Indian startup C2i optimizes power supply for AI data centers

Indian startup C2i has attracted $15 million in investment from the major venture fund Peak XV Partners. This significant sum testifies to the high potential of the technology aimed at solving one of the most acute problems of modern digital infrastructure – the energy deficit in data centers, particularly those focused on computationally intensive artificial intelligence tasks. The startup offers an innovative solution called "grid-to-GPU," which promises to radically reduce electricity losses during the power distribution stage within the data center.

The problem of data center energy consumption is not new, but with the exponential growth in demand for computational resources for training and operating artificial intelligence models, it has reached a new level. Modern graphics processing units (GPU), which are the heart of AI infrastructure, require enormous amounts of energy. Meanwhile, traditional approaches to electrical distribution within data centers lead to significant losses, reaching 15-20% of total consumption. These losses occur at various stages: from voltage transformation to the operation of numerous converters and cable systems. In conditions where existing electrical grids often cannot handle growing loads, and building new infrastructure is an expensive and time-consuming process, every saved kilowatt-hour becomes invaluable.

C2i's technology is aimed precisely at optimizing this critical link. Instead of relying on multi-stage transformations and distribution networks with high losses, the company is developing a system that delivers electrical power to GPUs with minimal losses. Although specific technical details have not yet been fully disclosed, the essence of the approach lies in more efficient management of energy flows, possibly using new materials, architectural solutions, or intelligent power management systems.

The goal is to come as close as possible to the ideal scenario where power from the electrical grid is delivered directly to GPU computing cores, bypassing unnecessary intermediate stages. This not only reduces overall energy consumption but can also contribute to increased equipment stability through cleaner and more predictable power supply.

The consequences of implementing such solutions could be quite significant. First, there is a direct contribution to improving the energy efficiency of AI infrastructure. Reducing losses means decreasing overall electrical consumption, which is critically important in the context of global efforts to combat climate change and the pursuit of sustainable development.

Second, there is a potential solution to the "bottleneck" problem in the form of physical limitations of electrical grids. If data centers can operate more efficiently, they will require less peak power, which could reduce the load on existing grids and postpone or even reduce the need for large-scale investments in their modernization. Third, more efficient power supply could lead to increased equipment lifespan and reduced operating costs.

For companies developing AI, this means the ability to scale their operations faster and with lower costs.

Thus, the attracted investments and C2i's innovative approach represent an important step in the development of AI infrastructure. As the world continues to strive for increasingly powerful and sophisticated AI models, ensuring their stable and energy-efficient power supply becomes no less important than the development of algorithms themselves. C2i's technology, aimed at optimizing the "grid-to-GPU" chain, has every chance of becoming a key factor that will allow us to overcome current energy barriers and open new horizons for scaling and deploying artificial intelligence.