# Network Stats Q3 2026: Higher Baselines, Shifting Patterns > Source: > Published: 2026-10-22 13:00:00+00:00 July set a new traffic record on the Backblaze network, but the peak is only part of the story. Across Q3 2026, elevated CDN traffic, expanding neocloud activity, and renewed migration bursts show how demand is changing across our regions. This quarter’s [Network Stats report](https://www.backblaze.com/blog/tag/networkstats/) looks at both the scale and shape of that demand.  While we’ve been [tracking trends since December 2023](https://www.backblaze.com/blog/backblaze-network-stats/), this is the third quarter since we [operationalized the dataset](https://www.backblaze.com/blog/network-stats-for-q3-2025-the-magnitude-of-ai-workflows/) and re-launched the series, allowing ourselves to make direct, quarter-over-quarter comparisons. Why? Because AI workloads were changing traffic patterns across Backblaze’s network, and reshaping the internet.  With three quarters of historical data now available, we’re moving beyond measuring traffic volumes. We’ve been able to spot trends and start drawing conclusions—how predictable or unpredictable those workloads really are, and what that means for infrastructure that supports the next generation of AI applications. We’ve added quarter over quarter percentage changes to our regional heatmaps, making it easier to distinguish a consistently busy connection from a rapidly growing one. We’ll also look at the growth in private interconnections and the larger-capacity links supporting increasingly variable workloads. #### Join the Webinar Want to hear more? Join Brent Nowak, Manager, Network Engineering, and Stephanie Doyle, Sr. Manager, Market Intelligence and Keeper of Stats, live on October 29, 2026 at 10:00 a.m. PT / 1:00 p.m. ET to walk through the data and spot the latest trends.  Can’t make it live? [Register anyway](https://www.brighttalk.com/webcast/14807/678149) and we’ll send you the recording. If you’re interested in some of the trends we’ve spotted in previous reports, you can review the past reports here: ## The floor is rising The stacked area graph below shows total traffic by network type over time updated with the most current data. - **CDN traffic:** Activity reached a significant peak in July 2026 before moderating slightly in August and September. Despite the post-July dip, baseline CDN traffic remains elevated compared to prior months and significantly higher than same-time last year. - **Hosting traffic:** Continued its characteristic steady volume through Q3 consistent with its long-term baseline pattern over the past year. - **Hyperscaler traffic:** Surged to a strong peak in July 2026, coinciding with the broader network volume spike, before slightly contracting in August and September while remaining at a higher overall footprint than Q1/Q2 levels. - **Neocloud traffic:** Showed massive expansion in July 2026, pushing its individual volume footprint to record heights. Though it retracted slightly in August and September, neocloud remains one of the heaviest ongoing drivers of network volume. - **ISP-regional traffic:** Continued as a primary driver of volume alongside neocloud, expanding into July 2026 and maintaining a strong band through September, contributing heavily to the overall peak in July 2026. - **Migration traffic:** This traffic includes one way migrations into our environment, frequently via[Universal Data Migration](https://www.backblaze.com/cloud-storage/solutions/cloud-migration) . Re-emerged visually with prominent, distinct bands appearing in July, August, and September 2026, showing another burst of migration activity similar to the periods seen in late 2025 and March 2026. In 2026 traffic aggressively increased through Q2 and Q3 2026, setting a new all-time peak in July 2026, driven primarily by surges in ISP-regional, AI-focused neocloud, and CDN activity. Additionally, Q3 saw a distinct resurgence in active data migration spikes, while CDN traffic established a substantially higher baseline than the previous year. ## Heatmaps: How and where data moves To better understand our network activity, we isolated variables like region and types of provider. New for this quarter, we’ve added a notation at the bottom of each cell to call out the quarter-over-quarter (Q/Q) percentage change. This gives us a dual view: the absolute traffic concentration for the quarter (indicated by color intensity) and how that specific network type and region evolved over the last three months. Large percentage shifts signal potential industry or client workflow changes, while “no change” (N/C) in a dark-shaded cell points to a steady, heavy baseline. #### Quick terminology refresher **Standard definitions****Total traffic volume:** Where did we send and receive the most traffic? **Magnitude:** Where were the data transfers with the most bits per unique IP address?**Uniqueness:** What does the number of distinct IP addresses look like?