Rising GPU power demands and efficiency losses are driving C2i Semiconductors’ $15 million push to deliver grid‑to‑core solutions for AI data centres worldwide.
C2i Semiconductors has secured $15 million in funding led by Peak XV Partners, with Yali Deeptech and TDK Ventures also participating. The semiconductor startup, founded in 2024 by former Texas Instruments executives, is developing power management solutions for AI data centres and cloud infrastructure.
The company is focusing on system-level innovations to improve how electricity flows from the grid to processor cores. Its ‘grid‑to‑core’ platforms aim to address efficiency and reliability issues across enterprise server architectures.
C2i previously raised $4 million in November 2024 in a Series A round led by Yali Capital and Intel CEO Lip‑Bu Tan. The startup plans to move quickly from design to silicon, with its first chip scheduled for tape-out in April and a second in July. Manufacturing will take place at Tower Semiconductor in Israel and GlobalFoundries in Singapore or Dallas.
The founders highlighted rising power demands in AI systems. GPUs that once consumed hundreds of watts now require several kilowatts, while hyperscale data centres operate at megawatt levels. Current conversion processes from 800 volts to 800 millivolts incur 15–18% energy loss, and instability can reduce GPU performance by up to 10%.
C2i claims its architecture can recover 8–10% efficiency, improve GPU performance by around 3%, and extend server lifetimes. The company is already in discussions with several enterprise customers to define components for next-generation platforms.
The startup estimates that power-related semiconductors represent a $50-75 billion market opportunity, with its targeted segment valued at $5-8 billion. It plans to establish a US office to strengthen customer engagement and later expand into Taiwan to support original design manufacturers (ODMs).
In a report by The Economic Times, Peak XV Partners managing director Rajan Anandan said power has become a bottleneck in scaling AI, adding that C2i’s approach could help extend GPU longevity and reduce costs across the industry.
