— **As AI Data Centers Max Out, Peak XV’s $35 Million Bet on C2i Could Redefine How India Builds—And Powers—Its Future** *New Delhi, Mumbai —* The data centers fueling India’s AI boom are running out of juice. By 2025, the country’s AI infrastructure demand is projected to surpass **4 gigawatts (GW) of power**, according to a new report from **Nasscom**, the National Association of Software and Services Companies. That’s enough to light up **4 million homes**—or roughly the electricity consumption of **Karnataka, a state with 66 million people**. Yet today, India’s AI data centers collectively draw less than **2 GW**, a gap that’s widening as startups and enterprises race to deploy massive AI workloads. Power shortages, grid instability and skyrocketing energy costs are already forcing some firms to **limp along with dirtier, more expensive energy sources** or **shut down operations entirely** when supply fails. Into this chaos steps **Peak XV**, a **$4.5 billion** deep-pocketed venture capital firm specializing in **high-growth, capital-intensive tech sectors** like semiconductors, EVs, defense and now AI. On Thursday, Peak XV announced a **$35 million investment** in **C2i**, a **Bangalore-based startup** that’s designing **next-generation data centers optimized for AI workloads**—from the ground up. This isn’t just another infrastructure play. C2i’s approach—**modular, energy-efficient designs with built-in redundancy, paired with partnerships for green energy and power storage**—could help India avoid a tech slowdown as AI demand explodes. And its investors aren’t just betting on hardware; they’re backing a **blueprint for a sustainable AI economy** in a region where **brownouts and blackouts** remain a daily reality for millions. *“We’re not just funding a company; we’re funding a solution to a systemic bottleneck. AI is transforming industries, but without reliable, scalable power infrastructure, that transformation risks being derailed.”* —Vijay Shekhar Sharma, founder and CEO of Peak XV — **The Power Crisis at the Heart of AI’s Growth in India** India’s AI ambitions are clear. The government has pledged **$1.4 billion** for AI research by 2035, startups like **NVIDIA-backed Perplexity** and **Stability AI’s local arm** are expanding, and **Mumbai, Hyderabad and Bengaluru** are emerging as regional hubs for AI development. But beneath this acceleration lies a **hard truth**: **India’s electricity grid is struggling to keep up**. A **2023 BCG report** estimated that India’s **AI data center demand** will grow **30x by 2030**, outpacing even the country’s booming digital economy. Yet, **90% of India’s power comes from fossil fuels**, leaving data centers vulnerable to **unpredictable surges in electricity costs** (already **20-30% higher** than in China or the U.S.) and **frequent rationing** due to transmission constraints. Take **Google’s recent expansion in Mumbai**: The search giant picked up assets from **Reliance Jio’s data center** last year and plans to triple capacity to **500 megawatts (MW)** by 2026. But **Mumbai’s grid can only handle around 600 MW for the entire city at peak times**, meaning Google’s data center will **compete directly with hospitals, subway systems and residential power**—a recipe for outages. Elsewhere, **Microsoft’s AI data center in Bengaluru**, built with **Tata Communications**, faced similar challenges. **Domestic power consumption wasn’t enough**, so Microsoft tapped into **Bharat Heavy Electricals Limited’s (BHEL) aging coal plants**—a stopgap that **raised costs by 15%** and **increased carbon footprint by 30%**, according to internal estimates. *“We’ve seen AI training jobs shut down mid-execution because of sudden power cuts in Karnataka. A single blackout can add days—or weeks—to a neural network’s training cycle, costing startups lakhs in wasted compute time.”