The race to scale artificial intelligence infrastructure is hitting an unexpected barrier — power. As energy demand surges across global data centers, AI data center power solutions are becoming critical for sustaining growth.

Recognizing this shift, Peak XV Partners has invested in Indian startup C2i Semiconductors, which is building plug-and-play power systems designed to reduce energy losses and improve the efficiency of large-scale AI infrastructure.

The Bengaluru-based company recently raised $15 million in a Series A funding round led by Peak XV Partners, with participation from Yali Deeptech and TDK Ventures. This brings the startup’s total funding to $19 million.

Power Emerges as the Biggest AI Infrastructure Challenge

As AI adoption accelerates, the demand for electricity to run data centers is increasing rapidly. According to industry estimates, global data center electricity consumption could nearly triple by 2035, while power demand may rise 175% by 2030 compared with 2023 levels.

The issue isn’t just generating electricity — it’s converting and delivering power efficiently. Inside data centers, high-voltage energy must be reduced thousands of times before reaching GPUs, a process that currently wastes around 15% to 20% of energy.

C2i’s technology focuses on solving this inefficiency through integrated AI data center power solutions that manage power conversion, control, and packaging as a single system.

C2i’s Grid-to-GPU Power Architecture

Founded in 2024 by former Texas Instruments power experts Ram Anant, Vikram Gakhar, Preetam Tadeparthy, Dattatreya Suryanarayana, Harsha S. B, and Muthusubramanian N. V, C2i is developing a “grid-to-GPU” power delivery platform.

The system redesigns power flow from the data center’s electrical infrastructure directly to processors, improving overall energy efficiency.

C2i estimates its platform could reduce end-to-end energy losses by roughly 10%, saving about 100 kilowatts per megawatt consumed. These gains could lower cooling requirements, improve GPU performance, and significantly reduce operational costs.

For data center operators, such improvements directly impact profitability and long-term infrastructure economics.

Why Investors Are Betting on AI Data Center Power Solutions

For Peak XV Partners, the investment reflects the growing importance of energy efficiency in AI infrastructure. Once data centers are built, electricity becomes the dominant ongoing expense.

Even small efficiency gains can translate into massive cost savings. According to the firm, reducing energy expenses by 10% to 30% could generate billions of dollars in savings across large-scale deployments.

C2i plans to test its first silicon designs between April and June, followed by performance validation with hyperscalers and data center operators. The startup has already assembled a team of 65 engineers and is preparing customer operations in the United States and Taiwan.

A High-Risk, High-Impact Technology Bet

Power delivery remains one of the most complex and entrenched components of data center infrastructure, typically dominated by large industry players.

Unlike companies that optimize individual components, C2i is redesigning the entire power system — a capital-intensive strategy that requires coordination across chip design, packaging, and system architecture.

The company’s success will depend on execution, market adoption, and early validation results expected in the coming months.

India’s Growing Semiconductor Opportunity

The investment also highlights the rapid development of India’s semiconductor ecosystem.

India’s expanding engineering talent pool, combined with government-backed incentives for chip design, has made it increasingly viable for startups to develop globally competitive semiconductor technologies domestically.

Industry observers believe this momentum could position India as a key player in next-generation AI infrastructure development.

As C2i begins validating its AI data center power solutions with customers, the coming months will determine whether its technology can help solve one of the most pressing challenges facing the AI industry today.