Ashburn vs Los Angeles data center: pick a coast, or pair both

If your users are east of Dallas, or in Europe, put your server in Ashburn. If they’re west of it, put it in Los Angeles. Dallas itself is a dead heat at 33 ms to both sites, which makes it the line to check your own traffic against.

And if you want two sites that share almost nothing, this is the pair. They sit on opposite coasts and separate power grids, with Chicago in between if you ever need a third.

Round-trip times: East Coast site against West Coast site

We measured this instead of guessing. On September 30, 2026, about 15 RIPE Atlas probes per city, each on a local ISP, pinged our Ashburn and Los Angeles sites. The table shows the median of their average round trips, which is close to what a typical user in that city gets. Raw results: Ashburn and Los Angeles.

Winnipeg had only 6 probes and Calgary 8, so read those two rows as rough. Calgary’s 4 ms gap is inside the noise.

Ashburn takes 9 of the 15 cities and Los Angeles takes 5, with Dallas tied. Inside North America it’s 7 for Ashburn, 5 for LA, one tie. The margins matter more. When a coast wins, it usually wins by 30 to 60 ms, so a wrong guess costs real time on every request.

The worst cases tell the same story. Ashburn’s worst North American number is Vancouver at 67 ms, and LA’s is Boston at 73 ms. Neither is a good one-server answer for the whole continent; for comparison, Chicago’s worst North American city is 48 ms.

For a second opinion with a month of data behind it, Microsoft’s Azure round-trip latency table puts its West US and East US regions 67 to 69 ms apart, depending on direction (median for the 30 days ending July 30, 2026). That’s Azure’s own cloud regions talking to each other, not our network, and not a figure for any link between our sites. The same table has UK South reaching East US in 78 ms and West US in 140 ms, which lines up with our London row.

Where the line falls: Dallas

Dallas is 33 ms from both. That’s the hinge.

Go a little west and LA pulls ahead: Denver is 34 ms to LA against 45 to Ashburn. Go east and it isn’t close. Atlanta is 17 ms to Ashburn and 46 to LA, and Miami is 29 against 60.

Here’s the test I’d run. Pull a month of analytics, group sessions by state and draw the line through Texas. Texas and everything east of it goes in the Ashburn column; Mountain time and the Pacific states go in the LA column. The bigger column is your primary site. If the two come out roughly even, you’re really a national audience, and that’s a different question (see the FAQ).

Europe doesn’t move the line. London and Frankfurt are 54 to 58 ms better from Ashburn, and nothing you do on the West Coast closes that gap. If Europe is a real share of your traffic, add it to the Ashburn column and stop thinking about it.

Ashburn: the East, Europe and the big clouds

I’ll keep this short, because our Ashburn vs Chicago comparison covers Ashburn in depth. Our Ashburn dedicated servers sit in Equinix DC3 in Loudoun County, Virginia, and we’ve run servers in Ashburn since 2014. AWS us-east-1 and Google’s us-east4 are in Northern Virginia and Azure’s East US is in Virginia too, so a bare-metal database feeding services in any of them stays a short hop away.

Ashburn’s entry build is the $159 Xeon D-1541 Special: 64 GB of RAM, a 1 TB SSD and a 1 Gbps unmetered port. It’s a different machine from LA’s $126 entry server, so don’t read the two prices as a location premium. The Special can be ordered in LA too.

Los Angeles: the West, the Pacific and the other cloud corner

LA owns the West Coast by wide margins: 28 ms from Seattle, 31 from Vancouver, 34 from Denver and under 1 ms across the metro itself. Phoenix, Las Vegas and the Bay Area aren’t in our probe set, but they’re a shorter trip to LA than to Virginia.

It has cloud neighbors too. Google Cloud’s us-west2 region is in Los Angeles, per Google’s region list. AWS runs two Local Zones in Los Angeles, us-west-2-lax-1a and us-west-2-lax-1b. Those aren’t a region of their own; they’re extensions of US West (Oregon). Put the database on bare metal in LA, run the stateless pieces in us-west2 or the LA Local Zone, and you’ve built the West Coast version of the classic us-east-1 hybrid.

Then there’s the Pacific, where Ashburn never comes out ahead. Trans-Pacific cables come ashore in LA’s South Bay: TeleGeography’s map lists Hermosa Beach as the US landing for JUPITER, to Japan and the Philippines, and for Southern Cross NEXT, to Australia and New Zealand. Serving Asia from Virginia means crossing the country before the ocean even starts. Our October 1 RIPE Atlas run puts numbers on it. Users in 13 of the 14 Asia-Pacific cities we tested reached LA faster than Ashburn, by 43 to 70 ms everywhere except Bangkok. Tokyo is 110 ms to LA and 172 to Ashburn, Sydney 145 against 202, and Seoul 139 against 209. The exception is Jakarta, a dead heat at 233 ms; it looks like many Indonesian networks send US-bound traffic toward the East Coast. Raw results: Ashburn and Los Angeles.

Our Los Angeles dedicated servers run in Digital Realty LAX10, a SOC 2 Type II building at 600 West 7th Street downtown. The entry build is $126 a month for a Xeon E3-1230 v1 with 16 GB of RAM, and in-stock builds are online the same day or the next. If you’d rather ship your own hardware, Los Angeles colocation is quoted per deployment and costs more than our Chicago space.

Running both: the coast-to-coast pair

Ashburn and downtown LA run on different utilities and different grids, and California’s earthquake risk stays in California. A storm in Virginia and a heat wave in Los Angeles are separate events.

Compliance doesn’t change between the coasts. We’re SOC 2 Type II certified. Customers building HIPAA or PCI environments run them on our servers under their own compliance program.

The cost is distance. Ashburn and LA are about 52 ms apart over the point-to-point circuits between our sites, so plan on asynchronous replication only. Synchronous replication would add roughly 52 ms to every commit, and your users would wait with it.

Chicago sits in between, about 24 ms from Ashburn and about 40 ms from LA over the same circuits. That makes it a natural third site, or a closer replica if 52 ms is too long for your write path. Every build in all three cities is on our dedicated servers page, and the private network between your servers is included at 1 Gbps, or $20 a month for 10 Gbps.

Run the pair active-passive. One coast takes all the writes; the other holds a replica and a deployment that’s ready to take traffic when you flip DNS. Two-coast active-active looks great on a whiteboard and usually ends in conflict-resolution code nobody wants to own.

The decision rule

• East of Dallas, or in Europe: Ashburn. You save 29 to 57 ms per round trip to the East Coast and eastern Canada, and 54 to 58 ms to London and Frankfurt.
• West of Dallas, or facing the Pacific: Los Angeles. You save 11 to 61 ms across Denver, Seattle, Vancouver and LA itself.
• Users on both coasts and no tolerance for downtime: both. Primary where most of your users are, replica on the other coast, asynchronous replication over the private network.

If you’re unsure which column you’re in, the Dallas test above settles it in an afternoon.

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

Want a pair across both coasts priced as one setup? Send the workload through our custom quote form and we’ll spec both ends.