India’s push to democratize artificial intelligence infrastructure is gaining momentum. The IndiaAI Mission has announced the addition of 3,850 GPUs through its third round of tenders, further strengthening the country’s AI compute resources.
This expansion includes Google’s sixth-generation Trillium TPUs and Nvidia’s H100 GPUs, bringing India’s total GPU count under the mission to 34,333 units well above its initial target of 10,000.
“No new bidders have been empanelled in this round,” said Abhishek Singh, CEO of IndiaAI Mission, confirming the expansion to ET.
What’s Inside the Third Tender?
Four companies cleared the technical round Ishan Infotech, Locuz Enterprise Solutions, Sify Digital Services, and Vensysco Technologies.
- Locuz: 1,300 Nvidia H100 GPUs
- Ishan Infotech: 50 Google Trillium TPUs
- Sify Digital Services: 2,500 GPUs (1,000 Google Trillium TPUs, 800 Nvidia H200 GPUs, 700 Nvidia L4 GPUs)
- Vensysco: Lowered prices, but offered no new GPUs
This is also the first time Google Trillium TPUs cutting-edge sixth-generation Tensor Processing Units available to Google Cloud customers are part of IndiaAI’s GPU cluster.
Rackbank Datacenters and Teleglobal International failed to clear the technical evaluation.
Pricing and Market Dynamics
The financial evaluation of the third tender has been completed, with minor adjustments in L1 (lowest bidder) prices. According to Singh, empanelled bidders from the first and second rounds will now be asked to match these updated prices.
Earlier, ET reported that GPU prices dropped significantly between the first and second rounds ranging from as little as ₹4 per hour to as much as ₹1,234 per hour.
This aggressive price correction means that IndiaAI is now offering GPUs at under $1 per hour, making it among the cheapest AI compute markets in the world.
Scaling Beyond Initial Targets
When the Union Cabinet approved the ₹10,000 crore IndiaAI Mission in March 2024, the target was to procure 10,000 GPUs. With this round, the program has already crossed 34,000 GPUs surpassing expectations by over threefold.
Earlier, in May 2025, IT Minister Ashwini Vaishnaw had announced the addition of 15,916 GPUs, taking the cluster size to 34,333. The new additions further strengthen India’s AI readiness.
Fueling India’s AI Startups and LLMs
The mission isn’t just about hardware it’s about enabling India’s AI startup ecosystem. IndiaAI provides GPUs at competitive rates to support companies building large language models (LLMs) and other AI products.
Notably, out of 506 proposals received by IndiaAI for building foundation models, 43 are dedicated to LLMs. Startups like Sarvam, Gnani, Gan, and Soket AI Labs are already leveraging these resources to develop indigenous models.
The government is also incentivizing this innovation with funding, incubation, and capital access, aiming to establish India as a global hub for AI model development.
The Road Ahead
With partners like AWS-backed Locuz and Vensysco in play, and with Google Trillium TPUs now entering the cluster, IndiaAI is strategically diversifying its compute base.
The addition of 3,850 GPUs may look incremental, but it symbolizes India’s commitment to scaling AI compute infrastructure at unmatched speed and cost efficiency.
By offering world-class GPUs at globally lowest prices, the mission is setting the stage for India’s AI startups to not only build domestic large language models but also to compete with global giants.
As IndiaAI scales further, the message is clear: India isn’t just catching up in AI infrastructure it’s aiming to lead the global compute economy.
Also Read: How Indian Space Tech Startups Are Powering the Global New Space Era