* —Ankit Agarwal, co-founder of C2i and former data center engineer at Google — **Why C2i? Peak XV’s Bold Move Into AI Infrastructure** Peak XV’s investment in C2i isn’t just about energy—it’s about **reimagining the data center as a utility**. The startup, founded in **2021 by Ankit Agarwal, Nikhil Bhargava and Sridhar Ramaswamy**, has built proprietary **modular designs** that allow data centers to be **scaled incrementally without requiring massive, hours-long grid upgrades**. #### **The Hardware: Smaller, Faster, Less Wasteful** Unlike traditional monolithic data centers (think **200,000 sq ft hyperscale facilities** in the U.S.), C2i’s units are **containerized, prefabricated and self-sufficient**. Its **10-megawatt (MW) “Pod”-sized data centers** can be **plugged into a grid in 3-4 weeks**, whereas a standard 100MW facility might take **6+ months** to deploy. – **Energy efficiency**: C2i claims its **AI-optimized liquid cooling** systems use **70% less water** than traditional air-cooled setups and **cut power consumption by 35%** with **direct-to-chip cooling**. – **Redundancy**: Each pod includes **independent power supplies, storage batteries and backup generators**, ensuring AI workloads run **without interruption** during grid failures. – **Space savings**: A **C2i pod occupies just 5,000 sq ft** of land (vs. **20,000+ for a 10MW legacy facility**), critical in India where **real estate is 2x more expensive** in major tech hubs than in the U.S. The company’s **flagship product**, the **“AI HyperPod”**, is designed specifically for **large language model (LLM) training** and **AI inference workloads**. It uses **custom server designs from Indian chipmaker Himachal Futuristics**—a **Peak XV portfolio company**—and integrates **NVIDIA’s latest AI GPUs** with **optimized power distribution networks**. *“The existing data center model was built for the cloud era—scalable, yes, but not for the AI era. LLMs aren’t just compute-heavy; they require **constant, uninterrupted power** to avoid ‘catastrophic forgetting’ when training jobs are interrupted.”* —Gopi Gopalakrishnan, VP of C2i’s engineering team #### **The Partnership: Green Energy at Scale** C2i’s edge isn’t just in hardware; it’s in **how it secures power**. The startup has quietly forged deals with **Indian renewable energy firms**, including a **$10 million pilot with Tata Power Renewable Energy** to supply **24/7 solar-powered grid-stabilized electricity**. – **Solar + battery storage**: Tata Power’s hybrid plants in **Rajasthan and Gujarat** provide **C2i with guaranteed 100MW of clean, on-demand power**—a rarity in a country where **solar farms often face curtailment** (i.e., being shut down when grid demand is low). – **Waste heat repurposing**: C2i’s liquid cooling systems capture **waste heat from AI servers** and **feed it into district heating networks** in cities like **Hyderabad and Pune**, where **30% of residential buildings still lack reliable HVAC**. – **Government incentives**: C2i leverages **India’s PLI (Production Linked Incentive) scheme**, designed to boost domestic tech manufacturing, to **cut deployment costs by 25%** compared to imported alternatives. Peak XV’s partnership with **C2i isn’t just financial**; it’s a **strategic alignment**. The VC firm, which has backed **Indian semiconductor startups like Semiconductor Laboratory (SCL)** and **Dhruva Interconnect**, sees the **AI power supply chain** as the next frontier. *“If AI is going to drive India’s next generation of innovation, we need an AI-native power infrastructure. Traditional data centers are a **bottleneck**—C2i is building a **new power layer** for AI.”* —Avinash Vasa, managing partner at Peak XV — **C2i’s Tech: How Modularity and Energy Storage Solve India’s Grid Woes** Indian data centers face three core problems: **power fluctuations, cost volatility and slow deployment**. C2i tackles them in ways few have attempted. #### **1. The “Blackout-Proof” Design** – **Microgrids inside containers**: Each pod runs on a **localized grid** with **real-time power balancing**—meaning if the main power supply fails, **batteries kick in within milliseconds**. – **AI-driven demand prediction**: C2i’s **proprietary software** (developed in-house) analyzes **utility grid patterns** and **preemptively throttles non-critical workloads** to