C2i Semiconductors Lands $15M to Tackle AI Data Center Power Bottleneck, Promising Massive Efficiency Gains

The relentless growth of artificial intelligence is creating an unprecedented demand for computing power, and increasingly, the availability of that power is becoming a major limiting factor. Data centers, the engines driving the AI revolution, are facing escalating energy consumption, prompting a surge in innovation aimed at boosting efficiency. C2i Semiconductors, a Bengaluru-based startup, is emerging as a key player in this arena, securing $15 million in Series A funding led by Peak XV Partners to revolutionize power delivery in AI data centers.
C2i's core mission is to address the significant energy losses that occur during power conversion within data centers. Currently, the process of stepping down high-voltage power to the levels required by GPUs can waste as much as 15% to 20% of the total energy input. Founded in 2024 by industry veterans from Texas Instruments, C2i is developing plug-and-play, system-level power solutions that treat power conversion, control, and packaging as a single, integrated platform, spanning from the data center bus all the way to the processor. This 'grid-to-GPU' approach promises to significantly reduce energy wastage and improve overall data center economics. The Series A round also included participation from Yali Deeptech and TDK Ventures, bringing the startup's total funding to $19 million.
The potential impact of C2i's technology is substantial. The company estimates that its solutions can cut end-to-end power losses by approximately 10%, translating to significant savings in kilowattage per megawatt consumed. This reduction in energy consumption directly impacts cooling costs, GPU utilization, and overall data center profitability. Industry projections paint a stark picture of escalating energy demand. A BloombergNEF report from December 2023 forecasts that data center electricity consumption will nearly triple by 2035, while Goldman Sachs Research anticipates a 175% surge in data center power demand by 2030 compared to 2023 levels. This increase is equivalent to adding another top-10 power-consuming country to the global grid, underscoring the urgent need for innovative solutions like C2i's.
Peak XV Partners, which spun out from Sequoia Capital in 2023, recognizes the pivotal role of power costs in shaping the economics of AI infrastructure. Rajan Anandan, managing director at Peak XV, emphasizes that energy costs become the dominant ongoing expense for data centers after the initial capital investment in servers and facilities. Even incremental efficiency gains can translate into substantial savings. C2i expects its first silicon designs to return from fabrication between April and June. The startup plans to validate performance with data-center operators and hyperscalers that have already expressed interest in reviewing the data. The company is also expanding its operations, establishing a customer-facing presence in the U.S. and Taiwan to support early deployments.
C2i's approach challenges the traditional power delivery landscape, which has long been dominated by established players. By taking a holistic view of power delivery, C2i is tackling a complex challenge that requires expertise in silicon, packaging, and system architecture. This capital-intensive approach is relatively rare for startups, and proving the technology in real-world production environments will be a crucial test. However, if successful, C2i's innovative power solutions could significantly reduce the energy footprint of AI data centers, paving the way for a more sustainable and efficient future for artificial intelligence.
Alex Chen
Senior Tech EditorCovering the latest in consumer electronics and software updates. Obsessed with clean code and cleaner desks.
Read Also
AI ID Photo App Generates Passport and ID Photos from Selfies on iPhone
A new iPhone app, AI ID Photo, leverages artificial intelligence to generate passport and government-compliant ID photos directly from user-submitted selfies. Available for free, the app aims to streamline the often tedious process of obtaining compliant identification photos, potentially saving users time and money.
HJX: Aims to Simplify Web Development with AI-Native Unified UI Language
A new compiled UI language called HJX has emerged, promising to unify HTML, CSS, and JavaScript into a single, streamlined file. HJX aims to eliminate boilerplate, configuration hassles, and framework lock-in, potentially revolutionizing how web developers build interactive interfaces. By compiling to clean HTML, CSS, and JavaScript, HJX strives for performance and maintainability.
CodeGraph CLI: Unleash AI-Powered Code Understanding and Generation Directly from Your Terminal
Tired of tedious code analysis? CodeGraph CLI offers a powerful solution: a command-line tool that uses graph-augmented Retrieval Augmented Generation (RAG) to let you chat with your codebase, perform semantic searches, analyze impact, and even generate code using AI. This innovative tool promises to streamline development workflows and unlock new levels of code comprehension.
Wildex Update Turns Wildlife Spotting into a Real-World Pokémon GO with AI and Wilderness Guides
Wildex, the app that gamifies real-world wildlife observation, has released a significant update, transforming nature walks into engaging augmented reality experiences. The update boasts an improved AI model for species identification, enhanced user interface elements, and the introduction of 'Wildboy,' a virtual wilderness guide, promising a richer and safer exploration experience.