eSIM for India: Why Aadhaar Is Closed to Visitors

India asks a question before it asks for money, and the question is about identity. Every mobile connection in the country is verified against a named person and an address before it works, and the fast route through that check runs on Aadhaar, the twelve-digit biometric identity number that only residents hold. A visitor cannot use it, so a visitor goes through the slow route, which is paper, a counter and a wait.

That is the whole reason a travel eSIM exists in this market rather than being a convenience. It is not cheaper than a local line and it does not carry a number. It is a data profile issued abroad to a roaming subscriber, and the verification regime has nothing to attach to.

The prices below are recomputed from the 186 India plans ten brands sell, on a snapshot taken on 24 October 2026. Two of them carry most of the argument. A month of 20GB costs $12.00, and a month of 50GB costs $43.00, so the larger bucket is the worse buy by a wide margin, which is the opposite of how most markets are priced.

Why the Aadhaar shortcut is closed to visitors

Because the number is issued to residents, and a tourist visa is not a route to one. Aadhaar is built for instant digital verification, and a resident activates a new line in minutes with it. The same line, bought by a passport holder, goes into a different queue.

The Department of Telecommunications requires the identity and address of every customer to be verified before a connection is activated, and that rule does not bend for a visitor. It also does not bend for money. No retailer can skip the check, and none of them pretend otherwise once the form is on the desk.

What a visitor supplies instead is a passport with at least six months of validity, a visa, a local address that a hotel booking normally satisfies, and a photograph taken at the counter. The documents are then checked in a back office rather than at the desk, which is where the delay comes from. Two hours is a good outcome. A day is normal. An arrival at two in the morning usually waits until the office opens.

There is a second limit that catches long stays. A connection issued against a foreign passport runs for the length of the visa or ninety days, whichever ends first. A traveller on a six-month visa still gets ninety days, and a traveller who extends a stay discovers the line expires before the trip does.

The keyword data for this market says something about how well this is explained. The phrase best eSIM for India carries a difficulty of 22 and the travel variant a difficulty of 17, which are the two lowest scores in the whole cluster. Nobody with authority is answering the question, which is why the search volume keeps going to pages that list brands without explaining the trade.

What the documents have to say

The passport and the visa must agree, and they must agree exactly. A middle name on one and not the other, an address that does not match the one on the visa, or an expiry date read the wrong way round is enough to send the application back.

That is the practical failure mode, and it is not appealable at the counter. A rejected application is re-entered at an official store with the documents copied exactly, which is a second morning rather than a second attempt on the same form.

What the paperwork costs in hours

A morning, in the ordinary case, and the morning is the part that a price comparison never shows. The counter opens when it opens, the queue forms before it, and the verification clears after the traveller has left the desk.

The cost lands hardest on the first day. An arrival in Delhi or Mumbai at an hour when the counters are shut leaves a traveller with no data for the ride, the map or the message home, and the airport wifi that fills the gap wants a code sent by text to a number that does not exist yet. A eSIM installed before the flight is on a network before the immigration queue moves.

There is a second cost that arrives at the end. A local line bought for a trip has to be kept alive or closed, and a line opened against a passport does not follow a traveller home in any useful way. For a two-week visit, that is a purchase with a built-in disposal problem.

What the counter sells is a number, and the number is worth something specific. Train bookings, food delivery, ride-hailing accounts and a long list of local forms verify by text to an Indian line, and none of them accept anything else. A traveller who needs one of those has to go through the queue. A traveller who needs maps, messages and a card payment does not.

What a gigabyte costs against Thailand

India is one of the dearer markets priced here, and the comparison that makes it visible is the one with the cheapest. The median Indian gigabyte costs $2.15, against $1.05 in Thailand, so the same money buys half the data. The best rate here is $0.57 against $0.24 there, a multiple of 2.4.

Each price is the cheapest plan that brand sells at that exact size, so the columns compare ten sellers rather than every window any of them lists. Two of the cheapest figures in that table are one-day plans, which makes them the wrong purchase for anyone staying a week.

Look at the spread rather than the price. The 5GB band runs from $7.50 to $14.00, a multiple of 1.87, and the 10GB band from $12.50 to $24.00, a multiple of 1.92. The 20GB band runs from $12.00 to $55.00, a multiple of 4.58. The small sizes sit inside a narrow band and the 20GB size does not, which is the shape that produces the inversion in the next section.

