AI is transforming data centers, making them larger and more complex than ever before. Companies like Ciena are seeing the impact firsthand. “Some of these large data centers are just mind-blowingly large, they are enormous,” says Gary Smith, CEO of Ciena.
“You have data centers that are over two kilometers.”
As data centers expand, the strain on direct-connect technology is growing. “You’re beginning to see these campuses get to quite decent distances, and that is putting massive strain on the direct-connect technology,” Smith notes. Ciena expects to start selling fiber-optic equipment to connect GPUs within data centers.
This equipment is similar to what is used in long-haul telecom networks but adapted for internal use.
AI reshapes data center infrastructure
The sheer size of these AI data centers is staggering.
Thomas Graham, co-founder of chip startup Lightmatter, says there are at least a dozen new AI data centers either planned or under construction. Each one will require a gigawatt of power to run. “Just for context, New York City pulls five gigawatts of power on an average day, so, multiple NYCs,” Graham explains.
By 2026, global AI processing is expected to require 40 gigawatts of power “specifically for AI data centers, so eight NYCs.”
As AI data centers continue to grow, traditional fiber-optic links will be increasingly utilized within these massive facilities. “Given the speed of the GPUs, and the distances that are now going on in these data centers, we think there’s an intersect point for that [fiber optics] technology, and that’s what we’re focused on,” says Smith. The future of AI is driving unprecedented changes in data center infrastructure.
Companies are racing to develop solutions that can keep up with the massive scale and complexity of these new AI-powered facilities.
April Isaacs is a news contributor for DevX.com She is long-term, self-proclaimed nerd. She loves all things tech and computers and still has her first Dreamcast system. It is lovingly named Joni, after Joni Mitchell.