**Regions****US-West:** Our largest and longest-running region**US-East:** Region with the most observed proximity to neocloud infrastructure**CA-East:** Our newest region in Canada. **EU-Central:** Our EU region. **Network Types****CDN:** Networks that use Backblaze as an origin store for content delivery. **Hosting:** Traditional hosting providers that runs workloads like physical or virtual servers for web, database, or application tasks.**Hyperscaler:** Large, traditional cloud providers.**ISP-regional:** Local or regional ISPs; think of these as the “last mile” paths as these networks are very close to customer equipment and efficient. **ISP Tier1:** National or international ISPs that carry our traffic long distances. Neocloud: AI-focused compute networks. ### Heatmap #1: Where did we send and receive the most traffic? ISP-regional traffic remains heavily concentrated in US-West—our darkest single cell alongside CDN—reflecting its status as our earliest deployed region with our largest internet exchange (IX) and server footprint. CDN traffic in US-East is a standout story this quarter, showing the highest overall volume on the heatmap alongside an impressive +85% Q/Q increase. Neocloud traffic shows heavy concentration across both US-East and EU-Central, with notable QoQ acceleration in US-West (+58%) and EU-Central (+46%). Additionally, EU-Central experienced dramatic growth in ISP-Tier1 traffic (+319% Q/Q), and US-East saw strong hosting traffic growth (+133% Q/Q). Keep in mind that, as we’ve said before, while we can draw connections to larger macro trends, this report isn’t a stand-in for global traffic patterns of all cloud providers—you can read that EU expansion as a shift in Backblaze’s customer set in that region. Since CA-East (Canada) is our newest region, its baseline traffic values remain relatively low; consequently, small shifts in traffic volume register as double-digit quarterly percentage changes (e.g., +44% in hyperscaler, +35% in neocloud). Contrast these low-volume percentage jumps with the +85% CDN growth in US-East or +46% neocloud growth in EU-Central—because those regions already handle massive traffic volumes, those high percentages represent substantial absolute bandwidth growth. A pattern for ISP-Tier1 traffic emerged in our EU-Central region, with an increase of +319%. The volume increase alone doesn’t tell us which workloads drove the change. But, magnitude and uniqueness can help us distinguish growth spread across many endpoints from heavier transfers involving a smaller set of addresses. ### Heatmap #2: Where were the data transfers with the most magnitude (bits per IP address)? Another metric we record is bits per IP or what we term “magnitude.” This combination of the amount of traffic transferred with how many actors are involved per network is a good proxy to measure how heavy or impactful individual data flows are. In short: - **High volume, many IPs:** Easier to distribute and load-balance across infrastructure. And many source and destination pairs means that we can traffic engineer at the WAN layer, sending some traffic over one provider and some over another. - **High volume, few IPs:** More difficult, but more interesting, from a NetEng perspective. While both US-East and EU-Central are regions with a high magnitude of neocloud traffic, we noticed a change in behavior over the latest quarter. There was a +37% increase in EU-Central, with an almost identical decrease in US-West. Other notable changes over the last quarter include the magnitude of CDN (+602%) and hosting traffic (+279%) in our US-East regions. Mirrored in the magnitude heatmap is the large increase in EU-Central ISP-Tier1 traffic. This points to new, impactful workflows in this region that equate to high volumes of traffic with a low amount of individual endpoints. ### Heatmap #3: How many unique addresses do we interact with? US-West continues to have the highest overall uniqueness—a quarter trend that has persisted for three quarters and is consistent with its larger data center footprint and substantial ISP-regional activity. As always, the consistency of the uniqueness metric is a good sanity check on our dataset. - US-West shows the highest overall uniqueness, driven by its larger number of data centers and mix of workloads. - Neocloud traffic, by contrast, tends to involve fewer, more persistent endpoints, consistent with AI pipelines that rely on stable, long-standing connections between storage and compute. A standout in the quarterly comparison numbers shows a large increase in the number of addresses for CDN networks in our CA-East region (+335%). ## Neocloud and hyperscaler traffic vs. predictive patterns This next set of charts shows a deeper dive into the metrics associated with neocloud and hyperscalers over time. The contrast between a more “traditional” workload (e.g., CDN, hosting, and ISP regional traffic) and emerging trends with neoclouds and hyperscalers is the easiest place to see the shift in network traffic profiles. The latter represents bursty, high magnitude traffic that reshapes conversations around network planning. ### Chart #1: What’s the magnitude of neocloud and hyperscaler traffic over time? Following a highly concentrated, low-magnitude baseline for both categories in January and February 2026, Q1 closed with a dramatic March surge where several individual neocloud networks