Peak XV Backs C2i to Solve AI Data Center’s Power Problem

When the founders of C2i Semiconductors left their jobs at Texas Instruments in 2024, they weren’t just starting another chip company. They were betting that power—not compute—would become the defining bottleneck of the AI era. Eighteen months later, that bet is looking prescient. Peak XV Partners has led a $15 million Series A round into the Bengaluru-based startup, joining a growing chorus of investors who believe the next trillion-dollar opportunity lies not in building bigger GPUs, but in powering them more efficiently.

“If you can reduce energy costs by, call it, 10 to 30%, that’s like a huge number. You’re talking about tens of billions of dollars.” — Rajan Anandan, Managing Director, Peak XV Partners

The Grid-to-GPU Challenge

C2i—which stands for control, conversion, and intelligence—is tackling a problem that most data center operators would rather not think about. Every watt of electricity that enters a facility must undergo multiple transformations before it reaches a GPU. High-voltage power steps down thousands of times, passing through converters, regulators, and distribution systems. Each handoff bleeds energy.

According to C2i co-founder and CTO Preetam Tadeparthy, this process currently wastes between 15% and 20% of the electricity that data centers consume. As voltage requirements climb—from 400 volts to 800 and beyond—the losses compound. For a hyperscale facility drawing hundreds of megawatts, that inefficiency translates into hundreds of millions of dollars in wasted energy over a facility’s lifetime.

C2i’s approach is to redesign power delivery as a single, integrated “grid-to-GPU” system. Rather than optimizing individual components, the company is building a platform that spans the entire power chain—from the data center bus to the processor itself. By treating conversion, control, and packaging as one system, C2i estimates it can cut end-to-end losses by roughly 10%.

Why Peak XV Bet on Power

For Peak XV, which split from Sequoia Capital in 2023, the investment represents a thesis about where value will accrue in the AI infrastructure stack. After the upfront capital costs of servers and facilities, energy becomes the dominant ongoing expense for data center operators. In a market where every percentage point of efficiency matters, a 10% improvement in power utilization isn’t just an operational win—it’s a competitive advantage.

The timing is critical. Data center electricity consumption is projected to nearly triple by 2035, according to a December 2025 report from BloombergNEF. Goldman Sachs Research estimates that data center power demand could surge 175% by 2030 from 2023 levels—the equivalent of adding another top-10 power-consuming country to the global grid.

The technical challenge is equally significant. Power delivery has long been dominated by entrenched incumbents with deep balance sheets and years-long qualification cycles. Redesigning the stack end-to-end requires coordinating silicon, packaging, and system architecture simultaneously—a capital-intensive approach that few startups attempt.

“All that translates directly to total cost of ownership, revenue, and profitability.” — Preetam Tadeparthy, Co-founder and CTO, C2i Semiconductors

India’s Semiconductor Moment

The C2i investment also reflects a broader shift in how semiconductor startups are built. Rajan Anandan compares the current moment to e-commerce in 2008—an industry just beginning to find its footing in India. The country has developed deep engineering talent, with a growing share of global chip designers based there. Government-backed design-linked incentives have lowered the cost and risk of tape-outs, making it increasingly viable for startups to build globally competitive products from India rather than operate only as captive design centers for multinational corporations.

C2i has already built a team of about 65 engineers and is establishing customer-facing operations in the U.S. and Taiwan. The company’s first two silicon designs are expected to return from fabrication between April and June, after which validation with data center operators and hyperscalers will begin.

The coming months will test whether C2i’s integrated approach can deliver on its promises. For Peak XV, the feedback loop should be relatively short. “We’ll know in the next six months,” Anandan said, pointing to upcoming silicon and early customer validation as the moment when the thesis will be tested.

If C2i succeeds, it won’t just be a win for the startup and its investors. It will be a proof point that the next generation of AI infrastructure can be built more efficiently—and that the companies solving the power problem may ultimately matter as much as the ones building the models that consume it.

This article was reported by the ArtificialDaily editorial team. For more information, visit TechCrunch.