Is Google Meet Live Translation Ready for Prime Time

In late 2025, Google disclosed the technical framework behind its real-time speech-to-speech translation system currently used in Google Meet and Pixel 10 devices. The end-to-end architecture is supposed to enable nearly instantaneous communication by maintaining the original speaker’s voice and tone with only a minimal delay. Well, not quite minimal.

While the January 27, 2026, beta rollout to the general business public revealed a mostly linguistically competent tool, latency was a noticeable ~10 seconds in Slator’s informal test using Google Meet. This delay makes it so that by the time people react to what someone has said in another language using their own cloned voice, the original speaker is probably three thoughts ahead.

At launch, bidirectional support is available between English and five European languages (Spanish, French, German, Portuguese, and Italian), but sessions are limited to one active language pair at a time. Additionally, only one-on-one, speak-wait-listen-wait conversations are possible. 

Multispeaker calls are not. And current users on meeting room hardware can listen to translations, but their own speech cannot yet be translated. There are also usage limits, and to encourage adoption, Google Workspace customers can have access to higher usage limits for at least 60 days. After that, standard per-user limits will apply.

Readers shared their thoughts on Google Meet live translation, and over half of them (53.9%) think it is not ready for prime time yet. Approximately one in five (19.2%) believe it works well, while those on Microsoft (15.4%) cannot test it. For the rest (11.5%), it works, but it is slow.

ElevenLabs’ Rich Valuation

In early February 2026, ElevenLabs reported a USD 500m Series D round of funding. Investors like Sequoia Capital and Andreessen Horowitz have backed the firm as it transitions from a creator-focused AI dubbing tool to an enterprise-grade platform.

The company’s valuation more than tripled from approximately USD 3bn following a Series C round in early 2025 to USD 11bn following the most recent round. Currently, the firm says it is generating over USD 330m in ARR, according to multiple social media posts.

ElevenLabs is transitioning into becoming a piece of foundational AI audio infrastructure: the ElevenCreative suite and its ability to handle over 70 languages with brand-consistent audio is touted on the company’s website as “a platform for creators and brands to generate, edit, and localize high-fidelity audio.”

High-profile projects, such as dubbing for a January 2025 interview with Ukrainian President Volodymyr Zelenskyy that garnered over seven million views on YouTube, have helped highlight the startup’s capabilities and diversify its user base. But is investor confidence enough to make sense of the USD 11bn valuation? That is the question Slator readers answered after the news hit the virtual presses.

Close to a third (29.8%) think the company is overvalued. A little over a quarter (27.0%) believe it probably does not make sense. Close to one in five people (21.6%) would make sense of the valuation if it is well executed. The rest are either unsure and think it is too early to say (13.5%) or convinced the valuation definitely makes sense (8.1%).

Fifth Wheel is a Translator App

David Duda and Hong Liang have been a couple since late 2019. He speaks English, she speaks Mandarin, and neither has learned more than 200 words in the other’s language. Enter Microsoft Translator, which knows all their secrets and helps them communicate on their smartphones. That is the backdrop of a New York Times story about speech-to-text translation and the shift in AI use cases to more personal applications. 

The article (which incidentally quoted Slator’s sizing of the language industry as a USD 31bn market) also mentioned how AI translation is still struggling with context, idioms, and incomplete or confusing messages, forcing users to monitor the output. 

Other limitations include ambient noise, connectivity issues, nonsensical long translations, or lost data as the speech-to-text engine fails to keep up with natural pacing. 

But users like Mr. Duda and Ms. Liang put up with these issues because these smartphone translation apps are easy to use and free in exchange for data for AI: in this case, the couple opted into a human-in-the-loop agreement, allowing Microsoft to review their private audio clips to improve the app.

Over a third of readers (33.4%) reacted favorably to the married couple using Microsoft Translator to communicate. More than a quarter (28.9%) think it would not work for them, and a similar size group (24.4%) finds it very difficult. The rest either applaud Microsoft (4.4%) or agree at once with all these views (8.9%).

Deadline Looming for High-Visibility Vacancy

The European Commission (EC) has set March 19, 2026, as the last date to submit applications for the position of Director of the Translation Center for the Bodies of the European Union (EU). 

Based in Luxembourg, the incoming Director will lead an organization of approximately 200 staff, providing vital multilingual services to over 70 EU agencies across all 24 official languages.

Beyond managing budgets and human resources, the successful candidate must represent the agency on the international stage and harmonize best practices across the broader EU translation ecosystem. A key responsibility of the Director will be to evolve the Center’s business model to navigate shifting trends within a language industry increasingly defined by language AI.

Most readers (54.1%) think the most important qualification for the next EU Translation Center boss is change management experience. Technical (AI) experience is the second most voted qualification (18.9%), with networking skills (10.8%) and other qualifications (16.2%) following.