DeepL moves into Silicon Valley, takes on Mixhalo staff and technology to scale up Voice AI
DeepL, the German Language AI company, has acquired the team and technology behind Mixhalo, the San Francisco-based platform that delivers real-time, ultra-low latency audio. The deal is designed to extend the capabilities of DeepL Voice, the company’s real-time AI voice translation product, into larger and more complex settings where speed, clarity and reliability are critical — including major events, conferences, customer support operations and enterprise workflows.
New capabilities for DeepL Voice
DeepL Voice already holds a leading position in real-time voice translation for spoken conversations, outperforming Microsoft Teams, Zoom and Google Meet on accuracy, fluency and reliability. According to an independent assessment by Slator published in 2026, DeepL Voice achieved a quality score of 96.4 out of 100 and a failure rate of just 4%, compared with a market average of 17%. With Mixhalo’s technology now part of the company, DeepL is integrating ultra-low-latency audio infrastructure directly into DeepL Voice to support live, large-scale environments. The integration will allow translated speech and captions to reach audiences clearly and instantly — from small meetings to crowds of tens of thousands — while preserving the pace and natural fluency of live speech.
“DeepL Voice is already changing how people and businesses work across languages every day, and the Mixhalo team and technology let us bring this to even larger, more complex settings,” said Jarek Kutylowski, founder and chief executive of DeepL. “The team has solved one of the hardest problems in live audio, which is delivering high-fidelity sound to thousands of people at once with basically zero latency. Together, we’re building the real-time Language AI layer for communication, so people can understand each other naturally wherever they are interacting, whether that’s in team meetings, customer calls or even major international events.”
How Mixhalo’s technology works
Mixhalo was founded in 2016 by Incubus guitarist Mike Einziger, violinist Ann Marie Simpson-Einziger and technologist Vik Singh, with the goal of reimagining how people experience live sound. Its platform streams high-fidelity, synchronised audio directly to attendees’ smartphones and headphones with imperceptible delay — using proprietary technology that behaves more like a radio network than traditional Wi-Fi, ensuring consistent quality regardless of the number of users. The company has deployed its system across major sports leagues, including the MLB and NASCAR; entertainment venues such as Red Rocks Park and Amphitheatre and the Sacramento Kings’ arena; large conferences including CES, Mobile World Congress, Salesforce’s flagship events, the Databricks AI Summit and Microsoft AI Tour Paris; and concerts and residencies for artists including Metallica, Aerosmith, Sting and Charlie Puth. Mixhalo had raised approximately $39 million in funding before the acquisition.
“We launched Mixhalo to deliver the best quality audio to live audiences at any size, all within 20 milliseconds,” said Vik Singh, co-founder of Mixhalo. “As we expanded into real-time translation, the hardest part was finding technology that could keep up with live speech without compromising quality. After testing technology from nearly all of the leading providers, DeepL was the only one that could deliver accurate translations at the speed we needed to achieve real fluency. We’ve already seen what this combination can do in real-world settings, and by joining DeepL, we can now bring this experience to even more audiences and customers globally.”
Integration already under way
DeepL and Mixhalo have already brought their combined technology to market. The DeepL Voice API is powering real-time translation across Mixhalo’s live-audio platform, and the teams are piloting customer support applications through integrations such as Amazon Connect. The ultra-low-latency audio infrastructure is being woven into DeepL Voice to support environments where even a fraction of a second of delay can break the flow of communication. Translated speech and captions must reach audiences instantly, preserving the pace of live speech — a challenge that Mixhalo’s radio-network-like architecture is designed to solve.
“Voice AI is ultimately a fight for latency, quality, and real-world reliability,” said Sebastian Enderlein, chief technology officer of DeepL. “The Mixhalo team has deep experience bringing APIs and audio infrastructure into live environments where there is no room for delay or failure. That expertise is incredibly valuable as we continue to scale DeepL Voice, and their presence in San Francisco brings us even closer to the customers, partners, and developer ecosystem shaping the next generation of AI products.”
Expanding US presence
The acquisition also marks the opening of DeepL’s first office in San Francisco, strengthening its footprint in the United States, its fastest-growing market. DeepL’s US customer base already includes NVIDIA, Cisco and Nasdaq, and nearly half of the Fortune 500 are users of the company’s language platform. DeepL’s broader Language AI suite — which includes Translator, Voice, Write, Documents and API services — has seen rapid adoption. The company’s machine translation market share grew by 44% in a single year, and a 2024 study by Forrester found that DeepL delivered a 345% return on investment, reduced translation time by 90% and cut workload by 50%. According to a 2024 industry survey by the Association of Language Companies, 82% of language service companies use DeepL, compared with 46% for Google, 32% for Microsoft and 17% for Amazon AWS.
DeepL has raised a total of $415 million over five funding rounds, including a $300 million Series C in May 2024 that valued the company at $2 billion post-money. Investors include Index Ventures, IVP, Atomico, Benchmark and ICONIQ Growth. The company has been exploring a potential initial public offering, with reports suggesting a possible timeline of late 2025 or 2026 and a valuation target of up to $5 billion. Internally, DeepL has undergone a restructuring that included approximately 250 layoffs as it focuses on AI-native teams and accelerates product development.
Audiences will be able to experience the enhanced DeepL Voice technology at GITEX Europe 2026 at the end of June, where DeepL will serve as the official translation partner, powering live German-to-English captions on the main stage. The company will also showcase its next-generation Voice AI capabilities live at upcoming US events including the Databricks Data + AI Summit, the Esri User Conference and Salesforce Dreamforce.
