Celero Communications, a budding technology startup based in Irvine, California, announced on Monday that it has raised $140 million to develop a groundbreaking chip designed to enhance the way artificial intelligence (AI) data centers communicate over long distances. Founded by networking veterans, the company aims to tackle one of the most pressing challenges in AI infrastructure: efficiently linking distant data centers.

Bridging the Distance Between Data Centers

The new chip under development by Celero is designed to convert the pulses of light transmitted through fiber-optic cables into the electrical signals that data centers rely on for processing information. This conversion is crucial because while fiber-optic networks can transport enormous amounts of data across continents, data centers themselves operate on traditional electrical ones and zeros.

The need for improved long-distance connectivity has grown as major tech giants, including Google and Meta Platforms, expand their sprawling data center campuses. Many of these campuses are built near plentiful power sources, but connecting them over hundreds of miles presents a unique technological hurdle. Celero’s approach is focused on solving this challenge while reducing energy consumption.

Leadership with Proven Expertise

Celero was founded by Nariman Yousefi and Oscar Agazzi, both of whom previously held senior positions at Marvell Technology, a prominent supplier of data center networking chips. Leveraging their deep experience in the field, the founders are developing chips with embedded algorithms capable of processing light signals more efficiently than existing solutions.

“Every time the customer sends a pulse of light, you want to send more information,” Yousefi explained in an interview. “That's how you get efficiency over the fiber optics.” By packing more data into each pulse and reducing power usage, Celero’s technology could significantly improve the cost-effectiveness and performance of long-range data center connections.

Details of the Funding Round

The $140 million funding consists of a $100 million Series B round led by CapitalG, Alphabet’s independent growth fund. This latest round follows an earlier, previously undisclosed $40 million investment led by Sutter Hill Ventures, with participation from Valor Equity Partners, Atreides Management, Maverick Silicon, and others.

James Luo, general partner at CapitalG and now a member of Celero’s board, highlighted the market opportunity. Existing networking chips designed for data center campuses struggle to operate over distances of hundreds of miles. Meanwhile, telecom-grade chips that can span these distances are costly and consume significant power, making them unsuitable for AI workloads, where efficiency is key.

“There's a clear gap in the middle,” Luo said. “We are seeing the emergence of a new class of technology, and Celero is at the forefront of that development.”

The Broader Context: AI and Fiber Optics

The investment comes at a time when the AI industry is driving unprecedented demand for high-speed, energy-efficient connectivity. Data centers are no longer isolated hubs; they must communicate seamlessly across regions to support AI models that require massive datasets. Celero’s innovation could bridge the gap between short-range, campus-focused chips and long-haul telecommunications chips, enabling faster and more efficient AI operations.

The startup’s strategy centers on maximizing the potential of existing fiber-optic networks without necessitating entirely new infrastructure. By embedding advanced algorithms directly into the chip, Celero seeks to increase the amount of data transmitted per light pulse, improving speed and reducing energy usage simultaneously.

Looking Ahead

As AI adoption continues to accelerate, the need for smarter, more efficient interconnections between data centers becomes increasingly urgent. Celero’s technology promises to fill a critical void in the market by combining speed, affordability, and power efficiency. With $140 million in fresh funding and leadership grounded in proven industry experience, the company is poised to play a pivotal role in shaping the next generation of AI infrastructure.