{"id":112590,"date":"2025-11-20T08:35:11","date_gmt":"2025-11-20T16:35:11","guid":{"rendered":"https:\/\/www.backblaze.com\/blog\/?p=112590"},"modified":"2025-11-20T08:35:14","modified_gmt":"2025-11-20T16:35:14","slug":"building-multimodal-ai-data-infrastructure-with-pixeltable","status":"publish","type":"post","link":"https:\/\/www.backblaze.com\/blog\/building-multimodal-ai-data-infrastructure-with-pixeltable\/","title":{"rendered":"Building Multimodal AI Data Infrastructure with Pixeltable"},"content":{"rendered":"\n
\"A<\/figure>\n\n\n\n
<\/div>\n\n\n\n

We’re approaching a fascinating inflection point in AI development. Research from Epoch AI<\/a> indicates that high-quality text data will be fully exhausted by 2026 to 2028. As recently as January, OpenAI co-founder Ilya Sutskever said<\/a> at a conference that all the useful data online had already been used to train models. Over 35% of top websites now block AI scrapers<\/a>. OpenAI is cutting deals with publishers like The Financial Times because freely available training data is running out.<\/p>\n\n\n\n

So what comes next? Multimodal data: video, images, audio, sensor readings. Data that captures how the physical world actually operates, not just how we describe it in text.<\/p>\n\n\n\n

Nvidia CEO Jensen Huang highlighted this shift when discussing Tesla’s AI advantage<\/a>. He noted that the company has a “phenomenal position” because Tesla is collecting massive amounts of real-world data through its AI-enabled factories and autonomous vehicles.”<\/p>\n\n\n\n

This real-world data, what some call “world data,” is multimodal at its core. It includes video from cameras capturing spatial relationships and motion, sensor telemetry recording physical interactions, images showing object states, and audio capturing environmental context. Video is particularly valuable because it captures temporal dynamics, depth perception, and how objects interact over time, insights that static text or images alone cannot provide.<\/p>\n\n\n\n

Here’s the insight most organizations miss: you’re already generating this data.<\/p>\n\n\n\n

Your organization is already producing multimodal data<\/h2>\n\n\n\n

Every single day, your organization produces massive amounts of multimodal data, including:<\/p>\n\n\n\n