Epoch AI
@epochai.bsky.social
We are a research institute investigating the trajectory of AI for the benefit of society.
epoch.ai
epoch.ai
How does math research change when the cost of trying your first dumb idea goes to zero?
University of Toronto mathematician Daniel Litt joins hosts Greg Burnham & Anson Ho to discuss what today’s models can and can’t do in math, and how far they are from doing high-quality research.
Video below!
University of Toronto mathematician Daniel Litt joins hosts Greg Burnham & Anson Ho to discuss what today’s models can and can’t do in math, and how far they are from doing high-quality research.
Video below!
January 29, 2026 at 8:11 PM
How does math research change when the cost of trying your first dumb idea goes to zero?
University of Toronto mathematician Daniel Litt joins hosts Greg Burnham & Anson Ho to discuss what today’s models can and can’t do in math, and how far they are from doing high-quality research.
Video below!
University of Toronto mathematician Daniel Litt joins hosts Greg Burnham & Anson Ho to discuss what today’s models can and can’t do in math, and how far they are from doing high-quality research.
Video below!
Was serving GPT-5 profitable?
According to jsevillamol.bsky.social, @exponentialview.skystack.xyz’s Hannah Petrovic, and Anson Ho, it depends. Gross margins were around 45%, making inference look profitable.
But after accounting for the cost of operations, OpenAI likely incurred a loss.👇
According to jsevillamol.bsky.social, @exponentialview.skystack.xyz’s Hannah Petrovic, and Anson Ho, it depends. Gross margins were around 45%, making inference look profitable.
But after accounting for the cost of operations, OpenAI likely incurred a loss.👇
January 28, 2026 at 11:20 PM
Was serving GPT-5 profitable?
According to jsevillamol.bsky.social, @exponentialview.skystack.xyz’s Hannah Petrovic, and Anson Ho, it depends. Gross margins were around 45%, making inference look profitable.
But after accounting for the cost of operations, OpenAI likely incurred a loss.👇
According to jsevillamol.bsky.social, @exponentialview.skystack.xyz’s Hannah Petrovic, and Anson Ho, it depends. Gross margins were around 45%, making inference look profitable.
But after accounting for the cost of operations, OpenAI likely incurred a loss.👇
Can AI solve math research problems that have eluded human mathematicians? Our new benchmark, FrontierMath: Open Problems, is designed to help find out.
AI hasn’t solved any of these yet, but the game is young!
AI hasn’t solved any of these yet, but the game is young!
January 27, 2026 at 4:34 PM
Can AI solve math research problems that have eluded human mathematicians? Our new benchmark, FrontierMath: Open Problems, is designed to help find out.
AI hasn’t solved any of these yet, but the game is young!
AI hasn’t solved any of these yet, but the game is young!
We’ve added new trends & figures categories to our Trends page!
Do you know:
• How fast LLM inference prices are falling?
• How fast compute stocks are growing?
• How long it takes to build a GW scale data center?
Find out on our Trends page:
epoch.ai/trends
Do you know:
• How fast LLM inference prices are falling?
• How fast compute stocks are growing?
• How long it takes to build a GW scale data center?
Find out on our Trends page:
epoch.ai/trends
January 23, 2026 at 9:56 PM
We’ve added new trends & figures categories to our Trends page!
Do you know:
• How fast LLM inference prices are falling?
• How fast compute stocks are growing?
• How long it takes to build a GW scale data center?
Find out on our Trends page:
epoch.ai/trends
Do you know:
• How fast LLM inference prices are falling?
• How fast compute stocks are growing?
• How long it takes to build a GW scale data center?
Find out on our Trends page:
epoch.ai/trends
Models that are good at math benchmarks tend to be good at coding and reasoning benchmarks too, pointing to a common factor driving AI capabilities.
We find that AI benchmark scores are nearly as correlated across domains (0.68) as within them (0.79).
We find that AI benchmark scores are nearly as correlated across domains (0.68) as within them (0.79).
