CtrlS Bets ₹10,000 Crore on an AI Hyperscale Factory in Bihar

Bodh Gaya is the place where the Buddha is said to have reached enlightenment under a bodhi tree some 2,500 years ago. Soon, it might also be home to one of eastern India's largest AI data centres — a ₹10,000 crore hyperscale facility that didn't exist on anyone's map a year ago.

On October 13, data centre operator CtrlS Datacenters switched on a new ₹400 crore facility in Patna's Patliputra Industrial Area, with Bihar Chief Minister Samrat Choudhary cutting the ribbon. In the same breath, the company signed a Letter of Intent (LoI) — essentially a formal, if non-binding, statement of intent to invest — with the state government for a much bigger project: a 150 megawatt (MW) "hyperscale" AI factory in Bodh Gaya — "hyperscale" just means a data centre built at a scale far beyond a typical corporate server room, with thousands of GPUs packed in to train and run AI models.

What's actually being built

The Patna facility is live now, with 12 MW of what the industry calls "IT load" — basically a measure of how much computing hardware a data centre can power and cool at once. The Bodh Gaya project is a different scale entirely: at 150 MW, it would be more than ten times larger, purpose-built for AI workloads rather than the general-purpose cloud hosting most Indian data centres were built for over the last decade.

CtrlS isn't a newcomer chasing a trend. The company already runs 19 data centres across nine Indian markets and has set a target of building more than 4.4 gigawatts of capacity nationally by 2030 — a number that gives a sense of just how aggressively Indian operators expect AI demand to grow.

Why Bihar, of all places

For decades, Bihar has been a state people left to find tech jobs, not one that built data centres. A few things changed that calculus:

• Land and power are cheaper and easier to acquire outside India's saturated data centre belts of Mumbai, Chennai, Hyderabad and the NCR.
• State governments, Bihar included, have been offering incentives to pull industrial investment into regions that have historically missed out on it.
• AI workloads don't need to sit next to financial exchanges or corporate headquarters the way older data centres did — a GPU cluster just needs land, power and connectivity.

That last point is the real shift. Cloud data centres clustered near business hubs because latency to enterprise customers mattered. AI training runs care less about shaving milliseconds off a trade and more about cheap, reliable megawatts — which is exactly what second-tier states can offer that Mumbai increasingly can't.

A hyperscale AI campus rising next to a 2,500-year-old pilgrimage site says a lot about how fast India's compute map is being redrawn.

Part of a bigger pattern

CtrlS's Bihar bet lands in the middle of a broader scramble to build AI infrastructure on Indian soil. Code24 covered TCS's $1 billion deal with TPG for a $2 billion AI data centre project last year, and similar announcements have followed from other operators chasing the same demand curve: Indian companies and the government don't want to keep renting compute from American hyperscalers indefinitely.

That matters for Indian developers and startups in a very concrete way. More domestic AI compute capacity could eventually mean lower latency and cheaper access to GPUs for Indian AI teams who currently queue for capacity on AWS, Azure or Google Cloud regions, some of which aren't even based in India. It also lines up with the Digital Personal Data Protection Act's push toward keeping sensitive data processed within the country — a data centre in Bihar is still a data centre in India, which matters for companies that need to demonstrate local data residency to regulators or enterprise clients.

The part that's still on paper

It's worth being clear about what has actually happened and what hasn't. The Patna facility is real and running. The Bodh Gaya "AI factory" is, for now, a signed intention with a price tag attached — no construction timeline, commissioning date, or technical blueprint has been made public. LoIs in Indian infrastructure have a mixed track record of turning into finished buildings on schedule, and ₹10,000 crore is a large enough number that execution risk is the obvious thing to watch.

If it does get built, Bihar goes from a state that exported engineers to one that houses the infrastructure those engineers' work increasingly depends on. That's a bigger deal for the state's economy than the headline investment figure alone suggests — construction jobs, power infrastructure upgrades, and a precedent for the next operator deciding where to place the next 150 MW. Whether Bodh Gaya becomes a genuine node on India's AI map, or just another LoI that quietly expires, is the thing worth watching over the next year, not the announcement itself.