Assumptions, model disagreement, and rethink triggers, updated weekly before your PM, risk, or IC discussion. Currently accepting one macro thesis and one AI-economy thesis for July.
4-model average
building category history
loading
42pt spread
Consensus sits at 50% across the four models and is — (loading). Models are diverging at a 42pt spread, so the average conceals live disagreement worth inspecting below. Category confidence is LOW (building category history).
Confidence reflects category-level track record. Stability tracks estimate movement. Models shows whether the four agree.
Loading belief trajectory...
Recent reports from cloud service providers indicate that AI-generated content usage has surged, with estimates suggesting that AI workloads will account for over 30% of cloud traffic by 2025. In 2022, AI-generated content (including text, images, and video) was estimated to generate around 15% of total data stored in major clouds, and this is expected to double by 2029 as generative AI tools become more integrated across industries.
The exponential growth in AI model training data, coupled with the increasing adoption of AI for content creation across various sectors, suggests AI-generated data will rapidly outpace the relatively linear growth of human-generated content storage. By 2029, the sheer scale of AI model datasets, which often require petabytes of information, will likely tip the balance.
Current AI-generated content represents <1% of total cloud storage data as of 2024, while human-generated content (photos, videos, documents) comprises ~90%+ of enterprise and consumer cloud usage. For AI content to exceed human content by 2029 would require AI storage to grow >100x faster than human content over 5 years—a structural shift lacking clear precedent. While AI model checkpoints, training datasets, and synthetic media are growing exponentially (estimated 40-50% YoY), human content generation (messaging, video streaming, IoT sensor data) is also growing at 25-30% YoY. The gap is too large and human content generation too entrenched for reversal in 5 years absent a discontinuous technology shift.
AI training datasets and synthetic media generation are growing at 35-45% CAGR per Epoch AI and Scale reports, yet inference traffic (ChatGPT, Midjourney) already generates >70% of current AI storage demand; total AI-related storage on AWS, Azure, and Google Cloud was ~4% of capacity in 2023 and would need to reach >50% by 2029 to overtake human-generated content under current 25-30% overall cloud growth rates.