January 23, 2026 at 9:03 PM
Models that are good at math benchmarks tend to be good at coding and reasoning benchmarks too, pointing to a common factor driving AI capabilities.
We find that AI benchmark scores are nearly as correlated across domains (0.68) as within them (0.79).
We find that AI benchmark scores are nearly as correlated across domains (0.68) as within them (0.79).
New record on FrontierMath Tier 4! GPT-5.2 Pro scored 31%, a substantial jump over the previous high score of 19%. Read on for details, including comments from mathematicians.
January 23, 2026 at 6:38 PM
New record on FrontierMath Tier 4! GPT-5.2 Pro scored 31%, a substantial jump over the previous high score of 19%. Read on for details, including comments from mathematicians.
xAI's Colossus 2 data center is running, but likely won't reach 1 GW of power until May, despite prior claims by Elon Musk.
Our updated analysis shows the facility lacks the cooling capacity to run 550,000 Blackwell GPUs at full power, even in winter conditions.
Our updated analysis shows the facility lacks the cooling capacity to run 550,000 Blackwell GPUs at full power, even in winter conditions.
January 19, 2026 at 10:30 PM
xAI's Colossus 2 data center is running, but likely won't reach 1 GW of power until May, despite prior claims by Elon Musk.
Our updated analysis shows the facility lacks the cooling capacity to run 550,000 Blackwell GPUs at full power, even in winter conditions.
Our updated analysis shows the facility lacks the cooling capacity to run 550,000 Blackwell GPUs at full power, even in winter conditions.
AI data centers can now use as much power as New York State uses on the hottest days of the years.
We find that data centers currently have a total capacity of around 30 GW.
We find that data centers currently have a total capacity of around 30 GW.
January 16, 2026 at 11:18 PM
AI data centers can now use as much power as New York State uses on the hottest days of the years.
We find that data centers currently have a total capacity of around 30 GW.
We find that data centers currently have a total capacity of around 30 GW.
How well did forecasters predict 2025 AI progress?
According to the AI Digest's survey, forecasters:
- Mostly nailed benchmark scores
- Underestimated risks from AI-enabled bioweapons
- Underestimated revenue by almost 2×
- Overestimated public concern about AI
Details in 🧵
According to the AI Digest's survey, forecasters:
- Mostly nailed benchmark scores
- Underestimated risks from AI-enabled bioweapons
- Underestimated revenue by almost 2×
- Overestimated public concern about AI
Details in 🧵
January 16, 2026 at 8:42 PM
How well did forecasters predict 2025 AI progress?
According to the AI Digest's survey, forecasters:
- Mostly nailed benchmark scores
- Underestimated risks from AI-enabled bioweapons
- Underestimated revenue by almost 2×
- Overestimated public concern about AI
Details in 🧵
According to the AI Digest's survey, forecasters:
- Mostly nailed benchmark scores
- Underestimated risks from AI-enabled bioweapons
- Underestimated revenue by almost 2×
- Overestimated public concern about AI
Details in 🧵
Our 2025 Impact Report is out.
The AI industry is scaling exponentially - investment, compute, data center buildouts. So, it turns out, is demand for making sense of it all.
See how we’ve kept up!
The AI industry is scaling exponentially - investment, compute, data center buildouts. So, it turns out, is demand for making sense of it all.
See how we’ve kept up!
January 16, 2026 at 6:12 PM
Our 2025 Impact Report is out.
The AI industry is scaling exponentially - investment, compute, data center buildouts. So, it turns out, is demand for making sense of it all.
See how we’ve kept up!
The AI industry is scaling exponentially - investment, compute, data center buildouts. So, it turns out, is demand for making sense of it all.
See how we’ve kept up!
We loved this quick visual rundown of the Frontier Datacenters hub.
Thanks to Rowan Cheung for featuring our project!
www.youtube.com/shorts/szAW...
Thanks to Rowan Cheung for featuring our project!
www.youtube.com/shorts/szAW...
