epoch.anthropic.inference_compute_usd.fy

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
2a23fc81cd16d222	$2.7B	grade C, evidence tier 5	2025-12-31	2026-01-21	5	estimated	api	approved	2026-09-11 10:55	200	9b2f4ca0bcc7	—	source page

Raw snippets and dispute text

2a23fc81cd16d222 {"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"}