Apple closes reverse acquisition deal with personalized podcast startup Huxe

Apple disclosed in a regulatory filing to the European Commission that it has agreed to hire staff and obtain a non-exclusive license to the intellectual property of Huxe, a personalized-audio startup that announced its shutdown in May. The transaction, first reported by MacRumors, is structured as a reverse acqui-hire — a model that lets large companies absorb key talent and technology without purchasing the corporate entity, thereby sidestepping some antitrust scrutiny.

The reverse-acqui-hire model has gained traction in recent years as a way to build AI capabilities quickly while attracting less regulatory attention. Instead of acquiring the legal entity, its assets, liabilities and capital structure, the buyer extends job offers to core engineers and signs a technology license. In Huxe's case, the filing does not specify which employees received offers, whether those offers were accepted, or what Apple intends to do with the technology.

Huxe was founded by developers who previously worked on the AI-generated podcast features of NotebookLM, recently rebranded as Gemini Notebook. On 21 May the company announced it was ceasing operations: pulling its app from the Apple and Google stores, halting the service and deleting user data. "The team is moving on to new things and we will not continue developing the product," the shutdown notice read, thanking users for "making it feel like we built this together."

On 9 June, shortly after Huxe's announcement, Apple notified the European Commission of the deal. A day before Huxe's closure notice, Spotify unveiled its own AI podcast-creation capabilities, raising the possibility that Apple is targeting similar functionality for its Podcasts app. Industry executives are convinced personalized audio will replace traditional podcasts — and perhaps even parent-child interaction — but it remains unclear how many ordinary listeners actually want it.

The regulatory filing is silent on Apple's concrete plans, the identities of the hires and the scope of the license. Given the team's NotebookLM background, their expertise likely lies in generating dynamic audio content from large language models, but without technical specifications or a product announcement that remains speculation.