The news: Luma AI has raised $900 million in new funding led by Saudi-backed AI company Humain, pushing Luma’s valuation above $4 billion and marking one of the largest investments to date in AI-generated video. Additional investors include AMD’s venture arm, Andreessen Horowitz, Amplify Partners, and Matrix Partners.

How Luma works: The AI service offers several features to create multimedia assets for a customer base that now spans major ad agencies, entertainment studios, tech companies, and professional creators—all leaning into AI video as a scalable production engine.

• Its multimodal “world models,” trained on text, video, audio, and imagery, are built to simulate physical environments with a realism that “traditional” LLMs can’t match, CEO Amit Jain argues.
• The company’s Ray3 model—released in September—can generate video, images, and audio from prompts. Jain said Ray3 outperforms OpenAI’s Sora 2 and benchmarks comparably to Google’s Veo 3, framing Luma as one of the top tier generative video models. Luma also plans deeper deployment across the Middle East to address cultural underrepresentation in AI-generated visuals.
• Luma has faced copyright scrutiny surrounding its Dream Machine tool, but Jain said the company uses internal detection models to block IP-infringing outputs and is continuously improving guardrails.

Why it matters: Video advertising is shifting from manual production to AI-assisted pipelines, and the data shows marketers are embracing these tools at speed.

• 60% of marketers say content creation is their top AI use case in video advertising, per Teads and MMA Global.
• 54.9% say AI is improving workflow automation for creative teams and 48.2% expect AI to improve video performance optimization, per an EMARKETER and Triplelift survey.
• AI video tools are rapidly becoming standard; 36% of marketers already use smart video/audio editing tools and 30% use video or animation generators like Luma or Sora, per HubSpot. Twenty percent of marketers globally now use genAI for full video generation, and 33% expect to adopt such tools within 6–12 months, per SAS.
• AI video unlocks personalization at scale; 31.7% expect AI to improve video personalization, per EMARKETER and Triplelift, and 33% already use AI-driven personalization in their video ad workflows, per HubSpot.
• Predictive video testing is emerging; 19% of marketers use predictive AI to pretest video concepts before production dollars are spent, per Teads and MMA Global.

Key takeaway for marketers: Luma’s mega fundraising round is a signal that AI video is becoming the backbone of modern creative production.

• As marketers adopt AI editing, narration, generation, and personalization tools at scale, video teams will need to rethink workflows around rapid iteration and multi-format output.
• The brands that embrace AI as core infrastructure—not an experiment—will be the ones producing faster, testing smarter, and tailoring video ads to every surface from social feeds to CTV.