This map reveals some of the hidden facts behind AI data centers 👀 #trendingshorts #ai #research
Epoch AI, a nonprofit research group, is using satellite imagery and public records to track the rapid expansion of AI datacenters across the United States.B...
www.youtube.com
January 13, 2026 at 10:13 PM
We loved this quick visual rundown of the Frontier Datacenters hub.
Thanks to Rowan Cheung for featuring our project!
www.youtube.com/shorts/szAW...
Thanks to Rowan Cheung for featuring our project!
www.youtube.com/shorts/szAW...
Frontier labs are investing massively in RL environments, yet most of what happens in this space stays behind closed doors.
@chrisbarber and @js_denain interviewed 18 people from RL environment startups, neolabs, and frontier labs. Here's what they found:
@chrisbarber and @js_denain interviewed 18 people from RL environment startups, neolabs, and frontier labs. Here's what they found:
January 12, 2026 at 8:43 PM
Frontier labs are investing massively in RL environments, yet most of what happens in this space stays behind closed doors.
@chrisbarber and @js_denain interviewed 18 people from RL environment startups, neolabs, and frontier labs. Here's what they found:
@chrisbarber and @js_denain interviewed 18 people from RL environment startups, neolabs, and frontier labs. Here's what they found:
Anthropic's data center in Indiana is likely the largest in the world today: 750 megawatts by our calculations. Soon, it will pass the gigawatt milestone.
How did they do it, and why do we think it's this big? 🧵
How did they do it, and why do we think it's this big? 🧵
January 9, 2026 at 10:55 PM
Anthropic's data center in Indiana is likely the largest in the world today: 750 megawatts by our calculations. Soon, it will pass the gigawatt milestone.
How did they do it, and why do we think it's this big? 🧵
How did they do it, and why do we think it's this big? 🧵
Total AI compute is doubling every 7 months.
We tracked quarterly production of AI accelerators across all major chip designers. Since 2022, total compute has grown ~3.3x per year, enabling increasingly larger-scale model development and adoption. 🧵
We tracked quarterly production of AI accelerators across all major chip designers. Since 2022, total compute has grown ~3.3x per year, enabling increasingly larger-scale model development and adoption. 🧵
January 9, 2026 at 10:41 PM
Total AI compute is doubling every 7 months.
We tracked quarterly production of AI accelerators across all major chip designers. Since 2022, total compute has grown ~3.3x per year, enabling increasingly larger-scale model development and adoption. 🧵
We tracked quarterly production of AI accelerators across all major chip designers. Since 2022, total compute has grown ~3.3x per year, enabling increasingly larger-scale model development and adoption. 🧵
Global AI compute capacity now totals over 15 million H100-equivalents.
Our new AI Chip Sales data explorer tracks where this compute comes from across Nvidia, Google, Amazon, AMD, and Huawei, making it the most comprehensive public dataset available.
Our new AI Chip Sales data explorer tracks where this compute comes from across Nvidia, Google, Amazon, AMD, and Huawei, making it the most comprehensive public dataset available.
January 8, 2026 at 8:47 PM
Global AI compute capacity now totals over 15 million H100-equivalents.
Our new AI Chip Sales data explorer tracks where this compute comes from across Nvidia, Google, Amazon, AMD, and Huawei, making it the most comprehensive public dataset available.
Our new AI Chip Sales data explorer tracks where this compute comes from across Nvidia, Google, Amazon, AMD, and Huawei, making it the most comprehensive public dataset available.
Since 2023, every model at the frontier of AI capabilities has come from the United States. Chinese models have trailed by 7 months on average.
January 2, 2026 at 9:28 PM
Since 2023, every model at the frontier of AI capabilities has come from the United States. Chinese models have trailed by 7 months on average.
It's been a big year for Epoch – we published more plots in 2025 than in all previous years combined!
December 31, 2025 at 9:59 PM
It's been a big year for Epoch – we published more plots in 2025 than in all previous years combined!
