R&D compute allocation
Epoch AI's annotated chart drew 445K views on X: 6.5x its typical post and #6 of 97 Epoch AI charts we track by views. Read Oct 10, 2026.
New data insight: How does OpenAI allocate its compute?
Why it spread
It spread because a top quarter source reach and a top quarter quote rate combined with 661 likes, 83 reposts, 56 quotes, 317 saves and 445,020 views, driven by many quotes and an independent author.
What worked
- The post leads with a plain question about how OpenAI allocates its compute, then answers it with a single striking split: most went to research and development, and only a minority of that went to final training runs of released models.
- It is self labelling and branded, with a source line, so readers can trust and cite it, which fits the many quotes and the top quarter quote rate.
- The framing is counterintuitive but simple, since a roughly $7 billion compute figure is paired with the claim that most of it funded experiments rather than released models.
What to borrow
We could take one large internal number, split it into the part that went to exploration and the part that went to final production, and state the split as a short question and answer with our own source line.
Careful
Do not copy the exact OpenAI framing or treat the split as a general rule, and do not over read the 56 quotes as agreement, since quotes can also be skeptical.
Written by our analysis model from the measured numbers above, not by Epoch AI.
Make it yours
We have not read this chart's numbers yet
Its own maker appears here once its numbers are read. Until then, these are other real charts in a similar format: each data page starts from that chart's numbers and takes yours.
API · Max
Use this chart via API
On the Max plan you can pull this chart's numbers with the provenance of each one, as JSON or CSV, with its redraw spec, or call it from an agent over MCP. Chart id 62c4d32d6f67.
| Path | Returns | Plan | |
|---|---|---|---|
| GET | /api/v1/charts/62c4d32d6f67 | Title, publisher, date, form, point count; with a Max key, the data link | Open |
| GET | /api/v1/charts/62c4d32d6f67/data | Every point with its provenance and confidence, plus the redraw spec. Add ?format=csv for CSV | Max |
| GET | /api/v1/charts/62c4d32d6f67/render?view=1 | The draw-only view this page uses | Open |
| POST | /mcp | Tool get_chart_data with this id, for Claude, Cursor and other agents | Max |
curl -s https://viral-chart.com/api/v1/charts/62c4d32d6f67/data \
-H "Authorization: Bearer $VIRALCHART_KEY"claude mcp add --transport http viralchart https://viral-chart.com/mcp \
--header "Authorization: Bearer $VIRALCHART_KEY"60 requests a minute per key; over that you get 429 with Retry-After. Up to 5 keys per account, made on your account page and shown once. Without Max, the data route answers 402 with a link to the plans. The numbers describe a chart its publisher owns, and every file carries a license note.
More from Epoch AI
All 97
Other giants using a annotated chart
All 102
Questions people ask
How far did Epoch AI's "R&D compute allocation" chart travel?
445,020 views, 661 likes, 83 reposts and 317 saves on X, read Oct 10, 2026. A normal post from @EpochAIResearch gets about 68.2K views.
What kind of chart is it?
A annotated chart. Most of OpenAI's 2024 compute went to experiments.
Can I use Epoch AI's chart?
Use the format, not the picture. The chart, logo and design belong to Epoch AI. ViralChart redraws the chart type with your numbers and words, and credits the original publisher.