IndiaAI Mission - State-Backed AI Startup Funding + Subsidized GPU Pool (38,000+ GPUs)
Source: https://indiaai.gov.in/hub/indiaai-startup-financing
Description
Create account to comment on specific lines or Sign in
+ 1 The IndiaAI Mission is a ₹10,371.92 crore (~$1.25 B) Government of India programme run by MeitY (IndiaAI Independent Business Division). It bundles two big things builders care about: (1) heavily subsidised AI compute on a national pool of 38,000+ GPUs at ₹65/hour for A100-class and ₹92/hour for H100-class (40% subsidy by default, 100% subsidy for selected foundational-model builders), and (2) risk capital through the IndiaAI Startup Financing pillar plus a separate $1.1 B state-backed venture capital Fund of Funds approved by the Cabinet in February 2026 (channelled via SEBI-registered AIFs into deep-tech / AI / advanced manufacturing startups).
No comments on this line yet.
+ 2
No comments on this line yet.
+ 3 This is India-only: you need an Indian-registered entity (DPIIT-recognised startup, MSME, or research institution) and Indian KYC (Aadhaar / Digilocker / ePramaan) to even sign in to the compute portal. If you qualify, it is one of the cheapest ways on the planet to train a serious model.
No comments on this line yet.
+ 4
No comments on this line yet.
+
5
No comments on this line yet.
+ 6
No comments on this line yet.
+ 8
No comments on this line yet.
+ 9 Don't confuse the two main offerings — they have very different processes:
No comments on this line yet.
+ 10
No comments on this line yet.
+ 11 TrackWhat you getHow to applyDifficulty
No comments on this line yet.
+ 12 IndiaAI Compute PortalSubsidised GPU hours (₹65/hr A100, ₹92/hr H100)Self-serve portal sign-up + KYC, auto-approval ≤5,000 GPU hoursEasy if Indian
No comments on this line yet.
+ 13 Foundational Model Call for Proposals100% free compute (no ₹/hr charge) for building Indic LLMs / foundation modelsCompetitive RFP — select few startups (e.g. Sarvam, Soket, Gnani, Gan, Fractal)Very hard
No comments on this line yet.
+ 14 Startup Financing PillarEquity / risk capital for deep-tech AI startupsOnline application via IndiaAI portal, multi-stage screeningHard
No comments on this line yet.
+ 15 $1.1 B VC Fund of FundsIndirect VC funding (LP into AIFs)Apply to private AIFs that receive the LP money — not direct from govtHard
No comments on this line yet.
+ 16 Global Acceleration ProgrammeSoft-landing in Europe via Station F + HEC Paris (10 startups/cohort)Cohort applicationHard
No comments on this line yet.
+ 17
No comments on this line yet.
+
18
No comments on this line yet.
+ 19
No comments on this line yet.
+ 21
No comments on this line yet.
+ 22 • Indian entity registered in India is mandatory for everything below.
No comments on this line yet.
+ 23 • DPIIT-recognised startup, MSME, registered research institution, academic faculty, or government body.
No comments on this line yet.
+ 24 • For deep-tech startup definition (revised 2026): up to 20 years from incorporation, turnover up to ₹300 crore, demonstrable R&D intensity.
No comments on this line yet.
+ 25 • Founders/team need to authenticate via Meri Pehchaan (Digilocker / Parichay / ePramaan) — i.e., Indian KYC. Foreign founders without an Indian entity cannot use the compute portal.
No comments on this line yet.
+ 26 • For the foundational model 100% subsidy: only a handful of startups have been selected via formal RFP rounds; not a self-serve track.
No comments on this line yet.
+ 27
No comments on this line yet.
+ 28 If you are a solo international vibe coder, this is not for you — look at NVIDIA Inception, Modal credits, or generic startup programs instead.
No comments on this line yet.
+ 29
No comments on this line yet.
+
30
No comments on this line yet.
+ 31
No comments on this line yet.
+ 33
No comments on this line yet.
+ 34 This is the part most Indian builders/startups will actually use.
No comments on this line yet.
+ 35
No comments on this line yet.
+ 37 • 38,000+ GPUs already provisioned across 14 empanelled cloud service providers (Yotta, Tata, Jio, CtrlS, E2E, NxtGen, etc.). Original target was 10,000; over-delivered. Another ~20,000 GPUs being added.
No comments on this line yet.
+ 38 • Mix of NVIDIA H100, H200, A100, L40S, L4, MI300X, plus newer GB200 / B200 announced post-AI Impact Summit 2026.
No comments on this line yet.
+ 39
No comments on this line yet.
+ 41 GPU classEnd-user price
No comments on this line yet.
+ 42 A100 / equivalent₹65 / GPU-hour (~$0.78)
No comments on this line yet.
+ 43 H100 / equivalent₹92 / GPU-hour (~$1.10)
No comments on this line yet.
+ 44
No comments on this line yet.
