Nvidia-backed video generation startup Luma AI just announced plans to build a massive 200-person London office by early 2027, making it the latest US AI company racing to lock down European talent. The expansion comes just two weeks after the startup closed a $900 million funding round that pushed its valuation past $4 billion, positioning it as one of the most valuable AI companies focused on video generation technology.
Nvidia-backed video generation startup Luma AI just pulled the trigger on a major London expansion that signals how aggressively US AI companies are now competing for European talent and market share. The Palo Alto-based company announced Tuesday it will hire around 200 employees across research, engineering, partnerships and strategic development at its new London base by early 2027.
The timing isn't coincidental. This expansion comes exactly two weeks after Luma closed a massive $900 million Series C funding round led by Saudi Public Investment Fund-owned AI company Humain, catapulting the startup's valuation north of $4 billion. That makes Luma one of the most valuable pure-play video AI companies in the world, with enough cash to fuel an international hiring spree.
"With this Series C raise and the upcoming build-out of global compute infrastructure, we have the capital and capacity to bring world-scale AI to creatives everywhere," CEO and co-founder Amit Jain told CNBC. "Launching across Europe and the Middle East is the logical next step in putting this power directly in the hands of storytellers, agencies and brands globally."
Luma's London office will represent roughly 40% of the company's total workforce, an unusually large international footprint for a US-based AI startup. But Jain has specific reasons for targeting the UK first. "London has some of the best people when it comes to research, given the universities here and institutions like DeepMind," he explained to CNBC. "We also consider London to be the entry point to the European market."
The company is building what researchers call "world models" - AI systems that can learn from video, audio, and images alongside text, unlike traditional large language models that primarily process written content. These multimodal systems are seen as potentially crucial for achieving artificial general intelligence, though they lag behind text-based models by about 12-18 months according to Jain.
Luma's core business targets marketing, advertising, media and entertainment companies through both an API and a content creation suite. The startup released its latest model, Ray3, in September, which Jain claims benchmarks higher than 's Sora video generator and performs at similar levels to 's Veo 3.
