In 2014 Amazon failed miserably with its Fire Phone. AI has given him the perfect excuse to go back to his old ways

In 2014 Jeff Bezos was still the absolute boss of Amazon. And it occurred to him that Amazon should have its own smartphone, so the company ended up launching the fire phone. That device boasted a 3D screen that overheated the battery and a catalog without the apps that everyone used, but it also it was very expensive for the time ($649). It was canceled 14 months later and became one of the most notorious failures of the recent history of technology. Now Amazon wants to try again with a project codenamed Transformer.

Same idea, different time. In reality, the underlying concept has not changed much since 2014. Here Amazon’s goal is to have its own device that has Alexa as an integral part of the experience, but that is also a gateway for purchases on Amazon, connects with Prime Video or its food delivery services. Or what is the same: it is the perfect mobile for those who make their lives revolve around the Amazon ecosystem. And be careful, because what failed in the past may have a chance now.

AI as an argument. The Transformer project wants to integrate AI functions to ensure that with it we do what theoretically will end up arriving sooner or later: that we simply ask for what we need so that the phone takes care of everything. Alexa would be a central component of that experience here, and the advantage that Amazon has is that it has the infrastructure and ecosystem that should allow doing something like this. AI agents begin to demonstrate their potential —we are seeing it with OpenClaw—and if Amazon can make that shopping experience easier, you may have a winning idea here.

A team with tables. The project is led by J. Allard from an internal unit called ZeroOne in which the objective is to create “revolutionary” devices. Allard was one of the fathers of the original Xbox at Microsoft and also of the Zune, the music player that tried to compete with the iPod and failed. But above him is Panos Panay, who headed the Surface division at Microsoft before joining Amazon. They are certainly two veterans with a lot of experience in the hardware field and know first-hand what it is like to compete with the market leaders from disadvantaged positions. Now they have a unique opportunity to shine, but the challenge is colossal.

The ‘dumbphone’ as a back door. One of the most curious twists of the project is that Amazon is not only exploring a conventional smartphone, but also a “dumbphone.” That is, a simple and minimalist mobile with limited functions inspired by the Light Phone and its interface “dumbed down“. The argument is striking: here it is not about trying to unseat the iPhone as the main device, and instead Amazon could position it as “a second phone.” Mobile phones with limited functions represented 15% of global mobile sales in 2025 according to Counterpoint Researchand although it is a small market, Amazon may have an interesting entry point there.

But. The context, however, complicates everything. Amazon comes to this project at a particularly difficult time for mobile hardware. The number of smartphones distributed (“sold”) will probably fall more than ever in 2026, and in fact the preliminary descent of that figure is 13% due especially to the RAM crisis.

AI hardware is the holy grail. To this we must add the fact that for now no AI hardware device has succeeded, and those who have tried have been an absolute failure. The Humane AI Pin and the Rabbit R1 have been painful lessons for an industry that now certainly wants to try again—let them tell it to OpenAI and its alliance with Jony Ive—. The project is underway, but it is not ruled out that Amazon ends up canceling it if the strategy changes or the numbers do not add up.

A golden opportunity. But what is certain is that Amazon has indisputable advantages in getting it right. For example, a service ecosystem with hundreds of millions of active users who already shop, watch content, and use Alexa, even if it’s just to set timers. The multi-million dollar investment in Anthropic and a relationship increasingly narrower with OpenAI they can also be key in this project. The question, of course, is whether an AI phone can really convince us to switch phones. And Amazon wants to have the answer to that question.

In Xataka | NVIDIA is doing better than ever. And there is also more competition ready to eat it than ever.