+ 45 For reference, on-demand H100 in the US is typically $2.50–$4/hr — so this is roughly 1/3 the global rate.
No comments on this line yet.
+ 46
No comments on this line yet.
+ 48 Selected startups (Sarvam AI, Soket AI Labs, Gnani.ai, Gan.ai, Fractal — list grows each round) receive 100% subsidy = compute is fully free for the duration of the build. You apply through specific Call for Proposals rounds for foundational/sovereign LLM work — not the standard portal flow.
No comments on this line yet.
+ 49
No comments on this line yet.
+
50
No comments on this line yet.
+ 51
No comments on this line yet.
+ 53
No comments on this line yet.
+ 54 1. Go to the IndiaAI Compute Portal linked from indiaai.gov.in/hub/indiaai-compute-capacity.
No comments on this line yet.
+ 55 2. Sign in via Meri Pehchaan using one of: Digilocker, Parichay, or ePramaan account (Indian government SSO).
No comments on this line yet.
+ 56 3. Pick your applicant category — Startup, MSME, Research / Academic, Student, IndiaAI Fellowship awardee, or Government / PSU.
No comments on this line yet.
+ 57 4. Fill the eligibility form and upload supporting docs (DPIIT certificate for startups, Udyam for MSMEs, institution affiliation letter for research, etc.).
No comments on this line yet.
+ 58 5. Submit a compute request specifying GPU type, number, and total GPU-hours needed.
No comments on this line yet.
+ 59 6. Approval rules:
No comments on this line yet.
+ 60 • ≤ 5,000 GPU-hours → auto-approved (instant or near-instant).
No comments on this line yet.
+ 61 • > 5,000 GPU-hours → manual review by the PMEC committee (Programme Management & Evaluation Committee). Expect days to a couple of weeks.
No comments on this line yet.
+ 62 7. Once approved, you get a voucher / allocation redeemable on the empanelled CSP of your choice. Use it like normal cloud GPU time — but billed at the subsidised rate.
No comments on this line yet.
+ 63
No comments on this line yet.
+ 64 Important:
No comments on this line yet.
+ 65 • You are paying — just heavily subsidised. This is not "free credits" in the way AWS Activate is; you still need a payment method on the empanelled CSP for whatever portion isn't subsidy.
No comments on this line yet.
+ 66 • For foundational-model 100% subsidy, the allocation is granted upfront, no out-of-pocket spend.
No comments on this line yet.
+ 67 • One approved compute request typically has a time window (use within the validity, then re-apply).
No comments on this line yet.
+ 68 • Non-Indian payment cards / non-Indian billing addresses won't work on most empanelled CSPs.
No comments on this line yet.
+ 69
No comments on this line yet.
+
70
No comments on this line yet.
+ 71
No comments on this line yet.
+ 73
No comments on this line yet.
+ 74 For deep-tech AI startups looking for funding rather than compute. Process:
No comments on this line yet.
+ 75
No comments on this line yet.
+ 76 1. Apply through the IndiaAI portal (online only).
No comments on this line yet.
+ 77 2. Stage 1 — Initial screening: Eligibility, completeness, quality of proposal.
No comments on this line yet.
+ 78 3. Stage 2 — Technical evaluation: Innovation, technical feasibility, responsible AI practices, team quality.
No comments on this line yet.
+ 79 4. Stage 3 — PoC review & shortlisting: Demo / proof-of-concept evaluation by expert panel.
No comments on this line yet.
+ 80 5. Selected startups receive funding (size varies per round) plus mentorship and ecosystem support.
No comments on this line yet.
+ 81
No comments on this line yet.
+ 83 Approved by the Indian Cabinet in February 2026. This is a fund-of-funds — government commits capital to private SEBI-registered Alternative Investment Funds (AIFs); those AIFs invest in startups. So you don't apply to the government for this money directly — you raise from one of the participating AIFs (which will be announced as deployment proceeds). Targets AI, deep-tech, advanced manufacturing.
No comments on this line yet.
+ 84
No comments on this line yet.
+
85
No comments on this line yet.
+ 86
No comments on this line yet.
+ 88
No comments on this line yet.
+ 89 Run in partnership with Station F (Paris) and HEC Paris. Each cohort selects 10 Indian AI startups for soft-landing into European markets — workspace at Station F, HEC Paris mentorship, EU partner introductions. Apply through periodic open calls announced on the IndiaAI portal.
No comments on this line yet.
+ 90
No comments on this line yet.
+
91
No comments on this line yet.
+ 92
No comments on this line yet.
+ 94
No comments on this line yet.
+ 95 For an Indian-registered AI startup:
No comments on this line yet.
+ 96
No comments on this line yet.
+ 97 • ₹65/hr A100 = roughly $0.78/hr vs ~$2/hr market → ~60% saving over commercial cloud.
No comments on this line yet.
+ 98 • ₹92/hr H100 = roughly $1.10/hr vs ~$3–4/hr market → ~65–70% saving.
No comments on this line yet.
