Chicago vs Los Angeles data center: national reach or the Pacific side

If a third or more of your users are west of the Rockies, or you run game servers, voice or anything real-time for the Pacific side, put your server in Los Angeles. Everyone else should pick Chicago. It beats LA to 12 of the 15 North American and European cities we measured and keeps every North American city in our test under 50 ms. Run both if you want the West Coast under 35 ms without giving up Chicago’s numbers in the East and Canada.

That’s the short version. Here’s the data, so you can hold it up against your own traffic.

Round-trip times: Chicago against Los Angeles

The numbers come from RIPE Atlas, the public measurement network run by the RIPE NCC. On September 30, 2026, our network team had probes in 15 cities, roughly 15 per city on different local ISPs, ping our Chicago and Los Angeles sites, and each figure below is the median of those probes’ average round trip. Anyone can pull the raw results for Chicago and Los Angeles.

Treat Winnipeg (6 probes) and Calgary (8) as rough. Every other city had 12 or more.

Chicago takes 12 of the 15. Inside North America it’s 10 of 13, and the margins in the Northeast and eastern Canada are big: 42 to 47 ms on New York, Boston, Toronto and Montreal. Atlanta goes Chicago’s way by 22 ms and Miami by 19. London and Frankfurt are both about 40 ms better from Chicago, though if Europe is your main audience, neither of these sites should be your first pick.

LA wins three. Los Angeles itself is under 1 ms against 48, Seattle is 16 ms better and Vancouver 17. Dallas and Denver lean Chicago, but only by 9 and 4 ms. Denver’s gap is small enough that your own users’ ISPs could flip it.

The number I’d look at hardest is the worst case. Chicago’s worst North American city is 48 ms (Los Angeles and Vancouver), while LA’s is 73 ms, to Boston. Averaged over the 13 North American cities, Chicago comes out around 31 ms and LA around 47.

Microsoft’s Azure round-trip latency table makes a decent sanity check from a completely different network. Its West US region in California and North Central US region in Illinois sit 51 to 52 ms apart, depending on direction (median for the 30 days ending July 30, 2026). Those are Azure’s regions, not our network, but they line up with the 48 ms our LA users see to Chicago.

Why Chicago is the one-server answer

If you can only run one box and your users are everywhere, the question isn’t who’s fastest to any single city. It’s who’s least bad to the farthest one.

Chicago never goes over 50 ms in North America. Denver is 30 ms, Dallas 24, Toronto 14, and Seattle, two time zones away, is 44.

Canada is where the gap gets wide. Toronto reaches Chicago in 14 ms and LA in 61. Winnipeg is 20 against 60, and Montreal 24 against 67. Calgary also lands closer to Chicago (38 against 52), though with eight probes I wouldn’t lean on the exact figure. Only Vancouver flips to LA. Outside British Columbia, a Canadian audience is a Chicago audience.

Interconnection doesn’t tip it either way. PeeringDB lists 253 networks on Equinix Chicago and 265 on CoreSite’s Any2West, an exchange that spans Los Angeles and Silicon Valley. Both metros have plenty of networks to swap traffic with, so the choice really is about geography.

Our Chicago site is our own facility in Arlington Heights, IL, with transit from Cogent, GTT, PCCW and Zayo; builds and specs are on the Chicago dedicated servers page. If you’re weighing Chicago against the East Coast instead, or you’re building around the futures exchanges, we covered that in Ashburn vs Chicago.

Why Los Angeles wins the Pacific side and real-time work

For a web page, 48 ms against under 1 is noticeable but survivable. For anything real-time, think in frames.

At 60 fps a frame lasts about 16.7 ms, so a 48 ms round trip is nearly three frames between a player’s input and the server’s answer. From an LA server to an LA player, it’s a small fraction of one. Voice and video calls feel the same delay as talk-over and awkward pauses. Chatty APIs get hit hardest, because round trips stack: an app that makes ten sequential calls per request turns a 47 ms gap into almost half a second. If that’s your workload and your players skew West, look at our game server hosting in LA.

Then there’s the ocean. Several trans-Pacific cables come ashore around LA: TeleGeography’s map shows JUPITER (Japan and the Philippines), Southern Cross NEXT (Australia and New Zealand) and SEA-US (Guam, the Philippines and Indonesia) landing at Hermosa Beach, PLCN (Taiwan and the Philippines) at El Segundo, and Unity (Japan) at Redondo Beach. An LA server cuts the cross-country leg out of a trip to Tokyo or Sydney. It doesn’t shrink the Pacific. We measured that side too. In a second RIPE Atlas run on October 1, users in all 14 Asia-Pacific cities we tested reached LA faster than Chicago, by about 30 to 45 ms in most of them. Tokyo is 110 ms to LA against 144 to Chicago, Sydney 145 against 185, Singapore 182 against 228 and Hong Kong 157 against 195. Raw results are public for Chicago and Los Angeles.

Downtown LA is also where West Coast carriers meet. The operator of One Wilshire, the carrier hotel at 624 S. Grand Ave., says its meet-me room holds 300-plus carriers and service providers. Our LA servers sit in Digital Realty LAX10 at 600 West 7th Street, a block away. The facility is SOC 2 Type II certified, and in-stock builds go live the same day or the next. Builds and order forms are on the Los Angeles dedicated servers page.

Risk: earthquakes and a hot grid vs Midwest weather

LA’s risks are well documented. The USGS puts the chance of a magnitude 6.7 or larger earthquake in the Los Angeles region at 60% over the next 30 years, based on the UCERF3 forecast. That’s not a reason to avoid LA. It’s a reason not to have only LA.

The grid is the other one. California’s grid operator called 11 statewide Flex Alerts in 2022, ten of them on consecutive days from August 31 to September 9, and demand set a record of 52,061 MW at 4:57 p.m. on September 6. The lights stayed on: CAISO says operators and consumers together avoided rotating outages. Still, a record peak in the middle of a ten-day alert run is the kind of week you plan around.

Inside the building, our LA facility runs N+1 redundancy on power and cooling. Outside it, the real answer is a copy of your data in another metro.

Chicago’s hazards are the Midwest kind: tornado warnings in spring, ice and deep cold in winter. None of them line up with California’s.

Our Chicago and LA sites are about 40 ms apart over the point-to-point circuits between them. That’s our internal link, not the path your users’ traffic takes. It’s plenty for asynchronous replication.

The decision rule

• West Coast users, real-time traffic or a Pacific audience: Los Angeles. A third or more of your users west of the Rockies, game servers, VoIP, or APIs that make many round trips per request.
• Users spread across the US and Canada, or one server for everyone: Chicago. Nothing in North America over 48 ms, and the Canadian numbers aren’t close.
• Both: primary where most of your users are, replica in the other city, linked over our private network. A 1 Gbps private network is included, and 10 Gbps is $20 a month on the order form.

Between these two cities, price is a wash: the entry build is the same Xeon E3-1230 v1 at $126 a month on both order forms. Ashburn’s entry is a different build, the $159 Xeon D-1541 Special with a 1 Gbps unmetered port, so don’t set it against the $126 figure. Every location’s builds are on our dedicated servers page.

Before you order, run mtr to both sites from a network your users actually sit on. A customer’s home connection will tell you more than your office fiber.

Not ready to order yet?

Tell us what you’re after and we’ll email you the build and the price. If it’s sold out, we’ll let you know when it’s back. Need it today? Call (800) 561-2656 and we’ll check what’s in stock.

FAQ

Tell us where your users are and what you run through the custom quote form, and we’ll say which site fits or price the pair.