Back
Epoch AI: Papers & Reports
webCredibility Rating
4/5
High(4)High quality. Established institution or organization with editorial oversight and accountability.
Rating inherited from publication venue: Epoch AI
Epoch AI is a frequently cited empirical research organization in AI safety discussions; their compute and scaling trend data informs both near-term capability forecasts and longer-term existential risk assessments.
Metadata
Importance: 72/100blog posthomepage
Summary
Epoch AI is a research organization focused on investigating trends in AI development, including compute scaling, dataset growth, algorithmic progress, and AI forecasting. Their blog and reports provide empirical analysis to inform predictions about AI timelines and capabilities trajectories. It serves as a key reference for quantitative research on the pace of AI advancement.
Key Points
- •Tracks historical and projected trends in AI training compute, model parameters, and dataset sizes
- •Produces forecasts and empirical analyses relevant to AI timelines and transformative AI risk assessments
- •Publishes research used widely in AI safety and governance communities to ground capability predictions
- •Covers algorithmic efficiency improvements alongside hardware scaling trends
- •Serves as a primary data source for arguments about compute governance and AI policy interventions
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| Epoch AI | Organization | 51.0 |
Cached Content Preview
HTTP 200Fetched Mar 20, 20267 KB
# Publications
### Featured
[\\
\\
Report\\
\\
Aug. 20, 2024\\
\\
**Can AI scaling continue through 2030?**](https://epoch.ai/blog/can-ai-scaling-continue-through-2030)
[\\
\\
Report\\
\\
Sep. 16, 2025\\
\\
**What will AI look like in 2030?**](https://epoch.ai/blog/what-will-ai-look-like-in-2030) [\\
\\
Report\\
\\
Feb. 10, 2026\\
\\
**Where Autonomy Works: Evaluating Robot Capabilities in 2026**](https://epoch.ai/blog/where-autonomy-works-evaluating-robot-capabilities-in-2026)
## Filter
Type
Report (42)
Update (24)
Paper (22)
Viewpoint (2)
Topic
Compute (32)
Trends (28)
Performance & Benchmarks (16)
Hardware (15)
Tools & Resources (13)
Organizational Updates (9)
Algorithmic Progress (8)
Scaling Laws (8)
Math (7)
Training Data (7)
Economics of AI (5)
Model Development (2)
Data Centers (2)
Energy (2)
Inference (2)
Accessibility (1)\+ More
Sort
- Newest
- Oldest
90 results
[\\
\\
Report\\
\\
Feb. 20, 2026\\
\\
**Expanding our analysis of biological AI models**\\
\\
We release a database of over 1,100 biological AI models across nine categories. We analyze their safeguards, accessibility, training data sources, and the foundation models they build on.\\
\\
By David Atanasov, Niccolò Zanichelli, and Jean-Stanislas Denain](https://epoch.ai/blog/expanding-our-analysis-of-biological-ai-models) [\\
\\
Report\\
\\
Feb. 13, 2026\\
\\
**What do “economic value” benchmarks tell us?**\\
\\
These benchmarks track a wide range of digital work. Progress will correlate with economic utility, but tasks are too self-contained to indicate full automation.\\
\\
By Florian Brand and Greg Burnham](https://epoch.ai/blog/what-do-economic-value-benchmarks-tell-us) [\\
\\
Report\\
\\
Feb. 10, 2026\\
\\
**Where Autonomy Works: Evaluating Robot Capabilities in 2026**\\
\\
We assess the current state of autonomous robotics by evaluating robot performance on concrete tasks across industrial, household, and navigat
... (truncated, 7 KB total)Resource ID:
8a3ab5bfcf7a96f8 | Stable ID: ZDc3NzNlZD