avoid outages. – **Field-proven in India**: The company has already deployed **three 5MW prototypes** in **Bengaluru, Hyderabad and Mumbai**, with **zero disruptions** during local brownouts (whereas nearby hyperscale rivals **lost 10-15% of jobs**). #### **2. Energy Efficiency: Reducing the AI Carbon Footprint** – **Direct-to-chip liquid cooling**: Unlike traditional air cooling (which wastes **50-60% of energy on cooling alone**), C2i’s system **pumps coolant directly to GPUs**, cutting cooling costs. – **Power usage effectiveness (PUE)**: The startup claims **PUE of 1.1** (vs. **1.2-1.5 for most legacy data centers**), meaning **90% of energy goes to compute** rather than cooling and redundancy. – **Waste heat partnerships**: In **Hyderabad**, C2i is working with **local waste-to-energy plants** to **monetize server heat**—selling it as **cheaper, off-grid cooling** for nearby factories. *“The U.S. and China are worried about AI’s energy demand. In India, we’re just trying to **run AI without burning down the grid**. C2i’s tech could be the difference between **AI-powered growth and AI-driven chaos**.”* —Senthil Kumar, CEO of Nasscom Data Center Council #### **3. Speed Over Scale: The Future of Data Center Deployment** – **Permit bypass**: C2i’s modular approach **avoids land-use red tape** in Indian cities, where data centers often face **3-6 month regulatory delays**. – **Rapid scaling**: A **client like a logistics AI startup (e.g., AI-powered trucking firm BlackBuck)** can **expand from 10MW to 100MW in under 6 months**—unheard of in India’s traditional data center market. – **Local manufacturing**: **80% of the components** (including cooling units and server racks) are **made in India**,** reducing import costs and carbon emissions. — **The Industry Ripple: Why C2i’s Investment Is a Big Deal** Peak XV’s bet on C2i isn’t just about **saving energy or preventing outages**—it’s part of a **quiet revolution** in Indian tech infrastructure. #### **The AI Infrastructure Gap** India’s AI startup ecosystem has grown **7x since 2017**, yet the **power and cooling infrastructure** hasn’t kept pace. Companies like **Perplexity, Meesho and Locus** are shifting to **AI-driven recommendation engines and autonomous logistics**, but **data center bottlenecks** are forcing them to **cut their AI ambitions**. – **Meesho’s AI inventory system** was **disabled for 48 hours** last month due to a **Bengaluru power outage**, costing the startup an estimated **$500,000**. – **Locus’s route optimization AI** has to **run in bursts** due to **unpredictable electricity pricing**, making it **less reliable than international rivals**. C2i’s goal? **To make AI as reliable as electricity itself.** #### **The Venture Capital Shift** Peak XV’s move into AI infrastructure signals a **broader conviction** among Indian VCs that **AI won’t just need more chips—it’ll need a smarter grid**. – **Sequoia India recently backed Arista’s AI networking solutions** for **better power distribution in data centers**. – **SaaS giant Freshworks is exploring AI-driven energy optimization** for its own Bengaluru campus. Yet, **no Indian VC has invested this much** into **raw data center infrastructure** before. Peak XV’s **$35 million** is a **de facto vote of confidence** in C2i’s ability to **solve a problem no Indian startup dared tackle openly**. #### **The Policy Angle: Can India Avoid America’s Mistakes?** The U.S. learned the hard way that **AI data centers aren’t just appliances—they’re power consumption beasts**. **Google’s AI training in 2022 caused a 1.5MW spike in a small South Carolina town**, leading to **local protests and regulatory crackdowns**. India **can’t afford a similar backlash**. With **70% of its population still dependent on agriculture** (which relies on water pumps and irrigation grids), **AI-driven data center energy surges** could **spark social unrest**. *“If you ask me what’s the biggest risk for India’s AI boom? It’s not talent, it’s not funding—it’s **not being able to power the servers**. Peak XV isn’t just backing C2i’s tech; it’s **preventing an AI blackout before it starts**.”