Indicators / series
epoch.anthropic.compute_spend_usd.fy
Source: Epoch AI data hub — AI companies (Epoch AI) · CC BY 4.0 · CSV
Epoch AI, 'Data on AI Companies' (CC BY 4.0), epoch.ai/data
| id | value | grade | as of date | published date | tier | audited vs reported | extraction method | review status | retrieved at | http status | content hash | supersedes id | url | flags |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0e2aa9f664a60e70 | $285M | grade C, evidence tier 5 | 2023-12-31 | 2023-11-10 | 5 | estimated | api | approved | 2026-09-10T01:28:41.972289Z | 200 | 90c5337eb305 | — | source page | |
| 3d0339072044d386 | $2.5B | grade C, evidence tier 5 | 2024-12-31 | 2024-07-24 | 5 | estimated | api | approved | 2026-09-10T01:28:41.972289Z | 200 | 90c5337eb305 | e95aa7fbad569d93 | source page | |
| d4728620ab99f64c | $2.7B | grade C, evidence tier 5 | 2025-12-31 | 2026-01-21 | 5 | estimated | api | approved | 2026-09-10T01:28:41.972289Z | 200 | 90c5337eb305 | 62fb114da668dbdd | source page |
Raw snippets and dispute text
- 0e2aa9f664a60e70 {"Amount": "285000000", "Category": "R&D cloud compute,Inference cloud compute", "Company": "Anthropic", "Confidence": "Likely", "Date": "2023-12-31", "Id": "Anthropic 2023 cloud compute", "Notes": "\"In May, Anthropic told investors it planned to spend $285 million on cloud services this year\"", "Period type": "Year", "Report date": "2023-11-10", "Source 1": "https://www.theinformation.com/articles/pro-weekly-ai-pushes-cloud-vendors-to-new-heights?rc=9mzoog", "Source Type": "Media report", "Total compute spend": "285000000.0"}
- 3d0339072044d386 {"Amount": "2500000000", "Category": "R&D cloud compute,Inference cloud compute", "Company": "Anthropic", "Confidence": "Likely", "Date": "2024-12-31", "Id": "Anthropic 2024 cloud compute", "Notes": "According to source 1, from July 2024, Anthropic estimated $2.5 billion in compute spend in 2024\n\nPer source 2 from February 2025, Anthropic's cash burn in 2024 was $5.6B (including all expenses), 3B higher than projected, and more than half of the cash burn (i.e. ~3B) came from a one-off payment to access data centers. This implies: 1) that this one-off payment accounts for the unexpected increase in cash burn, meaning their expense projections were otherwise accurate (ignoring any forecast error in revenue which is probably minor: Anthropic's ARR was only 1B at end of 2024) and 2) their total compute-related cash payments in 2024 were actually around $5.5B.\n\nThe nature of this one-off cash payment is unclear, but probably doesn't represent operating expenses that should be accounted for in 2024.", "Period type": "Year", "Report date": "2024-07-24", "Source 1": "https://www.theinformation.com/articles/why-openai-could-lose-5-billion-this-year?rc=9mzoog", "Source 2": "https://www.theinformation.com/articles/anthropic-projects-soaring-growth-to-34-5-billion-in-2027-revenue?rc=9mzoog", "Source Type": "Media report", "Total compute spend": "2500000000.0"}
- d4728620ab99f64c {"Amount": "2700000000", "Category": "R&D cloud compute,Inference cloud compute", "Company": "Anthropic", "Confidence": "Likely", "Date": "2025-12-31", "Id": "Anthropic 2025 inference", "Inference compute spend": "2700000000.0", "Notes": "\"Anthropic last month expected its costs for training its AI models for 2025 to be roughly $4.1 billion.\" In context this presumably means full R&D spending, not just non-research training, given reported $2.7 billion.\n\nGraph in same article indicates $2.7B inference costs, with a gross margin of 40% on $4.5B in revenue (1 - 40% * $4.5B = $2.7B)\n\n", "Period type": "Year", "Report date": "2026-01-21", "Source 1": "https://www.theinformation.com/articles/anthropic-lowers-profit-margin-projection-revenue-skyrockets?rc=9mzoog", "Source Type": "Media report"}