onllm.dev
Finance Minister Nirmala Sitharaman mentioned AI 11 times in her Budget 2026-27 speech — the highest count ever. But beyond rhetoric, the infrastructure commitments signal something concrete: India is actively subsidizing the compute stack.
If you're building with AI, the window is open.
The Compute Economics Are Shifting
Three things are happening simultaneously:
1. Hyperscalers are deploying at scale
- $90 billion in data centre investments announced
- Google committed $15 billion for a 1 GW facility in Visakhapatnam
- Tax holiday until 2047 for foreign cloud providers operating from India
2. Power infrastructure is being prioritized
- Basic customs duty waived on nuclear equipment
- Data centres getting policy attention as "critical infrastructure"
- The Economic Survey explicitly links clean baseload power to AI sustainability
3. The semiconductor supply chain is localizing
- ISM 2.0 launched with focus on equipment, materials, and Indian IP
- ECMS outlay nearly doubled to ₹40,000 crore (246 applications already in)
- 10 semiconductor projects worth ₹1.6 lakh crore approved across 6 states
Current data centre capacity sits at 1,280 MW with a 4 GW target by 2030. That's 3x growth in four years — and most of it AI-optimized.
What Builders Actually Get
| Initiative | Allocation | Impact |
|---|---|---|
| IndiaAI Mission | ₹1,000 crore (FY27) | Shared compute infrastructure, datasets platform |
| AI Centres of Excellence | ₹500 crore | Research partnerships, applied AI |
| Semiconductor Mission 2.0 | ₹40,000 crore | Domestic chip ecosystem, reduced import dependency |
| ECMS expansion | ₹40,000 crore | Component manufacturing, supply chain resilience |
| IT safe harbour threshold | Raised to ₹2,000 crore | Tax certainty for services companies |
The 5-year IndiaAI Mission total: ₹10,371 crore — covering compute, datasets, innovation centres, and startup financing.
The Talent Pipeline Is Expanding
- 15,000 AI labs in schools
- 10,000 technology fellowships at IITs
- AI training expanding to 500 universities
- 890+ active GenAI startups as of H1 2025 (tripled from 240 in prior year)
Deep-tech funding rose 78% in CY24. Karnataka hosts 39% of GenAI startups, Maharashtra 14%.
The Policy Stance Is Clear
The Economic Survey advocated for a bottom-up, application-led AI strategy — sector-specific models over chasing scale. This aligns with how most Indian builders work: solving real problems with constrained resources.
Key Survey positions:
- Shared compute infrastructure should expand
- Open and interoperable systems preferred
- Sector-specific, smaller models over massive foundation models
- Risk-based, proportionate regulation
Translation: The government is not trying to compete with OpenAI. It's trying to make it easier for you to build useful AI products.
Why Now
Latency drops as hyperscalers deploy Indian infrastructure. Compliance simplifies with unified safe harbour margins. Power gets cheaper as nuclear and renewable capacity expands. Talent becomes accessible as training programs scale.
The GPU bottleneck remains real — global supply constraints don't disappear with policy. But the surrounding infrastructure (power, policy, talent, capital) is being actively built.
The AI Impact Summit (February 16-20, 2026, New Delhi) brings 15 heads of state and major tech CEOs. Expect concrete deployment announcements.
For Indian AI builders, the policy environment has never been more aligned with execution. The hardware constraint exists everywhere. The difference now: everything else is being solved.
Sources: Union Budget 2026-27 Speech, Economic Survey 2025-26, Ministry of Electronics & IT