Summary

  • DE-CIX says traffic at its Chicago exchange grew almost 300% in twelve months, passed a 1 Tbit/s peak in August, and now ranks second in the city market.
  • Its figures describe adoption, connected capacity and facility reach. They do not establish that any participant has a physically independent route or a tested failover path.

On 13 October, DE-CIX said traffic at its six-year-old Chicago Internet Exchange had grown almost 300% in the previous year. The company described the exchange as the second-largest in the Chicago market and said it had climbed from 58th to 29th in North America in PeeringDB between 2023 and 2026. That ranking is the company’s account of PeeringDB data, not an independently recalculated comparison.

DE-CIX also said the exchange first passed 1 Tbit/s of peak traffic in August, with more than 85 networks “connected and connecting” and over 8 Tbit/s of connected capacity. Its platform is accessible from 18 data centers across metropolitan Chicago, including partner locations. These are signs of a growing interconnection option. Each describes a different thing: a peak is a momentary exchange traffic rate; connected capacity is what participants have provisioned; and a network count does not show how much traffic each one exchanges or whether its route is independent of another participant’s.

The company’s own public PeeringDB record provides a useful, but older, comparison point. It lists 84 peers, 96 connections, 51 open peers and 6.3 T of capacity, while showing a last-update date of 6 August 2024. Those fields do not share a confirmed measurement date or necessarily match DE-CIX’s newer “connected and connecting” and capacity definitions. They therefore cannot be treated as an October 2026 reconciliation. PeeringDB describes itself as a user-maintained interconnection database; the operator’s release remains the source for its own current claims.

DE-CIX says its platform spans 18 data centers and gives networks access to a wide range of content, cloud, AI, carrier and enterprise participants. Vice president Ed d’Agostino argues that a second large neutral exchange can create more path diversity and strengthen competition. That is a plausible design benefit, not proof that a given network can survive the loss of a facility, metro carrier, duct, power feed or route-server dependency. The exchange connects networks; each network retains its own routing policy, and the physical services that reach an exchange may share failure points.

For an operator, the useful test is therefore participant-specific. Which facilities host its own routers? Do two circuits use different buildings, entrance ducts, metro providers and power paths? Are routes learned through a route server, bilateral sessions, or both? Does a second session lead to a usable alternate route when the primary transport fails? The release does not answer those questions, and it does not claim that every participant has duplicated access.

The growth announcement shows that more networks are choosing to connect and that the platform has expanded its metro reach. It makes the case for examining Chicago’s second large exchange. It does not settle whether path diversity is present at the points where operators need it. The distinction matters because an alternative that exists in the market is only an operational alternative when a network can reach it independently and use it under stress.

Sources