Phoenix list · named 2026-08-29 · baseline 2026-08-28 close
Five household names that stopped falling.
Broken Stocks tells you when a stock is in structural decline. This page is the other half of the
question: which of those broken names has actually finished breaking? Each of the five below fell
35-70% over a year or more, washed out the people who bought the first bounce, and has now put in a
low that confirmed on the monthly chart. The businesses underneath kept compounding
the whole way down.
We're naming them in public on purpose. The baseline is the 2026-08-28 close; on
2027-08-28 we'll publish how each one did against that number, and against the S&P 500.
Price is checked live below. This is recognition of a pattern, not a recommendation to buy anything.
Scorecard rules. The list is frozen. No swaps, no additions, no quiet removals: if one of the five breaks down again it stays on the scorecard and counts against it. Each name is scored on price change from the baseline close; the five together are scored as an equal-weight basket against the S&P 500 over the same dates. Dividends are ignored on both sides.
1
Netflix (NFLX)
Communication Services · $340.3BBusiness health 89/100 · Strong
The decline
−51.5%$134.12 (2025-06) to $65.08 (2026-07), 13 months
Monthly low confirmed
next monthtook out $78.44, the high of the low month
Where in the cycle
Month 1 · week 6of the new monthly and weekly cycle
Daily closes, 2025-05-01 to 2026-08-31. Dashed green: the monthly trigger, the high of the low month, that a later month had to clear. Dotted gold: the close it was named at.
The trap
February printed a green outside bar at $75.01 and confirmed in March; April made a higher high then reversed on a long upper wick; May through July undercut it. Everyone who bought the February bottom was stopped out before the real low.
Why it fell
Viewing hours per subscriber turned down year over year, investors priced in AI disruption to content and heightened competition after a wave of media M&A, founder Reed Hastings left, and in July the company narrowed its full-year revenue guide. The stock fell for 13 months and gave back more than half its value.
Why it's turning
The Q2 engagement softness is being read as a content-timing gap rather than a broken model, with a stronger second-half slate ahead. The business kept compounding through the whole decline; the chart put in its low in July and August confirmed it.
The business
FY2025 revenue $45.2B (+15.9%), operating margin 29.5%, net margin 24.3%, free cash flow $9.5B (20.9% of revenue), return on equity 41%, debt/equity 0.51, share count down 3.4% over three years.
FY2021
FY2022
FY2023
FY2024
FY2025
Trend
Peers
Revenue $B
29.7
31.6
33.7
39.0
45.2
Revenue growth
18.8%
6.5%
6.7%
15.6%
15.9%
improving
4.2%
Operating margin
20.9%
17.8%
20.6%
26.7%
29.5%
improving
2.0%
Free-cash-flow margin
-0.4%
5.1%
20.5%
17.7%
20.9%
improving
9.3%
Net cash (cash − debt) $B
-8.7
-9.2
-7.0
-6.0
-4.4
improving
Diluted shares (indexed) FY2021 = 100
100
99
99
97
96
shrinking
+4.3% 3-yr
Fiscal years from 10-K filings (SEC EDGAR XBRL), pulled 2026-08-29.
Peers: DIS, WBD, ROKU, FOX — median of each peer's latest fiscal year; NFLX's own three-year share change is -3.4%.
Trend: latest fiscal year vs the average of the two before it; growth ±2 pts, margins ±1 pt, net cash ±10%, shares ±1%.
Download all five as a spreadsheet.
The setup
Awaiting the first weekly pullback low after the July trough. Five straight up weeks so far; the first higher weekly low that confirms is the setup.
Daily closes, 2025-06-02 to 2026-08-31. Dashed green: the monthly trigger, the high of the low month, that a later month had to clear. Dotted gold: the close it was named at.
The trap
The March low confirmed in April and made two higher highs before June undercut it by a few dollars, the classic wash-out just above the prior low.
Why it fell
AI capital spending shocked the market: quarterly capex ran above $30B, up more than 80% year over year, while Azure growth decelerated. A securities class action alleged the deceleration had been played down, a new federal AI software review program added regulatory noise, and the stock fell into a bear market, more than 20% under its June 1 peak at one point and down about 27% for the year.
Why it's turning
The spending is the moat argument in reverse: the capacity is being built for demand that's already contracted. Analysts stayed overwhelmingly bullish through the drop, and price confirmed the June low in July.
The business
FY2026 revenue $331.8B (+17.8%), operating margin 46.8%, net margin 40.3%, free cash flow $67.0B, return on equity 30%, debt/equity 0.09. The highest health score in the five, and nothing in the filing that explains a 37% drawdown.