The one place India looks cheap is at the very top of the shelf. A 60GB plan with 365 days of validity costs $34.00, which is $0.57 per gigabyte, the lowest rate sold into the country. It is a resident’s or a repeat visitor’s product rather than a holiday purchase, and it is still the best value on the page.

Why the bigger bucket is the cheaper one

At thirty days, the three middle sizes do not rank the way they should. 5GB costs $12.47, 10GB costs $13.00, and 20GB costs $12.00. The largest of the three is the cheapest, and the smallest is 47 cents behind it.

This is not a stale listing or a rounding artefact. The 20GB band is where the range is widest, with twelve plans from nine sellers spanning 4.58 times, and the cheapest entry has been pushed below the 10GB entry rather than merely towards it. Ubigi sells 20GB over thirty days for $12.00, and Roamic sells 10GB over the same window for $13.00.

The practical consequence is blunt. In India, do not buy a smaller allowance to save money, because the saving does not exist at thirty days. A traveller who estimates 10GB and buys 10GB pays a dollar more than the traveller who buys 20GB.

The one exception is the 50GB tier, which breaks the pattern in the other direction. At $43.00 for a month it costs $0.86 per gigabyte, worse than the 20GB month at $0.60 and much worse than the annual plan at $0.57. That tier exists for people who need the volume in thirty days, and it is the only size in the file where the price rises faster than the allowance.

What the annual plan is for

A 60GB plan with a year of validity at $34.00, which is cheaper in total than the 50GB month at $43.00 and delivers 10GB more. For a single two-week trip the volume is absurd. For a repeat visitor, a long stay or a trip measured in months, it is the cheapest purchase in the market by a distance.

Where the unlimited shelf tops out

98 of the 186 plans are flat rate rather than metered, which is 52.7 per cent, and the ceiling of that shelf is the highest price in the file. Holafly sells six plans, all unlimited, from $19.90 for three days to $111.90 for thirty. Nothing else sold into India costs more.

The floor is $48.00 for thirty days from Ubigi, so the same nominal product runs from $48.00 to $111.90, a multiple of 2.33. That is a narrow spread by the standards of this market, where metered sizes vary by 4.58 times, and it says something about what unlimited is worth here: the sellers agree with each other and disagree with the metered shelf.

The daily figures are what separate them. Ubigi varies its threshold plan by plan and publishes it on each plan page. Roamic publishes 3GB a day and 1 Mbps afterwards, on every plan. Saily publishes 5GB a day. Airalo and Jetpac publish 3GB. Nomad publishes 2GB. aloSIM slows to about 512 Kbps after 1.5 to 2GB. Yesim states 500MB to 2GB depending on the plan. Holafly publishes none, and independent testing in early 2026 measured its cut-off at 500MB to 1GB a day.

Read the cheapest unlimited entry against the month and the arithmetic turns. A single day of unlimited costs $4.54 at Yesim, which is the lowest unlimited price in the file and the most expensive way to buy a month, because thirty of them would cost $136.20. The entry price and the monthly price are not the same product measured two ways. They are two different purchases.

Which brand is cheapest per gigabyte

The ranking is roughly the reverse of the search volume. Airalo is the brand travellers look for by name and its best rate of $1.83 is more than three times Ubigi’s. Saily has the dearest best rate of the ten at $1.95, with eleven plans to choose from.

Ubigi is the only brand whose rate is built on a product a visitor can use more than once. Its 60GB year at $34.00 sets the $0.57, and the same brand sells 20GB over thirty days for $12.00, which is the cheapest thirty-day purchase in the file at any size above 5GB.

Two brands are worth naming for the opposite reason. Yesim sells 37 plans, the longest catalogue here, and its best rate of $1.46 is two and a half times Ubigi’s. Roamic sells 36 and holds 30 of the 98 unlimited plans, which makes it the largest single source of flat-rate capacity in the country, and its ceiling of $77.00 is well below Holafly’s.

The middle of the table is where most purchases land, and the differences there are small enough to ignore. Between Nomad at $0.86 and Roamic at $1.14 there is very little to choose, and the window matters more than the brand.

