Mercado Pago and Ualá already play. The wedge is phone insurance without punishment, and Nubank already built it in Brazil
TL;DR
Nubank will not be the first to sell digital insurance in Argentina. Mercado Pago has done it since 2022 and Ualá entered in 2026 with Allianz inside its cap table. My recommendation is to enter anyway, with the product Nubank already built in Brazil and that everyone here sells badly: phone insurance.
The insurance line where that product lives bills $228.3 billion pesos of premium in 9 months, according to my calculation on financial statements filed with the SSN, the insurance regulator. Bank-channel insurers dominate it, and it punishes the user with deductibles of 25% to 40%, or with coverage caps that do not pay for half a phone. The wedge: be the first to sell it without punishment. Arriving first is no longer in play, and inventing the product is not needed. What it does demand is paying an anti-fraud operating cost that today the whole industry passes on to the customer. That part of my thesis is a bet, still without data behind it.
Independent analysis, not affiliated with any of the companies mentioned.
Argentine digital insurance already has serious players: the country's two largest wallets, with global insurers behind them.
Mercado Pago sells life and personal accident insurance in Argentina since the deal announced in October 2022: Prudential as the insurer, Swiss Re on reinsurance and the Argentine insurtech Klimber on the build. It replicated what the same partnership had launched in Brazil a few months earlier.
Ualá launched the same two products in 2026 and collected more than 300,000 quotes in a few weeks. In March it closed a USD 195 million round led by Allianz X, the Allianz group's investment arm, at a USD 3.2 billion valuation. The Allianz X CEO said: "in the future insurance won't be sold, it will be integrated into everyday financial experiences".
The 300,000 quotes do not prove conversion, but they do prove the condition this thesis needs: insurance demand can be activated from a financial app, with no sales force.
From Nubank, meanwhile, there is less than it seems. It confirmed a hub in Buenos Aires in January 2026, inside a global USD 474 million office plan. Asked by the press, the company did not confirm whether it resumes financial operations in the country. There was also a report by Valor Econômico (August 2025) about a failed negotiation to buy Brubank, never confirmed by the parties; Nubank replied that it does not comment on rumors. It is an unconfirmed press report, and no conclusion in this case rests on it.
The fintech-insurer partnership stopped being an idea and became the industry pattern. Chubb with Nubank in Brazil, Prudential and Swiss Re with Mercado Pago, Allianz with Ualá. That a fintech can sell insurance is already proven. What remains is deciding which product to enter with.
In the AMBA metro area around 10,000 phones are stolen per day, 416 per hour, according to the consultancy BTR (2024). Device theft is 27% of all robberies in the country and the most frequent victims are between 14 and 20 years old. For scale: Peru sits around 4,800 per day and Colombia around 3,500.
There is no lack of offers. Phone insurance is sold by manufacturers, the three telcos, Mercado Libre through La Caja and at least eight banks, according to La Nación's survey of August 2023. Mercado Pago itself started here: it launched phone theft and damage coverage in 2022, before its life products. The product is everywhere; what is missing is a version that does not punish, and the punishment takes two distinct forms.
Telcos punish through the deductible. Movistar covers 70% of the device's value on theft or total destruction: the remaining 30% comes out of the user's pocket. Claro charges a 40% deductible of list price when replacing after theft, and 25% on total damage. In general they do not cover loss, and they cap at two to four events per year.
Banks punish through the cap. The 2023 survey shows lower deductibles (BBVA, 10% of the insured sum) but insured sums that did not keep up with device-price inflation: maximum coverage around $180,000 pesos at Banco Nación or $66,000 at Supervielle at the time, against phones that already cost several times that, so the friendly deductible protects decorative coverage.
Run the math the user runs. Say a phone costs 800,000 pesos and the deductible is 30%: after paying the premium all year and having the phone stolen, you put 240,000 pesos out of pocket. The coverage exists on paper. The feeling of being covered does not.
Up to here, press and product pages. This section comes from a different source: the financial statements insurers file with the SSN, published as open data, with the same method I used in my case study on insurance collections.
The phone insurance products surveyed operate as theft coverage, so their natural home in the statements is the "Theft and Similar Risks" line. That line billed $228.3 billion pesos of premium between July 2025 and March 2026, 1.21% of the market, and closed the first quarter of 2026 with 3.32 million active policies: 9% of all property policies in the country.
Two things come out of that chart. First: six of the line's ten leaders sell through bank channels (Provincia, Galicia, Qualia of the Petersen group, Supervielle, Nación, BIND), and the top 3 alone concentrates 46% of the premium. The line is distributed through branches and home banking. There is no mobile-first player among the leaders.
The second comes from crossing it with my previous case: annulments eat 13.6% of the line's issued premium, against the 10.6% median I calculated for the market's top 50 with the same source and method. The line loses more premium to annulments than the market that contains it. My reading, and this is my interpretation on top of a verified figure: a product sold as a bundle through a channel the customer did not choose, and that punishes at claim time, leaves the customer no reason to keep it when reviewing expenses.