FY2022
FY2023
FY2024
FY2025
FY2026
Trend
Peers
Revenue $B
198.3
211.9
245.1
281.7
331.8
Revenue growth
18.0%
6.9%
15.7%
14.9%
17.8%
improving
13.8%
Gross margin
68.4%
68.9%
69.8%
68.8%
67.9%
worsening
53.3%
Operating margin
42.1%
41.8%
44.6%
45.6%
46.8%
improving
31.3%
Free-cash-flow margin
32.9%
28.1%
30.2%
25.4%
20.2%
worsening
9.7%
Net cash (cash − debt) $B
-35.9
-12.5
-26.6
-12.9
-19.4
stable
Diluted shares (indexed) FY2022 = 100
100
99
99
99
99
stable
-0.8% 3-yr
Fiscal years from 10-K filings (SEC EDGAR XBRL), pulled 2026-08-29.
Peers: AAPL, GOOGL, AMZN, ORCL — median of each peer's latest fiscal year; MSFT's own three-year share change is -0.3%.
Trend: latest fiscal year vs the average of the two before it; growth ±2 pts, margins ±1 pt, net cash ±10%, shares ±1%.
Download all five as a spreadsheet.
The setup
First weekly pullback low after the trough confirmed the week of August 21 at $477.15 (trigger $492.66).
Technology · $115.9BBusiness health 90/100 · Strong
The decline
−70.2%$638.25 (2024-02) to $190.12 (2026-06), 28 months
Monthly low confirmed
two months latertook out $275.44, the high of the low month
Where in the cycle
Month 2 · week 10of the new monthly and weekly cycle
Baseline
$291.522026-08-28 close
Since named
+0.4%$292.79 on 2026-08-31
False bottom first?
No
high $638 · 2024-02low $190.12 · 2026-06confirm $275.44named $291.52
Jun 24Nov 24Apr 25Sep 25Feb 26Jul 26
Daily closes, 2024-01-02 to 2026-08-31. Dashed green: the monthly trigger, the high of the low month, that a later month had to clear. Dotted gold: the close it was named at.
Why it fell
A 28-month, 70% decline on one fear: that generative AI erodes the creative-software moat. CEO succession headlines and the CFO's June departure piled on, and the stock bottomed at $190 the week after a record quarter.
Why it's turning
The June quarter was a record and the guide went up while the stock was making its low; that's the fundamental-versus-price divergence this list exists to catch. Shares have recovered roughly half of the decline's last leg and the monthly confirmed in August.
The business
FY2025 revenue $23.8B (+10.5%), operating margin 36.6%, net margin 30.0%, free cash flow $9.9B (41% of revenue), return on equity 61%, share count down 9.3% over three years. The June quarter that followed was a record ($6.62B, +13%) with the full-year guide raised and AI-first ARR past $500M.
FY2021
FY2022
FY2023
FY2024
FY2025
Trend
Peers
Revenue $B
15.8
17.6
19.4
21.5
23.8
Revenue growth
22.7%
11.5%
10.2%
10.8%
10.5%
stable
16.6%
Gross margin
88.2%
87.7%
87.9%
89.0%
89.3%
stable
77.7%
Operating margin
36.8%
34.6%
34.3%
31.3%
36.6%
improving
21.0%
Free-cash-flow margin
43.6%
42.0%
35.8%
36.6%
41.4%
improving
34.0%
Net cash (cash − debt) $B
-0.3
0.6
3.5
2.0
-0.8
worsening
Diluted shares (indexed) FY2021 = 100
100
98
95
93
89
shrinking
-0.9% 3-yr
Fiscal years from 10-K filings (SEC EDGAR XBRL), pulled 2026-08-29.
Peers: INTU, NOW, ADSK, CRM — median of each peer's latest fiscal year; ADBE's own three-year share change is -9.3%.
Trend: latest fiscal year vs the average of the two before it; growth ±2 pts, margins ±1 pt, net cash ±10%, shares ±1%.
Download all five as a spreadsheet.
The setup
First weekly pullback low after the trough confirmed the week of August 21 at $251.20 (trigger $277.10).
Daily closes, 2024-11-01 to 2026-08-31. Dashed green: the monthly trigger, the high of the low month, that a later month had to clear. Dotted gold: the close it was named at.
The trap
The August 2025 low confirmed in September and made three higher highs into December before the stock rolled over and lost another 40% into June.
Why it fell
The whole seat-based software model went on trial: if AI agents replace the humans who hold licenses, what happens to per-seat pricing? Add slowing organic growth, a Morgan Stanley downgrade, sector-wide software selling and rising yields, and the stock fell from $268 to $146 in eight months.
Why it's turning
The market is starting to price the agent business as the growth engine rather than the threat. The June low confirmed on the monthly in August and the enterprise-software group is rallying with it.
The business
FY2026 revenue $41.5B (+9.6%), operating margin 20.1%, net margin 18.0%, free cash flow $14.4B (35% of revenue), debt/equity 0.24, share count down 4.1% over three years under a $25B buyback. Agentforce ARR is up 205% year over year to $1.2B. One flag: a current ratio of 0.76, normal for a subscription business that collects cash up front.
