NvidiaLuma AI, the video generation startup backed by , joins a growing wave of US tech companies setting up operations in the UK, with major plans to expand to London revealed on Tuesday.
The Palo Alto-headquartered startup plans to employ around 200 people, or about 40% of its workforce, in research, engineering, partnerships and strategic development at its new London location by early 2027.
The expansion comes two weeks after Luma announced a $900 million funding round led by Saudi Public Investment Fund-owned AI company Humane, valuing the company at more than $4 billion. The startup was previously backed by Nvidia.
Luma is building a “world model”. This is a class of AI models that can learn from video, audio, and images in parallel with text, such as OpenAI’s ChatGPT and Google Used by Gemini.
The startup is currently targeting the marketing, advertising, media, and entertainment sectors with its video model, which it sells as part of its content creation suite through an application programming interface (API).
“With this Series C funding and the upcoming buildout of our global computing infrastructure, we now have the capital and capabilities to bring world-class AI to creatives everywhere,” said Amit Jain, CEO and co-founder of Luma AI. “Launching in Europe and the Middle East is the logical next step in putting this power directly into the hands of storytellers, agencies and brands around the world.”
Jain told CNBC that the UK is a starting point for expansion in terms of access to talent.
“London has the best talent when it comes to research, given the universities here and institutions like DeepMind,” he says. “We also see London as the gateway to the European market.”
AI-generated images created by Luma’s Ray3 model (Luma AI)
Luma AI
Luma is the latest in a wave of North American AI labs partnering with the UK and Europe to leverage talent pools and revenue opportunities.
San Francisco-based Anthropic announced plans in November to open offices in Paris and Munich, months after starting hiring in London and Dublin. Canadian AI startup Cohere announced in September that it would open a Paris office as its EMEA headquarters, and OpenAI announced a new office in Munich in February.
Although the world model may not yet be as developed as the LLM, some researchers say it is just as important, if not more important, in the pursuit of artificial general intelligence (AGI).
“These kinds of visual models are currently about a year to a year and a half behind language models,” Jain said.
But he predicted that world models will eventually become the “natural interface” for AI in most everyday applications, noting the amount of time people spend watching video content every day.
including technology giants google, meta and Nvidia are both developing global models for a variety of use cases.
Luma released its latest model, Ray3, in September, and Jain told CNBC that its benchmarks are higher than OpenAI’s Sora and on par with Google’s Veo 3.
