What a gigabyte of travel data actually costs.

What this index measures

One number, held constant: the price of one gigabyte of usable mobile data for a traveller, in one region, from a provider that sells to consumers directly.

It is not a survey of headline prices. A headline price on a travel data plan is nearly meaningless, because the four products in this market are sold in four different units.

Some are metered gigabytes with an expiry date. Some are metered gigabytes with no expiry date. Some are day passes with an undisclosed throttle. Some are monthly subscriptions with a fair use policy that is not published as a number.

The index converts all of them to one unit, dollars per gigabyte, and then publishes the working so you can disagree with it.

The method

Six steps, run the same way every month. Anybody with a browser and a spreadsheet can repeat it, which is the point.

• Take the lowest publicly listed consumer price for a metered plan covering each destination, from each provider, on the measurement date. No affiliate rates, no promotional codes, no first purchase discounts.
• Divide the price by the included data to get a raw cost per gigabyte. Prices in other currencies are converted at the European Central Bank reference rate for that day.
• For each region, take the median across every destination in that region. A median rather than a mean, because one very expensive island would otherwise drag a whole region.
• Exclude unlimited day passes and subscriptions from the raw comparison, because a plan with no gigabyte count cannot honestly be expressed in dollars per gigabyte. They are discussed separately further down.
• Record the sample size, which is the number of destinations behind each regional figure, and print it in the table next to the number.
• Publish the date measured, and recompute on the 24th of the following month.

The number that is not on the sticker

A gigabyte you paid for and could not use is not a gigabyte. It is a donation.

Every major travel eSIM plan except ours carries an expiry date, usually 30 days. Industry estimates put wasted purchased travel data at 20 to 35 percent, and that waste is revenue for the seller.

So the raw index above understates the difference. Adjust for expiry at the conservative end of that range, 25 percent, and the effective cost per usable gigabyte moves in a way the sticker price hides.

The table below works the adjustment for a 10 GB Europe plan. The assumptions are stated so you can change them. Set the waste to zero and the effective column collapses back onto the raw one, which is the arithmetic every expiring plan in the table depends on.

Why regional prices differ at all

Travel data is bought wholesale from local carriers, and wholesale rates are set by conditions the traveller never sees.

Europe is cheap because roaming regulation forced wholesale caps across the single market, and because dozens of carriers compete inside one regulatory frame. This is why Europe sits at $0.89 while everywhere else is above it.

Asia is mixed. Singapore, South Korea and Japan are cheap and fast. Nepal, Myanmar and Mongolia are neither, and they pull the regional median up.

Africa and the Caribbean are expensive for the same reason: a small number of carriers per country, high backhaul costs, and in island markets a real dependence on submarine cable capacity.

Oceania is the most expensive region in the world for travel data, and the reason is simply that three carriers serve a continent and none of them are short of demand.

Cost per usable gigabyte, worked for a 10 GB Europe plan

Effective cost assumes 25 percent of an expiring plan goes unused, the conservative end of the 20 to 35 percent range that industry estimates give for wasted purchased travel data. If you use every gigabyte before expiry, the effective column equals the raw column and the advantage disappears. Listed prices taken on 24 August 2026.

What to read next to this index

Cost per gigabyte is one axis, and buying purely on it would be a mistake.

A cheap gigabyte at 1 Mbps is worse than an expensive one at 60 Mbps. Independent testing measured Airalo at 10 to 22 Mbps and Holafly near 7 Mbps before throttling and under 1 Mbps after.

Unlimited plans are excluded from the raw table because dollars per gigabyte is undefined without a gigabyte count. What they cost, and how much of each one runs at full speed, is in the unlimited table above.

Coverage breadth is a separate axis again. Ours is 197 destinations from one balance, each with a coverage page that names the carriers behind it, so the destination can be checked before the price is.

So read this index next to the speed figures rather than instead of them.

How often this updates, and what we commit to

Recomputed on the 24th of every month and republished the same day, with the measurement date stamped on the page and returned in the structured data.

Every region is recomputed in the monthly run, and the archive of previous months stays online.

The dataset is also readable by machines. The same figures are emitted as structured data alongside the table and exposed through our Model Context Protocol endpoint, so an assistant can quote them with the date attached.

If you find an error, write to press@orbislo.com with the listing you are looking at. We correct within 48 hours and note the correction on the page.

The three largest providers on the axes that are not price

Competitor figures are from third party testing and from each provider's public listings on 24 August 2026. Where a provider does not publish a number, we write that they do not publish it rather than estimating one on their behalf. The unlimited column was rechecked on 31 August 2026.

The cruise listing that does not add up

Maya's own cruise page shows 30 days for $50.89. Resellers list what appears to be the same product at $300.89.

Maya's own page was fetched again and carries the same global prices as the rest of their catalogue, with no cruise premium at all. The $300.89 figure appears only on reseller listings and is nowhere on Maya's own estate, so we treat it as stale rather than as a second price.

What is still open is not the number. Maya markets cruise coverage across twenty plus cruise lines and publishes nothing about whether it works at sea or only in port, and neither does anybody else selling a cruise plan. So this index counts destinations on land, the same way for every seller in it.

Questions about the method