IBM and the startup arm of India’s technology ministry signed a memorandum of understanding in New Delhi on October 13 that gives Indian deep-technology firms access to the company’s enterprise artificial-intelligence platforms, its data-sovereignty software, and technical guidance on quantum computing.
The agreement with MeitY Startup Hub, an initiative of the Ministry of Electronics and Information Technology, covers watsonx, IBM’s AI and data platform; IBM Bob, a coding agent built for enterprises; and Sovereign Core, an AI-ready software stack designed for organizations that must keep data inside a jurisdiction. Startups will also receive mentoring, technical skilling, and introductions to potential customers and investors.
The memorandum carries no funding, no target number of startups and no deadline, which is the usual shape of such agreements and also the thing that makes them hard to judge. IBM is competing in India against Microsoft, Google and Amazon Web Services for the loyalty of developers who will decide which AI platform their employers buy. Startups are the cheapest way to reach that audience.
“This collaboration reflects our shared commitment to fostering innovation and strengthening our country’s deep-tech startup ecosystem,” said Sandip Patel, managing director of IBM India and South Asia. He said the aim was to help founders “innovate faster, scale responsibly, and build solutions that address both national and global challenges.” Dr. Panneerselvam Madanagopal, chief executive of MeitY Startup Hub, said the partnership would strengthen indigenous capabilities and create opportunity for innovators building for what the government calls a developed India.
The focus says as much as the content. India’s startup economy is large and lopsided: consumer software, commerce and financial services soak up most of the capital, while deep technology — semiconductors, advanced materials, quantum devices, biotech hardware — attracts far fewer founders and far less patient money. Companies in those fields need laboratory time, engineering talent and a first enterprise customer, none of which a term sheet supplies. The MoU aims at the middle of that journey, pairing technical enablement with proof-of-concept work.
A second priority is geography. The plan singles out founders in tier-two and tier-three cities, part of an effort to spread an ecosystem that clusters in Bengaluru, Hyderabad, Pune and the national capital region. MeitY Startup Hub runs a nationwide network of incubators and has used similar arrangements with other vendors to reach outside those hubs.
Sovereign Core addresses a specific pressure. India’s data protection law, enacted in 2023, and the rules written under it have pushed companies to keep sensitive data under domestic control, and public-sector buyers have grown more explicit about where information may live. Software that lets an organization run AI models while meeting those constraints is an easier sale in New Delhi than a promise of scale alone.
Quantum is the longest-dated part of the agreement. India approved its National Quantum Mission in April 2023 with an outlay of about 6,003 crore rupees through 2030-31 and has set up four thematic research hubs, one of them devoted to computing. IBM has said it will deliver a fault-tolerant quantum computer, called Starling, by 2029, and its commercial argument depends on demonstrating that useful work exists before the hardware arrives. Startups in the program are prospective users of machines that do not yet exist.
The funding picture is less generous than the ambition. The Economic Times reported that the government halved the allocation for its IndiaAI Mission to 1,000 crore rupees in the 2026-27 budget from 2,000 crore the previous year, part of what officials described as a deliberate shift in sequencing. Fewer public rupees mean private vendors carry more of the cost of building an ecosystem, which is roughly how the IBM arrangement is structured.
IBM has been here before. In October 2023 it signed three memorandums with entities connected to the ministry covering AI, semiconductors and quantum technology. The company’s India operations include large research and engineering sites, and it has used those to anchor similar programs.
The other vendors are running comparable programs. Microsoft has attached its own AI skilling and startup support to Indian cloud regions, Google has funded AI research partnerships with Indian institutes, and Amazon has used credits to pull founders onto its cloud. Nvidia has gone further, tying chip supply and training programs to Indian firms and government labs. For a vendor, a memorandum costs almost nothing and puts its software in front of companies before they commit to a platform.
Inside the agreement, IBM Bob is the piece with the most immediate commercial relevance. Coding agents have become the busiest corner of enterprise AI, with Microsoft, Google, Anthropic and a crowd of startups selling tools that write and review software. India’s engineering services industry, which employs millions of programmers, is both the largest potential market for those tools and the group most exposed to them.
Semiconductors sit one step further out. India has committed roughly 76,000 crore rupees to its semiconductor mission and is building fabrication and testing plants in Gujarat and Assam, and the government wants a domestic supply of companies that can use what those plants produce. A design-tool and AI platform partnership is a small piece of that plan, but it is aimed at the layer where Indian firms could compete without owning a fab.
Memorandums of understanding are the cheapest instrument in technology policy, and their value shows up years later in products rather than in press releases. The test for this one is narrow and countable: how many participating startups turn a proof of concept into a paid enterprise contract, and whether any of them sell that product outside India. Everything else in the agreement is preparation.
