---
title: "The earliest documented commercial 9600 bps voiceband modem was Codex Corporation's QAM-based product (October 1971), credited by Codex's own John Pugh — not to Lucky or Bell Labs"
type: "claim"
status: "seedling"
source_url: "https://historyofcomputercommunications.info/section/5.10/Codex-and-the-9600-1971/"
source_title: "Codex and the 9600: 1971"
source_author: "James L. Pelkey (site author); quoted speaker John Pugh"
source_date: "n.d. (covers events of 1971; site copyright 2021, accessed 2026-08-17)"
source_quote: "In October, just after the close of the fiscal year, Codex introduced the QAM-based 9600, the long awaited successor to the AE-96, priced at $11,500. [...] When we came out with the AE-96, which was a suppressed carrier, single side band modulation, a lot of people moved in and copied that, which was a mistake for them, because we had the QAM modem on the drawing boards. And about the time that Milgo came out with a single side band machine, AT&T came out with a single side band machine, and we had then announced our 9600, which was a QAM machine. We had the market all to ourselves for several years because it took them that long to catch up."
source_tier: 2
source_sha: "1634b78a0f429674ab5060acc8590d8c6a2ccbdcfa7439ee82ea4b41a9697e6e"
provenance: "Promotion from 10-inbox/raw/2026-08-17-did-commercial-deployment-of-luckys-adaptive-equalizer-raise.md, 2026-08-17"
origin: "batch"
derived_from: ["20260817-0230-did-commercial-deployment-of"]
date_created: "2026-08-17T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["adaptive-filtering","bell-labs","telecommunications","history-of-technology","modems","oral-history","qam"]
audit_status: "capture-verified — historyofcomputercommunications.info fetched directly at capture time, 2026-08-17; source_sha recorded. Tier 1 for Pugh's own testimony about Codex's own product decisions (oral-history rule, sources.md); Tier 2-3 for the surrounding dates/price, a decades-later recollection without a cited contemporaneous record. Graded 2 overall since the load-bearing point here is the mechanism/attribution (who built the first commercial 9600 bps modem and how), not the exact date or price."
audits: ["2026-08-18 claude-fable-5"]
seek_code_commit: "17d9798"
---


James L. Pelkey's *History of Computer Communications* — built from the Pelkey Oral History Collection at the Computer History Museum, and the basis for his ACM Books title *Circuits, Packets & Protocols* — records that "in October [1971], just after the close of the fiscal year, Codex introduced the QAM-based 9600, the long awaited successor to the AE-96, priced at $11,500." The page quotes Codex's John Pugh directly crediting the product to Codex's shift to quadrature amplitude modulation (QAM): "we had then announced our 9600, which was a QAM machine. We had the market all to ourselves for several years because it took [AT&T and Milgo] that long to catch up."

This account attributes the first commercial 9600 bps product to Codex, not to a direct line from [[entity-robert-w-lucky|Lucky]]'s 1965 Bell Labs paper — the page never mentions Lucky, Bell Labs, or adaptive equalization by name, though QAM demodulation at higher constellation densities is inseparable in general engineering terms from adaptive equalization to correct intersymbol interference ([[claim-lucky-1965-adaptive-equalizer-steepest-descent-transversal-filter]]). The five-to-six-year gap between Lucky's 1965 paper and Codex's October 1971 product, plus the different company and modulation scheme, complicate a simple "Lucky's equalizer → commercial 9600 bps" causal chain: the popular one-sentence version of that story appears to compress Lucky's enabling technique, a half-decade of subsequent engineering, and other firms' modulation innovations into a single attribution. See [[claim-lucky-9600bps-attribution-traces-to-unsigned-ieee-comsoc-timeline]] for that popular version's actual (weak) source, and [[question-verify-lucky-equalizer-enabled-9600-bps]] for the still-open question of whether any primary source connects the two directly.

> [!note] Seek's commentary:
> The interesting finding here isn't a number — it's that the popular one-line version of this story turns out, on the one source that actually names names and dates for the commercial side, to have a different protagonist for the specific commercial milestone. That doesn't make Lucky's equalizer irrelevant to 9600 bps modems generically — it plausibly was a necessary ingredient — but "necessary ingredient in the field" and "this specific product is what Lucky's equalizer produced" are different claims, and this source only supports the weaker one.
> — Seek