Dislocation

Dislocation

Comcast has lost roughly half its market value — $45.14 on 6 November 2024 to a $21.92 trough on 23 July 2026, down 51.4% over 624 days. The fall came in identifiable legs on dated earnings events, but the trigger was never an EPS miss: the company beat consensus every quarter through the decline. What repriced was the broadband franchise — subscriber losses, then the ARPU and EBITDA cost of defending them. The volume gauge is present but modest at 2.5x.

The drawdown, quantified

Source: derived from daily price history; drawdown figures per fit_features.capitulation_gauge.drawdown (peak $45.14, trough $21.92, −51.4%, 624 days).

The peak-to-trough decline is 51.4% — deep, but short of the 60–70% forced-selling zone the framework hunts. It unfolded not as a single crash but as a staircase: two sharp legs on the FY2024 broadband print, a long plateau through 2025, then a second descent through 2026 to the July trough.

Source: derived from daily price history, month-end closes.

The legs, dated to their trigger events:

Source: derived from daily price history; event dates matched to Comcast earnings releases and transcripts.

Between Leg 2 (Jan 2025) and Leg 3 (Apr 2026) the stock spent fourteen months in a $31–37 band, absorbing a market-wide 6.6% fall on 4 April 2025 (the tariff selloff, not a Comcast event) but making no new low. The dislocation is two clustered event legs bracketing a long drift, not one continuous capitulation.

The trigger

The identifiable adverse event is the broadband subscriber print, and it arrived twice. At an early-December 2024 investor conference management flagged a soft fourth quarter; the stock fell 9.5% on 9 December 2024. The confirmation came on 30 January 2025, when Comcast reported net broadband subscriber additions of negative 139,000 for Q4 2024 — "disappointing and worse than what we indicated in early December" — with management adding they "see no signs of this changing in the near term" [1]. The stock fell 11.0% that day. Critically, the same call reported record 2024 revenue and $12.5 billion of free cash flow [2] — the repricing was about units and the competitive path, not the current-year cash number.

The competitive mechanism is named in the filings: 5G fixed-wireless networks from the mobile carriers and rapid fiber overbuild, compounded in 2024 by the end of the federal Affordable Connectivity Program subsidy [3] [4]. Over the full year 2025, total domestic broadband customers fell 711,000 to 31.3 million — a wider loss than 2024's 411,000 — and broadband penetration of homes passed slipped to 47.6% from 49.8% [5].

The 2026 legs carry a different, telling character. On 24 April 2026 the stock fell 12.9% — the single worst day of the whole decline — on a quarter where the news on units was good: broadband losses improved by 117,000 year-over-year to 65,000, the first year-over-year improvement since Q4 2020, and wireless posted record net additions. What sold the stock was the cost of that improvement. Adjusted EBITDA declined 9%, and broadband ARPU fell 3.1%, as simplified pricing and a bundled free-wireless-line offer diluted revenue per customer [6]. The final leg to the trough came on 23 July 2026, when Q2 broadband losses of 167,000 — themselves a 34,000 year-over-year improvement — were not enough to stem the slide [7].

The trigger, then, evolved. In 2024–25 it was the volume of subscriber losses; by 2026 it was the margin price of defending the base. Both are the same underlying fear — that fixed wireless and fiber have structurally changed the broadband franchise — expressed through different line items.

The fear gauge

The measured capitulation is present but not violent. fit_features records a volume spike of 2.46x, defined as the peak 20-day average volume during the fall divided by the median daily volume over the 180 days before the peak (about 17.3 million shares).

Source: derived from daily volume history; median pre-peak volume ≈17.3M shares (180 sessions before 6 Nov 2024).

Two features matter. First, where the sustained-volume peak sat: the heaviest 20-day window ended around 1 July 2026 — late in the fall, near the trough, not at the January 2025 trigger. To the extent there was emotional selling, it clustered at the bottom rather than at the first bad news, which is the pattern the framework looks for. Second, the single largest raw-volume days of the decline — 246 million shares on 20 March 2026 and 134 million on 29 June 2026 — fall on index-rebalance and quarterly-expiry dates, so they read as mechanical passive flow rather than fear. The genuine event-driven spikes (3–4.5x the pre-peak median on the earnings days) are real but sit well below the double-digit multiples that mark true capitulation. This is orderly, event-anchored repricing with a fear component, not a panic.

Who was selling

The direct evidence here is thin, and the honest statement is that seller composition cannot be pinned down. Official reported short-interest data for CMCSA is unavailable in this run — FINRA returned no position rows, so short level, change, and days-to-cover are all blank. No forced-seller disclosures (fund liquidations, threshold net-short filings) are staged. Comcast remains an S&P 500 and Nasdaq-100 constituent, so there was no index-exit forced-sale event.

What can be said comes from the trading record. Liquidity is deep — roughly $900 million of stock changes hands daily and annual turnover runs about 164% of the float — so no holder was trapped by an inability to exit. The concentration of the largest-volume sessions on rebalance and expiry dates points to passive and index-linked flow reweighting a shrinking market cap, alongside the informed event-day selling on the earnings prints. Absent short-interest and holder-flow data, the balance between anchored/forced sellers and informed sellers is a genuine gap rather than a finding.

Estimates versus price

Consensus earnings did fall, but the price fell roughly twice as far. Over the past six months, consensus FY2027 EPS was cut from $4.10 to $3.64 (−11%) and FY2028 from $4.53 to $3.96 (−13%), while forward revenue eased only about 2%. Over the same window the stock fell about 24%.

Source: FY2027 consensus EPS revision path (CapIQ, data/sp/estimates.json) vs CMCSA close on each revision date; derived, indexed to 100 at 26 Jan 2026.

The two lines tracked each other into April 2026, then diverged: from April to July the price dropped a further 18% while the EPS estimate barely moved. The wider frame is starker. Across the full 51% market-cap decline, consensus forward free cash flow held in a $13–17 billion band and forward revenue stayed near $120 billion; against the current market cap that forward FCF computes to a 16–19% yield before any framework adjustment (fit_features.consensus_forward_yield). A price cut roughly double the estimate cut is the framework's dislocation signature.

Bottom line

There is a real dislocation here: a 51% market-cap decline against forward free cash flow and revenue that barely moved, on dated, cited triggers, with the price outrunning the estimate cut by roughly two to one. But it does not present as the clean Centene-style single-event anchor. Comcast beat EPS every quarter of the fall; the reprice came in two clustered legs a year apart and reflects a persistent, structural fear about the broadband franchise — fixed-wireless and fiber competition — expressed first through subscriber counts and then through the ARPU and EBITDA cost of defending them. The capitulation gauge is modest (2.5x) and the depth stops short of the 60–70% forced-selling zone. Whether the damage to units and margins is temporary or structural is not settled here — that question belongs to Damage Math and the trial. This tab records what happened: a deep, event-anchored, but orderly and structurally-themed repricing.