+ 99 • Foundational model track: literal $0 compute spend during build — only viable for serious LLM/foundation model teams that can win the RFP.
No comments on this line yet.
+ 100
No comments on this line yet.
+ 101 For 1,000 H100-hours (a realistic small fine-tune):
No comments on this line yet.
+ 102 • Commercial: ~$3,000
No comments on this line yet.
+
103
• IndiaAI subsidised: ₹92,000 ($1,100)
No comments on this line yet.
+ 104 • Foundational model awardee: ₹0
No comments on this line yet.
+ 105
No comments on this line yet.
+
106
No comments on this line yet.
+ 107
No comments on this line yet.
+ 109
No comments on this line yet.
+ 110 • Indian KYC required. Aadhaar-linked Digilocker is the smoothest path. Foreign founders cannot self-onboard.
No comments on this line yet.
+ 111 • Compute is allocated, not gifted. Auto-approval works for small jobs (<5,000 GPU-hours), but anything bigger goes through committee — plan for delays.
No comments on this line yet.
+ 112 • Empanelled CSP variability. GPU availability and exact pricing vary by CSP (Yotta, Jio, Tata, etc.). Check which CSP has your preferred GPU type before requesting.
No comments on this line yet.
+ 113 • Foundational model 100% subsidy is highly competitive. ~500+ proposals received in early rounds; only a handful selected. Don't bank on it for your roadmap.
No comments on this line yet.
+ 114 • The $1.1 B VC fund money is intermediated through private AIFs — you don't apply to the government for cheques.
No comments on this line yet.
+ 115 • Programme runs through the Mission's term (notionally 5 years from 2024), so timelines / amounts may shift; always check the current portal state.
No comments on this line yet.
+ 116 • Region lock: Compute is hosted in Indian data centres — latency-sensitive workloads serving non-India users should account for this.
No comments on this line yet.
+ 117
No comments on this line yet.
+
118
No comments on this line yet.
+ 119
No comments on this line yet.
+ 121
No comments on this line yet.
+ 122 For the standard compute subsidy track, getting onto the portal as a DPIIT-recognised Indian startup is fairly mechanical (rank ~3 if you're already registered in India). But the whole programme as listed — including the headline foundational-model 100% subsidy and the $1.1 B fund — is gated behind institutional eligibility (Indian entity, formal recognition, competitive proposal review, AIF intermediation). For most readers of a global "free AI credits" list, the practical rank is 4 (Hard) because of the strict India-only eligibility and competitive nature of the headline benefits.
No comments on this line yet.
+ 123
No comments on this line yet.
+
124
No comments on this line yet.
+ 125
No comments on this line yet.
+ 127
No comments on this line yet.
+ 128 • If you only need general AI credits and you're not Indian, look elsewhere (this isn't for you).
No comments on this line yet.
+ 129 • If you are an Indian deep-tech founder: get DPIIT recognition first — it unlocks IndiaAI, plus SISFS, plus VC partner programmes.
No comments on this line yet.
+ 130 • Combine IndiaAI compute subsidy with Sarvam AI's startup credits and Bhashini datasets if you're building Indic-language AI.
No comments on this line yet.
+ 131 • Watch IndiaAI's Call for Proposals announcements — that's where new tracks (foundational models, applications, safety, datasets) open up periodically.
No comments on this line yet.
+ 132 • The IndiaAI Innovation Challenge runs separate funding + government contract opportunities — different application track from the financing pillar.
No comments on this line yet.
+ 133
No comments on this line yet.
+
134
No comments on this line yet.
+ 135
No comments on this line yet.
+ 136 Sources:
No comments on this line yet.
+ 137 • IndiaAI Startup Financing (official)
No comments on this line yet.
+ 138 • IndiaAI Compute Capacity (official)
No comments on this line yet.
+ 139 • PIB — Cabinet approves ₹10,300+ crore for IndiaAI Mission
No comments on this line yet.
+ 140 • PIB — Common computing facility / sovereign AI announcement
No comments on this line yet.
+ 141 • TechCrunch — India approves $1.1B state-backed VC fund (Feb 2026)
No comments on this line yet.
+ 142 • TechDodo — IndiaAI 38,000 free/subsidised GPUs guide (2026)
No comments on this line yet.
+ 143 • TechDodo — Rs 67/hour compute access guide
No comments on this line yet.
+ 144 • SME Futures — Rs 65/hour subsidy across 14 service providers
No comments on this line yet.
+ 145 • Outlook Business — 100% compute subsidy for foundational AI models
No comments on this line yet.
+ 146 • Inc42 — 100% compute subsidy to Sarvam et al.
No comments on this line yet.
+ 147 • NVIDIA Blog — India AI Mission Infrastructure
No comments on this line yet.
+ 148 • Storyboard18 — IndiaAI adds 34,000+ GPUs, 500+ proposals
No comments on this line yet.
+ 149 • DD News — ₹10,300 crore mission, 38,000 GPUs
No comments on this line yet.