How Much Power Will Frontier AI Training Demand in 2030? | Epoch AI
The power required to train the largest frontier models is growing by more than 2x per year, and is on trend to reaching multiple gigawatts by 2030.| Epoch AI
The power required to train the largest frontier models is growing by more than 2x per year, and is on trend to reaching multiple gigawatts by 2030.| Epoch AI
Epoch AI’s Biology AI Models Dataset is a collection of machine learning models with applications in biology for research about trends in the intersection between AI and biology.| Epoch AI
It’s good at involved computations, improving at proofs, and useful for literature search. It still favors low-level grinds and leans on background knowledge.| Epoch AI
Our database of benchmark results, featuring the performance of leading AI models on challenging tasks. It includes results from benchmarks evaluated internally by Epoch AI as well as data collected from external sources. The dashboard tracks AI progress over time, and correlates benchmark scores with key factors like compute or model accessibility.| Epoch AI
Leading AI supercomputers are becoming ever more energy-intensive, using more power-hungry chips in greater numbers. In January 2019, Summit at Oak Ridge National Lab had the highest power capacity of any AI supercomputer at 13 MW. Today, xAI’s Colossus supercomputer uses 280 MW, over 20x as much. Colossus relies on mobile generators because the local grid has insufficient power capacity for so much hardware. In the future, we may see frontier models trained across geographically distribute...| Epoch AI
The computational performance of the leading AI supercomputers has grown by 2.5x annually since 2019. This has enabled vastly more powerful training runs: if 2020’s GPT-3 were trained on xAI’s Colossus, the original two week training run could be completed in under 2 hours. This growth was enabled by two factors: the number of chips deployed per cluster has increased by 1.6x per year, and performance per chip has also improved by 1.6x annually.| Epoch AI
The private sector’s share of global AI computing capacity has grown from 40% in 2019 to 80% in 2025. Though many leading early supercomputers such as Summit were run by government and academic labs, the total installed computing power of public-sector supercomputers has only increased at 1.8x per year, rapidly outpaced by private-sector supercomputers, whose total computing power has grown at 2.7x per year. The rising economic importance of AI has spurred the private sector to build more a...| Epoch AI
As of May 2025, the United States contains about three-quarters of global AI supercomputer performance, with China in second place with 15%. Meanwhile, traditional high-performance computing leaders like Germany, Japan, and France now play marginal roles in the AI supercomputing landscape. This shift largely reflects the increased dominance of major technology companies, which are predominantly based in the United States.| Epoch AI
AI supercomputers have become increasingly expensive. Since 2019, the cost of the computing hardware for leading supercomputers has increased at a rate of 1.9x per year. In June 2022, the most expensive cluster was Oak Ridge National Laboratory Frontier, with a reported cost of $200M. Three years later, as of June 2025, the most expensive supercomputer is xAI’s Colossus, estimated to use over $7B of hardware.| Epoch AI
Epoch AI is a research institute investigating key trends and questions that will shape the trajectory and governance of Artificial Intelligence.| Epoch AI
We are releasing a public registry of optimized Docker images for SWE-bench. This allows us to run SWE-bench Verified in 62 minutes on a single GitHub actions VM.| Epoch AI
How much does it cost to train AI models? Looking at 124 ML systems from between 2009 and 2022, we find the cost has grown by approximately 0.5OOM/year.| Epoch AI
This Gradient Updates issue explores DeepSeek-R1’s architecture, training cost, and pricing, showing how it rivals OpenAI’s o1 at 30x lower cost.| Epoch AI
Our ML trends dashboard showcases key statistics on the trajectory of artificial intelligence, including compute, costs, data, hardware and more.| Epoch AI
Our new article explores whether deployment of advanced AI systems could lead to growth rates ten times higher than those of today’s frontier economies.| Epoch AI
Returns to R&D are key in growth dynamics and AI development. Our paper introduces new empirical techniques to estimate this vital parameter.| Epoch AI
As the capabilities of AI models have expanded, and as the recent paradigm of test-time compute scaling has taken off, the demand for AI inference has grown enormously. Inference revenue at major AI companies such as OpenAI and Anthropic has been growing at a rate of 3x per year or more, even as their models continue to become smaller and cheaper compared to 2023.| Epoch AI | Blog
