DeepL Buys Mixhalo to Translate Live Events Live

DeepL, the German translation company, has acquired Mixhalo, a startup that streams real-time audio directly to attendees' phones at live events. The deal was announced June 17; terms were not disclosed. The move pushes DeepL out of the text box and into the room — concerts, conferences and sports venues where people need to hear a translated version of what's being said, live, in their own earbuds.

Mixhalo, founded in 2016 by Incubus guitarist Mike Einziger, violinist Ann Marie Simpson-Einziger and CEO Vik Singh, built the plumbing for delivering low-latency audio to a crowd's phones. Its pitch for live events is squarely about the language barrier: instead of a single interpreter over a PA, each attendee gets a translated stream on their own device. The company had raised more than $39 million from backers including Fortress Investment Group, Founders Fund, Defy Partners and Cowboy Ventures.

1Why DeepL wanted it

Mixhalo was already a DeepL customer, so the acquisition folds in a partner that was using DeepL's translation under the hood. DeepL CEO Jarek Kutylowski framed it bluntly: 'Mixhalo will work as a solution and also a marketing use case.' In other words, it's both a product DeepL can sell and a live stage to show off real-time voice translation in front of conference audiences. The company is also opening a San Francisco office to expand its U.S. business.

This fits a clear arc. DeepL spent years as a text-translation tool, then started moving into voice — voice-to-text across 33-plus languages in 2024, and voice-to-voice translation for multilingual meetings in April 2026. Buying Mixhalo gives it the distribution layer for that voice tech in physical spaces, and drops it into a competitive market that already includes rivals like Wordly AI and Palabra.

2Why this matters for you

The interesting shift here isn't the acquisition itself — it's what it says about where translation is going. For most of us, translation has meant pasting text into a window and reading the output. What DeepL is assembling is the opposite: live, spoken, in-the-moment interpreting, whether you're in a conference hall or on a video call. My honest read is that real-time voice translation is still the harder, messier problem than text, and live events are an unforgiving place to prove it — latency and accents punish you instantly. But if it works even half as well as DeepL's text does, the 'I don't speak the speaker's language' barrier starts to quietly disappear, and that reshapes which talks, classes and meetings are actually open to you.