Understanding where the Internet physically connects is critical for infrastructure planning, policy decisions, and business strategy. This dashboard tracks Internet Exchange Points (IXPs) — the neutral facilities where networks meet to exchange traffic directly. Whether you're evaluating markets for data center investment, planning regional connectivity initiatives, or researching the state of global internet infrastructure, this live data provides the foundation.
Global Internet Exchanges? Physical locations where multiple networks connect to exchange traffic directly.
1,276
+9 vs last monthGlobal Facilities with an IX? Data centers or colocation facilities that host at least one Internet Exchange.
2,115
+9 vs last monthGlobal Networks / Peers? Autonomous networks (ISPs, content providers, enterprises) registered in PeeringDB.
34,459
+270 vs last monthData as of February 1, 2026 via PeeringDB
Historical Growth
Internet Exchanges by Region
Africa
Asia Pacific
Australia
Europe
Middle East
North America
South America
North American Internet Exchanges
| Rank | State | # of IXs? Number of unique Internet Exchanges present across all facilities. | # Facs w/ IX? Number of facilities with at least one Internet Exchange. | # Nets (Phys)? Number of unique networks physically present across all facilities. |
|---|
| Rank | Province | # of IXs? Number of unique Internet Exchanges present across all facilities. | # Facs w/ IX? Number of facilities with at least one Internet Exchange. | # Nets (Phys)? Number of unique networks physically present across all facilities. |
|---|
Subscribe to receive this monthly report via email
Get updates on global Internet Exchange trends directly in your inbox.
Frequently Asked Questions
What is an Internet Exchange? ▾
An Internet Exchange (IX or IXP) is a physical location where multiple networks connect to exchange traffic directly, rather than routing through third parties. This reduces costs, improves performance, and lowers latency for end users.
IXPs are typically housed in carrier-neutral data centers where no single network has control — creating a level playing field for all participants. Networks connect via switches and exchange traffic through peering agreements.
Learn more in the NV Research AI Interconnection book/interview →
What is the source of this data? ▾
All data on this page is sourced from PeeringDB, a freely available database of networks and interconnection facilities maintained by the global network operator community.
PeeringDB data is self-reported by network operators and facility owners, and is updated continuously. Our statistics on this page are refreshed monthly to track trends over time. The data on many other pages of this site is refreshed nightly.
What do "Facs" and "Nets" mean? ▾
Facs (Facilities) refers to the physical spaces where network equipment is housed. A single building or address may host multiple facilities. "Facs with IX" counts facilities that host at least one Internet Exchange.
Nets (Networks) refers to autonomous networks, also known as Autonomous Systems (AS). These include Internet Service Providers (ISPs), content delivery networks, cloud providers, and enterprise networks that have registered in PeeringDB.
Why does Internet Exchange location matter? ▾
The physical location of Internet Exchanges directly impacts network performance, cost, and resilience for nearby users and businesses:
- Latency: Local IXs reduce the distance data must travel, improving response times
- Cost: Direct peering at local IXs reduces transit costs for ISPs
- Resilience: Multiple local interconnection points prevent single points of failure
- Economic development: IXs attract network investment and digital businesses to a region
Regions without Internet Exchanges often see their traffic "tromboned" through distant cities, adding latency and cost.
Why are internet exchanges important for AI inference? ▾
Real-time inference and edge inference workloads require the lowest possible latency to deliver instant responses. Internet exchanges provide direct peering between networks, minimizing the number of network hops and reducing round-trip time.
This is critical for real-time inference applications like conversational AI and autonomous vehicles, as well as edge inference deployments that process data closer to end users rather than in centralized cloud data centers.
Cite This Data
When referencing this data in publications, reports, or presentations:
