data residency and local compliance needs, allowing us to partner with startups across different sectors and use cases—from infrastructure and models to applications—depending on what they require.”
But analysts see sovereignty concerns go beyond data localization.
“Sovereignty is about autonomy—being able to operate in open ecosystems without being locked into a single platform, and having the ability to switch without artificial entry or exit barriers,” said Suri.
Sovereignty should not be reduced to nationalist rhetoric or localisation alone, according to Suri, but grounded in systems that are privacy-preserving and designed to serve the public interest rather than entrench domestic private control.
