- 400G waves now make up the bulk of terabit-level bandwidth demand, according to a Zayo report
- While 400G is ramping, enterprises for now are still buying mostly 100G wave circuits
- Much of the interest for 400G is related to AI, but that’s not the only use case
The 400G wave is in full swing as enterprises hunger for more bandwidth in data center interconnect (DCI), metro networks and the like. And terabit capacity is climbing alongside wavelength demand, according to data from Zayo.
Indeed, 400G wave purchases now account for the highest amount of capacity purchased by terabit, Zayo noted in a report. For context, one terabit of capacity is enough to transmit data like millions of document files or full libraries of high-definition video “in milliseconds.”
Although the 400G figure is based on Zayo’s customer activity, “it aligns with a growing preference for higher-bandwidth transport solutions to support rapidly scaling workloads,” said Bill Long, Zayo’s chief product and strategy officer.
Enterprises today are still mainly using 100G wave circuits for transferring large amounts of data, but we can expect to see 400G wave adoption surge by 2029, Ciena and Vertical Systems Group (VSG) have noted.
Long similarly told Fierce while 400G usage is ramping, 100G waves remain “the most widely purchased bandwidth size” for industries like manufacturing, data centers and finance. 100G circuits also currently make up three-quarters of the U.S. wave market, per VSG’s data.
The usual suspects are leading the charge for more waves. Zayo found nearly 62% of all wavelength capacity in 2024 was purchased by just ten buyers – all hyperscalers or carriers.
A typical “big deal buyer” (i.e., a customer seeking more than 1 terabit of wave capacity) made 15 purchases on average in 2024, said Zayo, compared to 11 such purchases in 2020.
Dark fiber – and lots of it – is another key element of the bandwidth equation. Zayo’s report noted long-haul dark fiber purchases increased 52.6% year-over-year from 2023 to 2024, and operators like FiberLight and Windstream have noticed similar purchasing trends.
Who needs 400G – and why?
We’re well aware hyperscalers and tech companies are interested in 400G to support AI workloads. But other sectors have plenty of 400G use cases as well, Long pointed out, and they’re not necessarily AI-specific.
Manufacturers, for example, increasingly rely on real-time data from thousands of IoT sensors for robotics, quality control and environmental monitoring. They can also use high-throughput, low-latency backhaul for digital twins, creating “large-scale 3D modeling and simulation data” to help with predictive maintenance.
The finance sector meanwhile requires 400G to support growing trading volumes, Long explained, and hospitals similarly need that much bandwidth to securely move large medical files and patient records across their networks.
“Regardless of AI adoption, financial firms must consistently retain and transfer large volumes of sensitive data to meet regulatory requirements, making high-capacity transport like 400G critical to both performance and compliance,” he added.