IBM Says Optics Breakthrough Can Enhance AI Training

IBM said new optics technology research could improve how data centers train AI models.

The researchers, the tech giant said Monday (Dec.9), have come up with a new process for co-packaged optics (CPO), the next generation of optics technology, to allow for connectivity within data centers “at the speed of light” through optics to complement existing short reach electrical wires.

“As generative AI demands more energy and processing power, the data center must evolve — and co-packaged optics can make these data centers future-proof,” Dario Gil, IBM senior vice president and director of research, said in a news release.

“With this breakthrough, tomorrow’s chips will communicate much like how fiber optics cables carry data in and out of data centers, ushering in a new era of faster, more sustainable communications that can handle the AI workloads of the future,” Gil added.

The release noted that while data centers use light-powered fiber optics for their external communications networks, racks in data centers still primarily communicate via copper-based electrical wires, which may have spent more than half of their time idle, incurring substantial expense and energy.

IBM researchers have found a way to bring optics’ speed and capacity inside data centers, with the potential to “significantly increase the bandwidth of data center communications, minimizing GPU downtime while drastically accelerating AI processing.”

This could mean lower costs for scaling generative artificial intelligence (AI), faster AI model training for large language models (LLMs), and a dramatic reduction of energy use at data centers, “saving the energy equivalent of 5,000 U.S. homes’ annual power consumption per AI model trained,” the company said.

AI-powered data centers could have a transformative aspect in a number of areas, Tim Peters, chief marketing officer at software company Enghouse Systems, told PYMNTS in an interview published earlier this year.

“AI-enabled data centers will exponentially accelerate transaction speeds by optimizing data flow and reducing latency,” Peters said. “AI has already shown the ability to increase supply chain forecasting accuracy by up to 20-50%. These efficiencies will boost eCommerce platforms’ ability to handle surges in demand, allowing retailers to deliver superior customer experiences. With the rise of 5G, this infrastructure will enhance real-time product recommendations, improve voice-activated shopping and personalized customer journeys at unprecedented scales.”