Budget 2026 must focus on tax breaks and R&D push to support innovation, says Shorthills AI’s Paramdeep Singh

With Finance Minister Nirmala Sitharaman just days away from presenting the Union Budget 2026–27, technology companies and startups are urging the government to pivot from building in-house AI capabilities toward creating an environment that enables private-sector innovation, according to Paramdeep Singh, co-founder of Shorthills AI.

Singh said the budget should prioritise smart tax incentives and performance-based R&D subsidies to help Indian companies absorb the high costs of building artificial intelligence systems, particularly around computing hardware and electricity.

“The budget should focus on tax schemes and R&D subsidies that encourage corporates to take the risk,” Singh said in an interview with Business Today.

Shifting government’s role from builder to enabler

Singh emphasised that the government should step back from directly running AI labs and instead focus on creating the conditions for private companies to build and scale advanced technologies.

“The government’s role should be to maintain law and order and create a seamless, enabling environment, not to run AI labs,” he said.

He added that India’s AI ambitions will depend less on state-led research and more on whether startups and enterprises can access capital, infrastructure and predictable policy support.

Budget 2026 should focus on hardware and power costs

Singh identified hardware and electricity as the two most critical cost barriers for Indian AI startups.

When asked about making AI compute more accessible, he said Budget 2026 should address two critical costs. Hardware and power.

Emphasising higher customs duties on advanced GPUs and TPUs, Singh said, “providing specific subsidies for AI chips would immediately help startups compete.” He added, “The government should look at power subsidies specifically for running high-density data centres to keep operations viable.”

He said startups need financial support to afford expensive AI infrastructure, especially computing power, and that lowering costs while prioritising domestic innovation would help companies scale.

Data, compute and power: The real threat

Singh said access to data, computing resources and reliable electricity will ultimately determine whether India can build globally competitive AI systems.

“If the budget ensures easy access to data, computing power, and electricity, Indian innovators will be able to deliver breakthrough results on their own,” he said.

He outlined three steps to strengthen India’s AI infrastructure. The first, he said, “is to find a mechanism to share anonymised, high-quality data from these public goods with corporate players.”

“Second, we need to address compute costs, specifically through tax breaks on high-end computing devices.”

“Third, and often overlooked, is the availability of consistent, high-quality power required to run the data centres that house this compute.”

A blueprint for “AI Parks”

Singh also proposed developing dedicated AI Parks through public-private partnerships, modelled on the Software Technology Parks of India framework that helped power India’s IT services boom.

“By offering subsidies and infrastructure in these zones, the government can encourage private players to focus on R&D,” Singh said.

“These structural reforms are essential if we want Indian organisations to move beyond just using AI to actually create high-quality generative AI models,” he added.