Union Budget 2026: From GPUs to power, what the AI ecosystem expects

As India prepares for the Union Budget 2026, the narrative is no longer just about being a larger participant in global trade but about becoming a shaping force within it—leveraging artificial intelligence (AI) as the primary power lifter.

Industry experts, founders, and investors expect the upcoming budget move beyond symbolic gestures and provide the structural plumbing necessary for a sovereign AI ecosystem, including compute access, cybersecurity resilience, and government-led demand.

Building the backbone

The primary anxiety shared by Indian innovators is the reliance on foreign suppliers for the hardware required to train and develop AI models. While the India AI Mission has made strides in allocating funds for graphic processing units (GPUs), founders argue that demand far exceeds the current supply of about 34,000 subsidised units.

Industry experts and investors say the lack of predictable public compute access is the single greatest constraint on scaling deeptech ventures in India.

“If the Budget does only one thing for AI, it should stop spreading money thin and go all-in on subsidised national AI compute. Talent and capital were never India’s bottleneck. But compute is,” observes Rahul Borude, Co-founder and CEO of StampMyVisa, an AI-powered platform that automates visa application processes.

To address this, some are calling for a National AI Compute and Innovation Stack that functions like an open public infrastructure.

“A multi-year commitment to a National AI Compute and Innovation Stack would be the most confidence-building signal—ideally with phased allocations tied to compute credits, shared clusters, and public-private co-investment vehicles,” adds Amit Chand, Founder of BYT Capital, an early-stage deeptech VC.

Beyond pure compute, the physical infrastructure of data centres requires immediate fiscal attention. 

Deloitte partners S. Anjani Kumar, Debashish Banerjee, and Shrenik Shah suggest that India is strategically positioned to capture 10%-15% of this global AI value, with its increasing focus on AI advancements.

They recommend allowing full GST input tax credits on data centre capital assets and providing a 20-year conditional tax holiday for qualifying developers. This must be coupled with energy security, as energy-intensive AI workloads require reliable and sustainable power.

“Sustained growth of digital infrastructure depends on policy frameworks that address three core priorities. First, access to reliable, affordable power, especially for energy-intensive AI workloads, requires long-term visibility on availability and pricing,” notes Manoj Paul, Managing Director of Equinix India, a global data centre and colocation provider for enterprise network and cloud computing.

Paul also advocates for accelerated fibre deployment and regulatory clarity to position India as a regional data hub, while recommending that infrastructure expansion be paired with renewable energy investment policies to ensure long-term sustainability.

As this infrastructure expands, cybersecurity resilience becomes the bedrock of trust. The fragmented approach of multiple regulators has led to gaps in the national strategy.

“A centralised command, building on the framework of the National Security Council Secretariat, with a cybersecurity focus, can streamline efforts and deliver measurable outcomes. Attackers contribute well via their tool kits, whereas the defenders keep fighting individual battles,” explains Harjinder Singh Takher, a Partner at Deloitte.

There is also a strong push for a secure government cloud and sectoral response teams to handle industry-specific threats, such as financial fraud in banking versus life-threatening incidents in industrial IoT.

Market dynamics

A significant shift in the pre-budget discourse is the demand for the government to act as a paying customer rather than just a capital provider. Founders argue that budget allocations for AI solutions in healthcare, defence, and logistics sectors would provide the revenue visibility that de-risks early-stage funding.

Maitreya Wagh, Founder and CEO of Bolna—a voice AI platform—proposes a concrete policy shift. “Mandate that 10% of government AI and IT procurement comes from startups, with simplified tendering and predictable payment cycles. Within a year, this would move dozens of Indian AI systems into production, generate India-scale usage data.”

Ankush Sabharwal, Founder and CEO of CoRover, a conversational platform driven by generative AI, echoes this sentiment by focusing on problem identification.

He argues, “The solutions already exist; what is required is for departments to clearly articulate their problem statements. This process may need to be escalated, along with the allocation of specific funds, to enable departments to adopt and embrace AI solutions.”

This demand-side support is often viewed as more effective than R&D tax incentives, which typically benefit established firms more than loss-making startups.

On the fiscal front, Deloitte partners Peeyush Vaish and Madhava Yathigiri highlight the need for competitive trade policies. “It will be important for India’s Budget strategy to focus on improving cost competitiveness and supporting exporters through the rationalisation of input tariffs and enhanced trade facilitation measures,” they explain.

This includes operationalising voluntary post-clearance revisions under the Customs Act and providing legislative clarity on the transfer of intellectual property rights.

Furthermore, the technology sector is seeking relief in Safe Harbour Margins to reflect the margin compression caused by AI disruption. For India to retain its most successful founders, capital market regulations must also evolve.

“Aligning long-term capital gains treatment for unlisted technology investments with listed peers and enabling dual-class share structures would materially improve founder retention and capital formation,” says Khadim Batti, Co-founder and CEO of Whatfix, an AI-powered digital adoption platform. Such measures would ensure the next generation of global champions is built and listed within India.

Sustainable edge

As India moves to build its own large language models (LLMs) and sovereign datasets, regulatory clarity becomes a competitive advantage.

“Regulatory clarity around data ownership, IP protection, and data security is the more critical lever. Clear rules around Indian data and IP are essential for building sovereign AI systems,” says Rahul Agarwalla, Managing Partner of SenseAI Ventures, an early-stage fund investing in AI-first startups.

Instead of a heavy-handed regulatory approach, industry leaders suggested a use-first, guardrails-next strategy that encourages experimentation while maintaining model safety.

“The most practical near-term approach is to position AI compute infrastructure as critical national digital infrastructure—simple, foreseeable, and broadly available to startups, research institutions, global capability centres, and businesses, while tying it directly to mass-scale upskilling initiatives,” notes Pratap Daruka, CFO of Tredence, a global data science solutions provider.

This upskilling is vital to closing the implementation gap. While India leads in AI skill penetration, there is a shortage of professionals capable of deploying large-scale solutions. Deloitte has recommended introducing project-based AI learning modules from primary schools to universities to ensure a continuum of literacy.

Technology readiness has outpaced human readiness, says Batti of Whatfix, adding, “India must move beyond basic digital literacy and focus on enabling deep, domain-specific human-AI collaboration across the workforce. If professionals are equipped to operate AI-led workflows with confidence, accountability, and sound judgment, India can emerge as the world’s most trusted implementation hub.”

Finally, the cybersecurity workforce also requires standardisation. To address a shortfall of 5 lakh professionals, experts suggest creating India-specific security certifications as a policy requirement for job applications. By institutionalising these skills and updating the National Cybersecurity Policy to reflect the complex threat landscape, India can enhance its reputation as a trusted global destination for business.

Experts believe that by weaving together compute infrastructure, secure digital borders, and a workforce trained for the AI era, the government can ensure that India remains an architect of the global technology landscape.