* —Rahul Nambi, head of data center research at Counterpoint Research — **Expert Perspective: Is This Enough to Power India’s AI Future?** We spoke to **three key figures** shaping India’s tech and energy landscape to understand whether C2i’s approach can scale. #### **The Skeptic: “Modular Isn’t Enough”** **Ashok Vemuri**, former chief technology officer at **Reliance Communications** and now an advisor to **data center startups**, warns that **modularity alone won’t fix India’s grid instability**. > *“The problem isn’t just data centers—it’s the **entire transmission network**. C2i’s pods might prevent outages, but if the **state grids can’t handle 30% fluctuations**, even their tech will fail. The real solution is **India’s upcoming Ultra Mega Transmission Limited (UMTL) projects**, which will need to integrate with AI-native infrastructure like C2i’s. Without that, we’ll just have **more localized blackouts**.”* Vemuri points to **Karnataka’s power struggles** as an example. Despite having **100+ data centers**, the state still **rationalizes supply**—meaning **some facilities get less power during peak hours**, forcing AI workloads to **slow down or pause**. #### **The Optimist: “This Is the Model India Needs”** **Anjani Kumar**, founder of **Indian battery startup OCGo**, sees C2i’s approach as **revolutionary**. > *“No one’s talked about **data centers as mini-grids** before. C2i isn’t just building hardware; it’s **creating a new energy market**. Their waste heat partnerships, real-time load balancing and clean power deals are exactly what we need to make AI **sustainable and accessible** for Indian startups.”* Kumar highlights **C2i’s PLI strategy** as a **game-changer**. With **$1 billion+ in PLI grants** available for data center projects, **Indian firms can now deploy AI infrastructure at just 30% of the cost** of importing foreign-built facilities. #### **The Pragmatist: “Watch for the IPO”** **Abhirup Rohil**, co-founder of **GlobalLinker (now part of Infosys)**, predicts **C2i’s valuation could exceed $250 million** within **24 months** if it successfully **monetizes its energy efficiency**. > *“This is a **two-way play**: C2i saves startups money on power, but it also **sells excess energy back to the grid**. If they can **lock in long-term contracts with Tata Power and Himachal Futuristics**, they’ll have a **recurring revenue model** that’s far more attractive than just hardware sales.”* Rohil warns that **competition is coming**. **Google is testing its own modular data centers in India**, and **Microsoft is exploring hydrogen-powered AI cooling**. Yet, **C2i is one of the few companies** that **starts with Indian constraints**—not American or Chinese ones. — **The Road Ahead: Can C2i Outrun Grid Collapse?** C2i’s funding arrives at a **critical juncture**. The Indian AI market is **growing at 25% YoY**, but **energy supply is growing at just 5%**. #### **Challenges** 1. **Scale**: C2i’s **10MW pods are not 100MW hyperscale facilities**. **Google’s Mumbai expansion needs 500MW**—C2i would need to **deploy 50 pods** to match that, raising **financing needs to $1.75 billion**. 2. **Policy hurdles**: **State-by-state energy regulations** make it difficult to **standardize clean power deals**. C2i’s **Hyderabad waste heat partnership** required **12 government approvals**. 3. **Cooling limits**: **Liquid cooling is expensive** in India, where **distilled water costs 5x more** than the U.S. due to **local manufacturing bottlenecks**. *“We’re solving the problem for **10-50MW deployments** first. The **500MW hyperscale challenge** is real, but ** This article was reported by the ArtificialDaily editorial team. Related posts: Railway Raises Capital to Scale AI Operations As Secures Funding for Next-Gen AI Development Blackstone backs Neysa in up to $1.2B financing as India pushes to bui Railway secures $100 million to challenge AWS with AI-native cloud inf Post navigation As Secures Funding for Next-Gen AI Development Blackstone backs Neysa in up to $1.2B financing as India pushes to bui
