- Avnish Prakash said his Dubai-based startup provides AI-powered content translation services in over 150 languages.
AMSTERDAM: The CEO of Dubai-based artificial intelligence translation platform CAMB.AI said on Friday that his company is working to make language barriers a thing of the past.
Speaking at the IBC 2025 technology show in Amsterdam, Avnish Prakash said he envisions “a world free of language (barriers).”
Prakash continued: “We work with languages – with their diversity – and we want to completely remove the language barrier from the equation.”
CAMB.AI is a cutting-edge AI-powered speech synthesis and translation company that Prakash says enables translation and localization of content — be it video, audio or text — into over 150 languages.
Prakash said CAMB.AI is already working with major clients including Comcast, MLS, Google and IMAX to “localize everything from content to communication, sports to entertainment, healthcare and even spirituality.”
He added: “When you're on the right side of the language barrier, when you can consume the best content in the world that's available in a certain language, life feels great. It feels very fair. But when you're on the other side of the barrier, life isn't so fair, right? Language shouldn't divide.”
According to Prakash, English is spoken by less than 17 percent of people worldwide, yet most content is created with English speakers in mind.
“Times are changing,” he continued. “We now have some of the biggest OTT (over-the-top) hits that are not English-language originals. There are more foreign-language films at the Oscars. Indian and Saudi sports leagues are thriving.”
Founded in 2023, CAMB.AI has become a major player in the real-time localization and multilingual translation space. The startup recently teamed up with Vox Cinemas to translate a film originally shot in Arabic into multiple languages, including Chinese.
The system also worked on live sporting events, including NASCAR races and Major League Soccer games, translating live commentary into multiple languages, which Prakash called “a first in human history.”
The company has released two language models: MARS (a text-to-speech model) and BOLI (a neural machine translation model).
Prakash explained that instead of providing “literal translation,” the technology processes “content that has meaning and significance,” providing translations in seconds.
“We don't have to wait for something to be ready,” he said, arguing that the system works “as long as we keep the soul of the original content – and that's what sets us apart.”
“We’re basically mimicking the flow, the swirl, the emotion, or the prosody of the original content,” he continued. “If there’s a stutter, we’re going to stutter. We’re doing the processing from scratch (training) and staying true to what’s happening right there in the moment.”
