Understanding Amazon’s Alexa Failure in the New Age of AI

In September 2023, a new version of the artificial intelligence (AI) behind Alexa, one of Amazon’s most popular products, was presented. At the event, it was demonstrated, for example, that the assistant will respond to commands with a more natural and accurate voice than before.

The problem is that these updates never reached the half-billion Alexa devices sold worldwide. According to former company employees interviewed by Fortune, the demo was done with technology that wasn’t ready for implementation. That means Amazon is starting to fall behind in the new age of AI compared to competitors like OpenAI, Google, Microsoft, Meta, and even Apple.

One reason for the failure is that Alexa’s new AI was trained with just a few bits of data. What’s more, the team behind the voice assistant doesn’t have access to the vast amounts of Nvidia’s most advanced specialized chips used to train and run AI models. These devices are highly sought after for AI processing, becoming the true “brains” of data centers and supercomputers responsible for executing the most powerful AI on the planet.