Does a data profile break UPI

It does, and this is the one place where a data-only product costs a traveller something real rather than merely a number. UPI is the payment rail that a very large share of Indian transactions run on, and it is bound to an Indian bank account and an Indian mobile number. A profile with neither cannot use it.

What still works is most of what a short visit needs. Cards are taken at hotels, at restaurants aimed at visitors and at larger shops. Cash works everywhere and remains the fallback for the small transactions where UPI is otherwise universal. Ride hailing and food delivery run on the account a traveller brought, with payment by card.

What does not work is the local layer. A market stall with a QR code, a train ticket bought at a counter that wants a UPI reference, a small guesthouse that prefers a transfer. Those are the moments where a local number would have paid for itself, and they are the reason the queue at the counter still has a queue.

What UPI actually needs

Three things, and a travel eSIM supplies none of them. An Indian bank account, an Indian mobile number registered to the same person, and an app that verifies against both. There is no workaround that does not begin with the registration described above.

Does the signal hold outside the cities

Every plan sold into India rides Jio and Airtel together, which the data records for all 186 plans, and both are national carriers. In the metros and along the main rail corridors the connection is good, and a traveller moving between Delhi, Agra, Jaipur and Mumbai will not notice which of the two is serving them.

It thins in three places. The hill states, where terrain is the constraint rather than the network. Parts of the north-east. And the remoter interior districts, where the distance between masts is measured in tens of kilometres. Where the two carriers diverge, a plan that carries both is worth more than one carrying a single carrier, and every plan here carries both.

The practical advice is to treat a long rural leg as a case for a larger allowance rather than a different provider, because the provider is not the variable. The handset is, more than most visitors expect. A phone bought outside India may not support the bands the two carriers use, and the compatibility check settles that by model and region in about a minute.

Can you install it before you land

Yes, and the profile sleeps until it first attaches to an Indian network, so installing a fortnight early costs nothing in validity. What it buys is the arrival, and the arrival is the part of an Indian trip that punishes preparation least kindly.

Three settings decide what happens after landing. Point mobile data at the profile and leave roaming on, since a travel plan registers as a visiting subscriber. Name the profile, because two lines called Travel is a coin flip in a settings menu. Leave the home line enabled for texts with its own roaming switched off, because the one-time code for the card purchase and the bank alert both go there. Running both lines is the arrangement that keeps a number and a data profile in the same handset.

The purchase itself is worth doing at home for the same reason. The card has to clear a foreign merchant, and the authentication code goes to the home number, which is in the slot until the profile is installed. A purchase started after the home SIM comes out stalls at the last step.

Emergency calls are the exception to all of it. 112 connects from any network the handset can see, with a plan or without one.

Does the plan reach Nepal and Sri Lanka

No, unless the product names the country. An India-only plan is a single-country product and it stops at every land border, which matters here more than in most of Asia because the itineraries that start in India so often do not end there.

The usual pairings are Nepal by land from Varanasi or Delhi, Sri Lanka by air from Chennai, and the Gulf on the way home. Each needs its own product or a regional bundle whose country list names it. The Asia coverage list compares which bundles name which countries, and the labels do not agree: products sold under the same description carry different lists.

A regional bundle is priced for its widest country, and India is one of the dearer markets in the region, so the arithmetic often favours two separate plans. For a short second stop that is almost always the cheaper route.

What fails in the first week

Three causes account for almost every dead connection, and none of them is the seller. Data is pointed at the home line. Roaming is off for the profile. The profile attached to a network but not to data.

Each takes under a minute to correct, and each is worth ruling out before writing to support. What to send when writing is the order number, the handset model, the step that fails and a photograph of the settings screen.

The reason to write on day one is the replacement rather than the refund. A profile reported inside twenty-four hours is usually replaced at no charge, because it is measurably unused. The same report on day five meets a plan that has already run for five days.

One failure is specific to this market and looks like a fault. A handset that shows bars and no internet is usually still holding the home line for data, because the phone kept its original data route when the second profile arrived. Switching the data line explicitly, rather than trusting the automatic choice, fixes it.

The comparison by trip length prices the window against the days actually spent, which is the faster way to size a trip than estimating apps one at a time, and the fair-use table carries the published daily threshold for every unlimited plan sold here.