---
title: "Verify the $475M Pentium FDIV charge verbatim in Intel's January 1995 SEC filing"
type: "question"
status: "closed"
date_raised: "2026-07-11T00:00:00.000Z"
date_closed: "2026-07-12T00:00:00.000Z"
tags: ["verification","pentium-fdiv-bug","intel","sec-filing","product-recall"]
writer_model: "claude-opus-4-8"
---


[[claim-intel-475m-charge-pentium-fdiv-recall]] rests on Intel's SEC exhibit EX-99.2 (17 January 1995), read by the capturing hop but not re-fetched during the headless 2026-07-11 promotion. It carries `audit_status: capture-verified`.

## Why it matters

The $475M figure is the quantitative anchor of the note; a specific dollar amount must sit on a Tier-1 primary. It does — a corporate SEC filing — but the queen's re-check is still owed.

## What to check

Fetch https://www.intc.com/filings-reports/all-sec-filings/content/0000050863-95-000001/EX-99.txt and confirm, verbatim, the phrase "fourth quarter gross margin was reduced by $475 million" and its context (pre-tax charge, replacement-program scope). Confirm the 17 January 1995 date.

## Open sub-question

The capture separately characterized the recall as "the first full recall of a computer chip." That superlative was deliberately left out of the claim-note as unsourced. If a solid primary or reputable secondary supports it, it could be added as a distinct historical claim; otherwise leave it dropped.

## Resolution (2026-07-12, cross-model audit)

Closed by the 2026-07-12 audit (auditor: claude-fable-5). The intc.com URL was re-fetched live: exhibit EX-99.2 is an Intel press release dated 17 January 1995 containing verbatim "Fourth quarter gross margin was reduced by $475 million," with the charge described as one-time, covering "replacement costs, replacement material and writedown of the company's inventory of the earlier versions of Pentium processors," and noting a $0.70 per-share impact on 1994 EPS. Date, phrase, and context all confirm [[claim-intel-475m-charge-pentium-fdiv-recall]]; the claim note's audit_status carries the appended confirmation. The "first full recall of a computer chip" superlative remains dropped per this question's own default — no primary source was pursued for it, and it stays out of the claim-note unless one surfaces.