spiked massively. Moving into Q2 2026 (April through June), while the absolute highest neocloud peaks compressed slightly downward compared to that March anomaly, the overall volume of high-magnitude neocloud workflows multiplied significantly, resulting in a much denser cluster of active endpoints staying consistently high quarter-over-quarter. Hyperscaler endpoints experienced a steady and noticeable upward move over the course of Q2, with multiple data points breaking out of their typical floor by May and June. In Q3 2026 (July through September), neocloud activity maintained its high-volume spread, led by a prominent breakout peak in July alongside a tightly packed cluster of mid-to-high endpoints. Activity moderated somewhat into August before rebounding in September with a spread of datapoints. Meanwhile, hyperscaler endpoints sustained their elevated performance floor into July before tapering into August and September, returning closer to baseline levels. Ultimately, neocloud retained its dominant, high-magnitude presence across Q1, Q2, and Q3, while hyperscalers peaked mid-year before consolidating into late Q3. ### Heatmap #1 and #2: How dynamic are neocloud and hyperscaler traffic patterns? US-East remains a major concentration of neocloud traffic, with growth from March through July. Across a broader reporting period, however, neocloud activity extends across US-East, US-West, and EU-Central. The regional picture is becoming more distributed. Hyperscaler traffic shows more variation from last quarter’s heatmap, including a modest shift toward US-East in Q3. Activity is still spread across our three regions. For network planning, those regional changes matter alongside total growth. We need to understand where demand is building, how concentrated it is, and how quickly it can change. These observations describe traffic on Backblaze’s network. We’re able to draw parallels to macro trends in the AI landscape, and it’s possible, even probable, that there’s a macro trend about geographical dispersion of AI data. But the changes also reflect growth in Backblaze’s customer base, changes in individual customer workflows, or a combination of all three factors. Because a small number of high-volume workflows can materially affect the results, we’ll keep tracking whether these patterns persist across future quarters. ### Heatmap #3, #4, and #5: How dynamic are CDN, hosting, and ISP-regional traffic patterns? We’re grouping CDN, hosting, and ISP regional types together because they represent a “steady-state” for us as network operators. These patterns are predictable, spread out over time, and generally do not change month-to-month. For Q3, we saw a large amount of CDN in US-East, causing a deep shade on our heatmaps not seen before. Hosting traffic is continuing to show a new pattern, with more activity in our EU-Central region starting in April into September. ## Bigger traffic calls for bigger connections Beyond mapping endpoint networks, our team evaluates *how* data travels across infrastructure. As a quick refresher, our connectivity methods include: - Transit via a Tier1 ISP - Internet Exchange (IX) - Private Network Interconnections (PNIs) Anticipating long-term network demand remains a core responsibility for the Network Engineering team. Throughout the past year, we have actively expanded our direct PNI footprint, allowing partners to maintain higher sustained throughput across these private links with no scaling cost. The chart below illustrates traffic distribution categorized by connection method for this year. Beyond steady expansion with transit carriers, our team continues to scale dedicated PNI interconnects. To accommodate this growing demand, we are shifting baseline provisioning standards—moving from traditional dual 100 Gbps link configurations to multi-100 Gbps and high-density 400 Gbps deployments. This has enabled us to handle and grow with the demands of variable traffic profiles associated with neocloud infrastructure. Below is a one year view of the growth across our different transit types, with the amount of traffic over PNI links leading the pack, with a +79% increase this year. Traffic over our internet carriers on transit links also has a sizable increase noted at +61% change this year. ## Conclusion Q3’s July peak stands out, but the more interesting story is what surrounds it: higher traffic baselines, growing neocloud demand, and regional shifts in how data moves. Those patterns are helping us understand the capacity our network needs. Growth across private interconnections and transit links shows why network planning involves more than adding bandwidth to a single path. Expanding our connectivity options and deploying larger capacity links helps us support sustained demand alongside bursts of concentrated activity. As the dataset grows, we’ll keep testing which patterns hold, investigating what changes, and sharing what we learn. What would you like us to explore next? Share your questions and observations in the comments, and subscribe to the Backblaze [Developer Newsletter](https://info.backblaze.com/tech-community-sign-up) for more technical reporting and practical guidance.