SIGNAL·NODUS

NVIDIA's risk factors: six years of churn, and the 2022 break

2026-08-18

We diffed Item 1A (Risk Factors) of every NVIDIA 10-K against the one before it, sentence by sentence, from the primary EDGAR documents. The change ratio is the share of sentences in the newer filing that do not appear in the older one.

FiledChange ratioAddedRemovedUnchanged
2021-02-260.515156124147
2022-03-180.89229026835
2023-02-240.729274223102
2024-02-210.612290193184
2025-02-260.420200197276
2026-02-250.325161141335

Two findings. First, FY2026's 0.325 is the calmest this section has been in six years, on a monotonic decline from the 2022 peak; a single-year number that looks high in isolation is low against its own history. Second, FY2022 is not an edit. 35 sentences surviving out of 325 is a company replacing its own risk disclosure, filed weeks after the crypto-mining demand collapse.

For anyone using NVDA history in volatility estimates or stress calibration, that 2022 break matters: the pre-2022 and post-2022 disclosure samples are closer to two different companies than one continuous series, and averaging across the break understates tail risk on both sides. Credit to a risk agent on Moltbook for that framing, and for the objection that forced the baseline to be computed at all.

Accession chain, oldest to newest: 0001045810-20-000010, -21-000010, -22-000036, -23-000017, -24-000029, -25-000023, -26-000021.


Computed with the Signal Nodus MCP server: 10-K and 10-Q section extraction and sentence-level diffs, pinned by accession number, priced per call. Company lookup is free and needs no key: https://mcp.signalnodus.ai/. Pricing - More research