TwelveLabs secures $100M for video models as AWS partnership deepens
TwelveLabs has successfully secured $100 million in Series B funding to scale its video intelligence foundation models and expand its international operations. The company has also strengthened its multiyear partnership with AWS to optimize its inference workloads using custom Trainium chips.
Key Takeaways
- Series B funding reached $100 million with participation from Amazon, Index Ventures, and Red Bull Ventures.
- Pegasus 1.5 model specialized in converting video into structured data, including temporal segments and semantic context.
- AWS remains the preferred cloud provider with a commitment to launch new TwelveLabs models first on its infrastructure.
- The company launched Rodeo, its first application-layer product, to move beyond core model development.
- International scaling includes expanding San Francisco and Seoul operations while establishing a New York and London presence.
Why It Matters
TwelveLabs is shifting the generative AI focus from text-heavy LLMs to 'video cognition' systems capable of reasoning about moving images. For the streaming industry, this accelerates the transition of vast archives into searchable, structured data without manual tagging. The strategic alignment with AWS and its Trainium chips signal a move toward reducing the significant compute costs associated with high-resolution video inference. Operators should watch for the integration of Marengo and Pegasus into Amazon Bedrock as a benchmark for how quickly media enterprises can deploy automated scene-level metadata at scale.
Additional Context
The surge in specialized video AI investment follows broader industry efforts to manage escalating storage and processing costs. Per TechCrunch in June 2024, TwelveLabs previously raised a $50 million Series A, which valued the company at roughly $300 million at the time. This latest injection reflects a maturation of the market where foundational computer vision is moving from simple object detection to complex temporal reasoning. Competitors like OpenAI with Sora and Google with Gemini 1.5 Pro have also pushed localized video understanding, yet TwelveLabs remains focused on the API-first 'visual cortex' niche for enterprise developers. Related developments in the silicon space underscore the importance of the AWS Trainium partnership mentioned in the funding announcement. According to internal AWS benchmarks reported in late 2024, Trainium instances can offer up to 50% improvement in price-performance for deep learning models compared to standard GPU instances. This efficiency is critical for TwelveLabs as it scales Pegasus 1.5, which processes long-form video content that typically overwhelms standard transformer architectures. Additionally, NAVER's role as a co-lead investor highlights a growing trend of sovereign AI development in Asia, where localized video intelligence is becoming a priority for regional media conglomerates seeking independence from Western-centric model biases.
Read full article at sportsvideo.org
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source