Here is the fact my recommendation is anchored on. Nubank would not have to build anything: Nu Celular Seguro exists in Brazil since February 2022, with Chubb, 100% in-app, covering theft and damage. Of the 2 million policy portfolio Nubank and Chubb announced in June 2024, more than one million is Nu Vida; the rest splits across the portfolio's other products: phone, home and installment protection. Entering Argentina with the phone product would be porting the piece of the portfolio that best fits the country's most massive risk.
One fact works against this recommendation:
Not even Nubank itself does a low deductible. Its Brazilian product page does not publish the deductible (you have to simulate in the app) and third-party sources place it around 25% of the device's value. My recommendation is a bet against the consensus of the whole industry, including the Nubank of Brazil.
The consensus exists for a reason that deserves respect before betting against it. The high deductible is the category's fraud control: when replacing the phone costs you 30%, inventing a theft loses its charm. Lowering the deductible removes that brake, and what replaces it is operations: checks of the IMEI, the phone's unique serial number, against the police report, behavioral signals in the app, repeat-claim blocking, fast resolution for the legitimate case and selective friction for the suspicious one. It is expensive and slow to build. For that same reason, once it works, nobody copies it in a quarter. The anti-fraud capability is not the product's cost: it is the moat.
And there is a signal that users feel the punishment: the line's 13.6% annulment rate. A product cancelled more than the market average is a product that is not retaining, and that is the incumbent the version without punishment would compete against.
Up to here every number came from a source or from a calculation on a source. The ones that follow are my projections.
The next step I propose is a 14-day smoke test with in-app traffic, before building anything: banner, landing, simulated quote, waitlist. The go criterion is set before running it, at 5% conversion from impression to waitlist. My model projects 8%. If the real number comes in under 5%, the thesis gets revised.
For volume at 24 months I built three scenarios anchored on Brazil's real trajectory, where the Nubank and Chubb attach rate went from 1.6% (1.4 million policies over 88 million customers in 2023) to 2.2% (2 million over 92 million in June 2024). My model projects between 15,000 and 120,000 active policies, with 60,000 as the base case.
The range is huge. The attach rate explains little of it: in Brazil it moved from 1.6% to 2.2% in a year and a half, so it lives in a narrow band and moves slowly. The distance between 15,000 and 120,000 policies comes from how many customers Nubank has in Argentina at launch.
The product: port Nu Celular with one deliberate difference from Brazil. A low, fixed deductible in pesos, shown on the purchase screen and not after simulating. Coverage for the device's real value, without decorative caps. Resolution committed in hours. And the anti-fraud operation treated as a core capability of the business, because it is what pays for everything above.
The timing: the sequence stays, because without a customer base there is no channel and without a channel the attach rate does not exist. What moves earlier is the preparation: the insurance partner and the product design close now, while the license and the base are still pending. When Nubank entered insurance in Brazil, Nu Vida launched in December 2020 and by February 2021 it already had 90,000 policies, with about half of those customers buying life insurance for the first time. That snapshot covers the first 90 days and cannot be extrapolated on its own, but it shows what happens when the product arrives ready to a base that already exists.
| Claim | Status |
|---|---|
| 10,000 phones stolen per day in AMBA | Published figure (BTR Consulting, 2024) |
| Deductibles: Movistar 30%, Claro 25-40% | Official product pages and La Nación (August 2023) |
| Bank caps ($180,000, $66,000) | La Nación survey (August 2023) |
| Theft-line premium: $228.3B pesos in 9 months | My calculation on SSN statements |
| 3.32M active policies in the line | SSN statistical annex, Q1 2026 |
| Line annulments: 13.6% vs 10.6% market median | My calculation on SSN statements |
| Nu Celular in Brazil since 2022; 2M policies with Chubb | Chubb and Nubank announcements |
| Nu Celular deductible around 25% | Third-party sources; Nubank does not publish it |
| Ualá's 300,000 quotes; Allianz X round | Published figure (Fortuna/Perfil) |
| Failed negotiation for Brubank | Unconfirmed press report |
| 8% waitlist conversion in the smoke test | My model, with the go bar at 5% |
| 15,000 / 60,000 / 120,000 policies at 24 months | My model |
The status column governs how each number reads. My calculations on SSN statements use the same source and method as my collections case: direct-insurance premium and annulments at subline level, July 2025 to March 2026.
If Nubank ports Nu Celular exactly as it is in Brazil, with its deductible around 25%, it enters the market but the wedge I recommend does not exist: it would be one more offer in the pile, competing on brand against a bank channel that already owns distribution. It is the most likely scenario against this analysis, because it requires nothing unusual: Nubank repeating its own playbook is enough.
If the smoke test comes in under 5%, the demand I assume does not hold and the recommendation falls entirely. If Ualá or Mercado Pago launch a phone product without punishment first, the wedge closes. If the anti-fraud operation turns out more expensive than I assume, the low deductible stops being an advantage and becomes the problem.
One more scenario: Nubank not returning to operate in Argentina at all. The Buenos Aires hub is real and confirmed. That it is the first step of a financial comeback is my reading, and the company, when asked, did not confirm it.
analysis/nubank-ramo-robo.py reproduces
every own calculation and charts/nubank-seguros-argentina.py regenerates the
three charts. The download URLs for the SSN datasets sit in each script's
header.If you read this far, something about the problem caught your interest. I like discussing these decisions with people who live them: write me what you would do differently, or let's book a call.
See all projects