Indian conglomerate Adani plans very slow $100 billion AI datacenter build

Adani plans massive AI infrastructure build

Adani Group will spend $100 billion to build 5 gigawatts of AI-ready datacenters across India by 2035. The industrial conglomerate intends to power these facilities entirely with renewable energy to establish what it calls sovereign infrastructure. This investment aims to position India as a primary creator of artificial intelligence rather than a mere consumer of foreign technology.

The company confirmed the plans during a period of intense competition for compute capacity in the region. Adani currently lacks secured land for all these facilities, but the roadmap targets high-density compute clusters. These sites will feature advanced liquid cooling systems and power architectures designed for next-generation AI workloads.

Adani’s timeline remains conservative compared to the aggressive spending cycles of American hyperscalers. While Adani plans to spend $100 billion over the next nine years, Amazon, Google, and Meta will spend a combined $635 billion on AI infrastructure in 2026 alone. The scale of US investment dwarfs the local effort, even as Adani leverages lower labor costs in the Indian market.

Big Tech targets the Indian market

Global technology giants are already pouring capital into India to capture the growing demand for cloud services. Amazon Web Services (AWS), Google, and Microsoft have collectively committed $67 billion to Indian AI infrastructure over the next few years. These companies are moving faster than Adani to establish a physical presence in the country’s growing tech hubs.

Adani faces the challenge of building a massive footprint while these established players already operate existing regions. The conglomerate expects its datacenters to host compute capacity specifically for Indian-language Large Language Models (LLMs). This focus on localized data initiatives serves as a defensive strategy against the dominance of Western AI models.

The 5GW target represents a massive energy requirement that few companies can fulfill independently. Adani’s existing portfolio in renewable energy provides a vertical advantage for powering these "bit barns." The company plans to integrate its solar and wind assets directly with the datacenter sites to reduce operational costs and meet sustainability goals.

• Total Investment: $100 billion through 2035
• Power Capacity: 5 gigawatts of renewable energy
• Target Workloads: High-density AI and Indian LLMs
• Cooling Tech: Advanced liquid cooling systems

India wants sovereign AI control

Prime Minister Narendra Modi recently declared that AI represents a civilizational inflection point for the nation. Speaking at the AI Impact Summit, Modi emphasized that India must become one of the top three AI superpowers globally. He framed the technology as a tool to expand human capability rather than a threat to the existing social foundation.

The government wants to avoid a future where Big Tech companies hold a monopoly over India’s digital economy. Modi’s vision includes Indian startups achieving high valuations through "made-in-India" AI models. This sovereign approach focuses on using AI as an enabler of opportunity for citizens rather than an instrument of external control.

To support this vision, the government is actively intervening in the hardware market. The National AI Mission currently operates a cluster of 38,000 GPUs that it rents to local companies at subsidized rates. These startups pay approximately ₹65 ($0.72) per hour to access the high-end compute needed to train and run models.

Government expands GPU rental programs

Minister of Electronics and Information Technology Ashwini Vaishnaw announced plans to add another 20,000 GPUs to the government’s rental system. This expansion will bring the total state-supported pool to 58,000 units. The initiative aims to lower the barrier to entry for small developers who cannot afford to build their own facilities.

Nvidia is currently the primary hardware provider for these national initiatives. The chipmaker is partnering with various Indian cloud operators to build "AI factories" across the country. These partnerships allow local firms to deploy H100 and Blackwell chips to compete with global cloud providers.

Adani’s entry into the market adds a significant local player to this hardware race. However, the conglomerate must compete with both the government's low-cost rental programs and the massive scale of US hyperscalers. The success of Adani's $100 billion bet depends on its ability to deliver 5GW of power faster than its rivals can lease it.

Technical hurdles for Indian datacenters

Building 5GW of capacity requires more than just capital and hardware. India’s climate presents significant challenges for traditional air-cooled datacenters, which often struggle with high ambient temperatures. Adani’s commitment to liquid cooling suggests a shift toward high-efficiency designs that can handle the heat generated by dense GPU racks.

These facilities must also navigate India's complex power grid and land acquisition laws. Adani’s status as a major energy producer gives it a unique edge in securing the high-efficiency power architecture needed for AI. The company intends to use its own power transmission networks to ensure the reliability of the 5GW footprint.

The focus on Indian-language LLMs also requires massive localized datasets. Sovereign infrastructure ensures that this data remains within national borders, addressing privacy and security concerns raised by the government. Adani’s datacenters will likely become the primary hubs for these "national data initiatives."

• Current GPU Pool: 38,000 units
• Planned GPU Addition: 20,000 units
• Rental Cost: $0.72 per hour
• US Hyperscaler 2026 Spend: $635 billion

The race for AI supremacy

Adani is positioning itself as the infrastructure backbone for a new era of Indian technology. By combining energy production with compute capacity, the group hopes to capture the entire AI value chain. This strategy mimics the vertical integration seen in US tech giants but applies it at a national scale within India.

The competition is already fierce as local cloud operators sign deals with Nvidia to secure the latest Blackwell chips. Adani will face pressure to accelerate its 2035 timeline if it wants to remain relevant in a market moving at the speed of software. The next decade will determine if India can successfully build its own AI stack or if it will remain dependent on American platforms.

For now, the government continues to pitch AI as a problem-solver for the masses. By subsidizing GPU access and encouraging massive private investment, India is attempting to leapfrog traditional development cycles. Adani’s $100 billion commitment is the largest private signal yet that the nation is serious about its sovereign AI goals.