AI traffic now represents nearly one-quarter of network flows; new metric reveals order-of-magnitude shift in data movement patterns.
Backblaze, Inc., the high-performance cloud storage platform for the AI era, announced the release of its latest Network Stats Report, revealing how AI workloads are reshaping global data flows to the extent that the company is building a new metric.
“The way data moves is fundamentally changing,” said Brent Nowak, Technical Lead Network Engineer at Backblaze. “As AI training and inferencing accelerate, data is no longer just growing—it’s flowing differently.”
The report examines the unprecedented rise of AI-driven data gravity and provides a first look at the telemetry metrics shaping digital infrastructure in the age of generative AI.
“The way data moves is fundamentally changing,” said Brent Nowak, Technical Lead Network Engineer at Backblaze. “As AI training and inferencing accelerate, data is no longer just growing—it’s flowing differently. Our Network Stats dataset helps quantify that shift so customers can better understand, anticipate and design for it.”
The Rise of AI-Driven Data Movement
Backblaze’s Q3 2025 dataset highlights a striking pattern: traffic associated with “neoclouds” (cloud providers offering compute, GPU or other AI-related services) already represents nearly one-quarter of total ingress and egress across Backblaze’s global network. Historically, CDN traffic dominated as Backblaze’s B2 Object Storage offering grew by hosting enterprise backup and storage. Now, AI-related workloads represent a new class of traffic reshaping the company’s network profile.
Unlike traditional workloads, AI workloads move in short, intense bursts tied to dataset replication, model training, and inference pipelines. This activity is concentrated in a smaller number of IP addresses—fewer talkers, bigger flows—reflecting the high-density, high-throughput nature of AI compute clusters.
To help quantify this shift, Backblaze introduced a new metric: the “magnitude” of traffic, which reveals how much traffic each unique endpoint carries as a proxy for network intensity. The report shows that AI-related traffic flows are an order of magnitude more intense than traditional patterns, offering an early signal of how the infrastructure beneath the AI economy is taking shape.
Evolving Network Transparency for the AI Era
This marks the beginning of a new phase for Backblaze’s Network Stats series, which now joins its long-standing Drive Stats reports as a regular, quarterly transparency initiative. The dataset tracks bits, packets, flows, IP address volume, traffic categories and protocol-level analysis—all with average, 95th percentile and maximum values.
Beyond aggregate metrics, Backblaze will begin publishing its raw network telemetry data quarterly to support independent analysis by researchers, partners, and the broader cloud ecosystem.
The Macro Trend: AI Data Gravity
The findings underscore a macro trend: AI-driven data mobility is reshaping the relationship between storage and compute. As organizations adopt multi-cloud architectures, data is increasingly following performance, cost, and proximity cues rather than remaining centralized in a single hyperscaler.
This architectural shift demonstrates how the open cloud breaks down walled gardens. Organizations no longer need to centralize everything in a single hyperscaler to avoid egress fees. Instead, they can combine best-of-breed solutions and scale workloads without compromising accessibility or budget control.
“Transparency around how data actually moves is critical for the next generation of builders,” said Gleb Budman, CEO at Backblaze. “Our mission is to make customers unstoppable by solving their toughest data storage challenges. By publishing these network insights, we’re helping customers understand not only how Backblaze operates, but also how the global data ecosystem is changing.”
Also Read: The End Of Serendipity: What Happens When AI Predicts Every Choice?
Comments are closed.