SWE-bench Verified| Epoch AI | Blog
Our director explains Epoch AI’s mission and how we decide our priorities. In short, we work on projects to understand the trajectory of AI, share this knowledge publicly, and inform important decisions about AI.| Epoch AI
We project how many notable AI models will exceed training compute thresholds. Model counts rapidly grow from 10 above 1e26 FLOP by 2026, to over 200 by 2030.| Epoch AI
This Gradient Updates issue explains Moravec’s paradox and offers a speculative picture of how hard various economic tasks are to automate based on the paradox.| Epoch AI
Available evidence suggests that rapid growth in reasoning training can continue for a year or so.| Epoch AI
This Gradient Updates issue explores how AGI could disrupt labor markets, potentially driving wages below subsistence levels, and challenge historical economic trends.| Epoch AI
Since 2010, the length of training runs has increased by 1.2x per year among notable models, excluding those that are fine-tuned from base models. A continuation of this trend would ease hardware constraints, by increasing training compute without requiring more chips or power. However, longer training times face a tradeoff. For very long runs, waiting for future improvements to algorithms and hardware might outweigh the benefits of extended training.| Epoch AI
Our database of over 500 supercomputers (also known as computing clusters) tracks large hardware facilities for AI training and inference and maps them across the globe.| Epoch AI
Our public datasets catalog over 2400 machine learning models. Explore data and graphs showing the growth and trajectory of AI from 1950 to today.| Epoch AI
AI supercomputers double in performance every 9 months, cost billions of dollars, and require as much power as mid-sized cities. Companies now own 80% of all AI supercomputers, while governments’ share has declined.| Epoch AI
This Gradient Updates issue explores how much energy ChatGPT uses per query, revealing it’s 10x less than common estimates.| Epoch AI
Progress in pretrained language model performance outpaces expectations, occurring at a pace equivalent to doubling computational power every 5 to 14 months.| Epoch AI
We introduce a compute-centric model of AI automation and its economic effects, illustrating key dynamics of AI development. The model suggests large AI investments and subsequent economic growth.| Epoch AI
AI’s biggest impact will come from broad labor automation—not R&D—driving economic growth through scale, not scientific breakthroughs.| Epoch AI
AI’s “train-once-deploy-many” advantage yields increasing returns: doubling compute more than doubles output by increasing models’ inference efficiency and enabling more deployed inference instances.| Epoch AI
While scaling compute is key to improving LLMs, post-training enhancements can offer gains equivalent to 5-20x more compute at less than 1% of the cost.| Epoch AI
We’ve compiled a comprehensive dataset of the training compute of AI models, providing key insights into AI development.| Epoch AI
FrontierMath is a benchmark of hundreds of unpublished and extremely challenging math problems to help us to understand the limits of artificial intelligence.| Epoch AI
We’ve expanded our Biology AI Dataset, now covering 360+ models. Our analysis reveals rapid scaling from 2017-2021, followed by a notable slowdown in biological model development.| Epoch AI
We clarify that OpenAI commissioned Epoch AI to produce 300 math questions for the FrontierMath benchmark. They own these and have access to the statements and solutions, except for a 50-question holdout set.| Epoch AI
This Gradient Updates issue goes over the major changes that went into DeepSeek’s most recent model.| Epoch AI
We are hosting a competition to establish rigorous human performance baselines for FrontierMath. With a prize pool of $10,000, your participation will contribute directly to measuring AI progress in solving challenging mathematical problems.| Epoch AI
We are launching the AI Benchmarking Hub: a platform presenting our evaluations of leading models on challenging benchmarks, with analysis of trends in AI capabilities.| Epoch AI
Data movement bottlenecks limit LLM scaling beyond 2e28 FLOP, with a “latency wall” at 2e31 FLOP. We may hit these in ~3 years. Aggressive batch size scaling could potentially overcome these limits.| Epoch AI
We investigate four constraints to scaling AI training: power, chip manufacturing, data, and latency. We predict 2e29 FLOP runs will be feasible by 2030.| Epoch AI
If trends continue, language models will fully utilize the stock of human-generated public text between 2026 and 2032.| Epoch AI
We characterize techniques that induce a tradeoff between spending resources on training and inference, outlining their implications for AI governance.| Epoch AI