The scale of decentralized AI training runs over the internet has grown exceptionally fast in recent years.
Could they catch up to the frontier of compute? 🧵
Could they catch up to the frontier of compute? 🧵
December 29, 2025 at 5:06 PM
The scale of decentralized AI training runs over the internet has grown exceptionally fast in recent years.
Could they catch up to the frontier of compute? 🧵
Could they catch up to the frontier of compute? 🧵
Benchmarking is crucial to keep track of AI progress. However, benchmarking is hard: each step of the pipeline involves moving parts that can affect the final headline result.
Let's dive deeper:
Let's dive deeper:
December 23, 2025 at 10:24 PM
Benchmarking is crucial to keep track of AI progress. However, benchmarking is hard: each step of the pipeline involves moving parts that can affect the final headline result.
Let's dive deeper:
Let's dive deeper:
AI capabilities accelerated in 2024! According to our Epoch Capabilities Index, frontier model improvement nearly doubled, from ~8 points/year to ~15 points/year.
December 23, 2025 at 8:11 PM
AI capabilities accelerated in 2024! According to our Epoch Capabilities Index, frontier model improvement nearly doubled, from ~8 points/year to ~15 points/year.
Looking back on 2025: what were our most popular short-form research posts?
We published 36 Data Insights and 37 Gradient Updates newsletters.
Here's what our website readers found most interesting: 🧵
We published 36 Data Insights and 37 Gradient Updates newsletters.
Here's what our website readers found most interesting: 🧵
December 23, 2025 at 4:56 PM
Looking back on 2025: what were our most popular short-form research posts?
We published 36 Data Insights and 37 Gradient Updates newsletters.
Here's what our website readers found most interesting: 🧵
We published 36 Data Insights and 37 Gradient Updates newsletters.
Here's what our website readers found most interesting: 🧵
We benchmarked several open-weight Chinese models on FrontierMath. Their top scores on Tiers 1-3 lag the overall frontier by about seven months.
December 22, 2025 at 6:57 PM
We benchmarked several open-weight Chinese models on FrontierMath. Their top scores on Tiers 1-3 lag the overall frontier by about seven months.
We took an updated look at the evidence on AI adoption.
Key finding: AI adoption has continued to grow faster than almost any other technology in history — but the drivers of this have started to change. 🧵
Key finding: AI adoption has continued to grow faster than almost any other technology in history — but the drivers of this have started to change. 🧵
December 20, 2025 at 12:13 AM
We took an updated look at the evidence on AI adoption.
Key finding: AI adoption has continued to grow faster than almost any other technology in history — but the drivers of this have started to change. 🧵
Key finding: AI adoption has continued to grow faster than almost any other technology in history — but the drivers of this have started to change. 🧵
Over the past month we've partnered with @blueroseorg to survey 5,660 Americans about their AI usage habits. Today we're releasing the results.
A key takeaway: A majority of Americans use AI on a weekly basis, with 35% using ChatGPT, 24% Gemini, and 13% Meta AI.
A key takeaway: A majority of Americans use AI on a weekly basis, with 35% using ChatGPT, 24% Gemini, and 13% Meta AI.
December 20, 2025 at 12:01 AM
Over the past month we've partnered with @blueroseorg to survey 5,660 Americans about their AI usage habits. Today we're releasing the results.
A key takeaway: A majority of Americans use AI on a weekly basis, with 35% using ChatGPT, 24% Gemini, and 13% Meta AI.
A key takeaway: A majority of Americans use AI on a weekly basis, with 35% using ChatGPT, 24% Gemini, and 13% Meta AI.
Gemini 3 Flash scored 36% on FrontierMath Tiers 1–3, comparable to top models. It scored comparatively less well on the harder Tier 4.
December 19, 2025 at 5:47 PM
Gemini 3 Flash scored 36% on FrontierMath Tiers 1–3, comparable to top models. It scored comparatively less well on the harder Tier 4.