IndiaAI Mission (Rs 10,371.92 crore national AI programme)

The Union Cabinet approved the IndiaAI Mission on 7 March 2024 with a five-year outlay of Rs 10,371.92 crore.¹¹⁰ It followed NITI Aayog's 2018 national strategy, which proposed research centres, a shared compute platform and a data marketplace but carried no budget of its own (see the separate entry on that strategy).

The Mission is implemented by IndiaAI, an Independent Business Division of the Digital India Corporation (DIC), a body under MeitY.¹⁰¹² The Rajya Sabha reply of 13 February 2026 describes its purpose as the development of "the overall AI ecosystem in the country".¹ MeitY's governance guidelines describe the problem it is meant to address as the concentration of AI resources (compute, data and models) in a few global players, and say the Mission aims to counter this by democratising the benefits of AI.⁷

The Mission has seven pillars and an overheads line. The allocations below are from the government's own reply to the Rajya Sabha; the shares are calculated by the author:¹³

The compute pillar was initially planned around 10,000 GPUs; planned capacity rose to 18,693 and then to more than 38,000 (see the entry on IndiaAI compute capacity).⁵⁷ The datasets pillar is meant to make public-sector datasets AI-ready, and the safety pillar funds technical work on machine unlearning, bias mitigation, privacy-enhancing tools and explainable AI.³⁷

Budget path. The 2024-25 Budget Estimate was Rs 551.75 crore, which was revised down to Rs 173 crore, a 69% cut. For 2025-26 the Budget Estimate was Rs 2,000 crore, revised to Rs 800 crore, and for 2026-27 it is Rs 1,000 crore.¹⁵ Parliamentary data show that the Mission had received Rs 21.79 crore of the 2024-25 revised estimate and Rs 379.15 crore of the 2025-26 revised estimate as on 9 February 2026, about 3.9% of the five-year outlay in total.¹⁴

MediaNama's analysis noted that the first-year cut might reflect aggressive bidding for GPUs, a phased rollout or reallocation (for example to the Rs 500 crore Centre of Excellence for AI in education announced in the 2025 Budget), though it said these explanations were unconfirmed.⁵

Institutions. MeitY is the nodal ministry, and IndiaAI, under the Digital India Corporation, has its own Chief Operating Officer.⁷¹² The Safe and Trusted AI pillar hosts the IndiaAI Safety Institute, which the governance guidelines recommend be resourced to provide technical expertise on AI safety.⁷ A drafting committee that MeitY constituted in July 2025, building on an advisory group under the Principal Scientific Adviser, produced the governance guidelines.⁷ In April 2026 MeitY set up an AI Governance and Economic Group, chaired by the Minister of Electronics and IT, to coordinate AI policy; it includes ministers and senior officials but not the sectoral regulators that the guidelines had proposed.⁹

Compute. Planned GPU capacity rose from 10,000 to 18,693 by early 2025.⁵ By 31 August 2025 more than 38,231 GPUs were available, and a secure cluster for 3,000 next-generation GPUs was being built for sovereign and strategic uses.⁷ By February 2026 the government reported more than 38,000 GPUs and 14 empanelled providers in Mumbai, Navi Mumbai, Hyderabad, Bengaluru, Noida and Jamnagar. Users of the IndiaAI Compute Portal included 114 academic researchers, 47 startups and MSMEs, 36 early-stage startups, 58 government entities, 32 students, 10 early-stage researchers and 8 IndiaAI Fellows.¹ In October 2026 MeitY's C-DAC issued a fresh request for empanelment for up to 10,000 AI compute units.¹¹

Foundation models. A call for proposals opened in February 2025. Sarvam AI was the first to be selected, in April 2025, and by March 2026 twelve organisations and consortia were funded, with support covering mainly compute costs. BharatGen, an IIT Bombay-led consortium, received Rs 1,058.52 crore and Sarvam AI Rs 246.72 crore. Models from Sarvam, BharatGen and Gnani were launched at the India AI Impact Summit in February 2026.²

Datasets, applications and skills. AIKosh, the national datasets platform, and the IndiaAI Compute Portal were launched on 6 March 2025 along with an AI competency framework for public officials and an AI start-up acceleration programme with Station F and HEC Paris.⁶ By August 2025 the datasets platform had onboarded about 1,500 datasets and 217 models; 30 sectoral applications had reached the prototyping stage; and the skilling programmes supported more than 500 PhD fellows, 8,000 undergraduates and 5,000 postgraduates.⁷ In February 2026 the government reported 27 India Data and AI Labs set up and 543 more identified.¹

Safety. Under the Safe and Trusted AI pillar the Mission has supported technical work on machine unlearning, bias mitigation, privacy-enhancing tools and explainable AI, and has established the IndiaAI Safety Institute (see the separate entry).⁷

Compute-heavy design. At a MediaNama roundtable on 11 December 2024, participants asked whether a 44% compute share was justified, one arguing that datasets create lasting assets while compute loses value after use, and another calling 44% "probably the bare minimum" given growing inference demand. A participant from the Takshashila Institution said how funds are spent matters as much as the total and criticised tender-based allocation.⁸

Slow releases. Only about Rs 401 crore of the Rs 10,371.92 crore outlay had been received by February 2026, after the 2024-25 and 2025-26 revised estimates were cut. MediaNama reported that this raised doubts about the Mission's ability to operate at scale.⁴

Scale against private investment. MediaNama noted that private commitments from Amazon and Microsoft of more than USD 50 billion far exceed public spending under the Mission, which could shape who controls compute capacity and innovation pipelines.⁴

Safety funding and opacity. The safe and trusted AI pillar's Rs 20.46 crore is about five times smaller than the Mission's overheads line. MediaNama also noted that the December 2024 breakdown did not explain how dataset quality, curation or anonymisation would be ensured, nor whether startup financing would be grants, equity or loans.³⁴ The reply that lists users of the compute portal gives counts of users, not utilisation rates, and MediaNama noted that access counts are not the same as usage.¹⁴

As of 15 October 2026 the Mission is ongoing. The last parliamentary data in the sources reviewed (9 February 2026) show Rs 400.94 crore received against a Rs 1,000 crore Budget Estimate for 2026-27 that had not yet drawn funds. The government has not announced a second phase or a change to the Rs 10,371.92 crore outlay in the sources reviewed.¹⁴

Recent developments include the April 2026 AI Governance and Economic Group, the first launches from the foundation-model programme, and MeitY's October 2026 call for further GPU providers.⁹²¹¹ The Mission was approved for five years from 2024.