FY2022
FY2023
FY2024
FY2025
FY2026
Trend
Peers
Revenue $B
26.5
31.4
34.9
37.9
41.5
Revenue growth
24.7%
18.3%
11.2%
8.7%
9.6%
stable
16.4%
Gross margin
73.5%
73.3%
75.5%
77.2%
77.7%
improving
77.5%
Operating margin
2.1%
3.3%
14.4%
19.0%
20.1%
improving
19.9%
Free-cash-flow margin
19.9%
20.1%
27.2%
32.8%
34.7%
improving
30.8%
Net cash (cash − debt) $B
-5.1
-3.6
-1.0
0.4
-7.1
worsening
Diluted shares (indexed) FY2022 = 100
100
102
101
100
98
shrinking
+3.5% 3-yr
Fiscal years from 10-K filings (SEC EDGAR XBRL), pulled 2026-08-29.
Peers: NOW, ORCL, WDAY, INTU — median of each peer's latest fiscal year; CRM's own three-year share change is -4.1%.
Trend: latest fiscal year vs the average of the two before it; growth ±2 pts, margins ±1 pt, net cash ±10%, shares ±1%.
Download all five as a spreadsheet.
The setup
First weekly pullback low after the trough confirmed the week of August 21 at $189.91 (trigger $211.07).
Daily closes, 2024-05-01 to 2026-08-31. Dashed green: the monthly trigger, the high of the low month, that a later month had to clear. Dotted gold: the close it was named at.
The trap
November's low confirmed in December and made two higher highs before the stock slid under it in June by less than two dollars.
Why it fell
Two years of traffic worries: transactions flat to slightly down, average check slipping, food and labor costs rising as a share of revenue, and heavier promotions to win customers back. The market decided the growth story was over and cut the stock by 60%.
Why it's turning
Comparable sales and transactions are accelerating again, J.P. Morgan upgraded to Overweight on valuation plus that trajectory, and the June low confirmed on the monthly the very next month.
The business
FY2025 revenue $11.9B (+5.4%), operating margin 16.2%, net margin 12.9%, free cash flow $1.4B, return on equity 54%, zero debt, share count down 4.3% over three years. Comparable sales turned positive again in Q2 2026 (+2.2%, transactions +1%).
FY2021
FY2022
FY2023
FY2024
FY2025
Trend
Peers
Revenue $B
7.5
8.6
9.9
11.3
11.9
Revenue growth
26.1%
14.4%
14.3%
14.6%
5.4%
worsening
6.2%
Operating margin
10.7%
13.4%
15.8%
16.9%
16.2%
stable
19.6%
Free-cash-flow margin
11.1%
9.8%
12.4%
13.4%
12.1%
stable
13.3%
Net cash (cash − debt) $B
0.8
0.4
0.6
0.7
0.4
worsening
Diluted shares (indexed) FY2021 = 100
100
100
99
98
96
shrinking
-2.4% 3-yr
Fiscal years from 10-K filings (SEC EDGAR XBRL), pulled 2026-08-29.
Peers: MCD, SBUX, YUM, CAVA — median of each peer's latest fiscal year; CMG's own three-year share change is -4.3%.
Trend: latest fiscal year vs the average of the two before it; growth ±2 pts, margins ±1 pt, net cash ±10%, shares ±1%.
Download all five as a spreadsheet.
The setup
A weekly pullback low confirmed the week of August 21 at $31.64 (trigger $33.88); it's the third since the trough, so the cleanest entry has already printed.
Every U.S. common stock over $10B in market value (about 900 names) is run through the same rules-based read used for the Broken List, then filtered the other way:
A deep, long decline. At least 30% peak to trough over at least three months. A one-month crash is a broken stock; a year-long grind is a washed-out one.
A confirmed monthly low. The lowest month of the decline, followed by a later month that traded above that month's high while the low held. Until that happens, a low is a guess. A stock that only confirms on the weekly chart is tracked but doesn't make the list.
Early in the new cycle. Ranked by how few weeks and months have passed since the low, measured against each stock's own history of how long its cycles run.
A false bottom before the real one. A bonus, not a requirement: an earlier low that confirmed, rallied for two or more months, then broke. That's the pattern that stops out the first buyers, and it's why the real low is quiet.
A business that kept working. A 0-100 health score computed from SEC filings (growth, margins, cash flow, balance sheet, share count) for the top-ranked charts. It informs the pick; it doesn't make it.
From the ranked list we chose five by hand, favoring the strongest businesses with the cleanest structure. Late confirmations count the same as fast ones; a low that consolidates for a few months before it clears the trigger is still a confirmed low. Everything else on this site stays what it is: the Broken List is price action only, and a name can be on both lists at once.
What this setup has done historically
We tested the rule instead of asserting it. Every U.S. stock with usable daily history from 2016 to 2026-08-28 — 3,585 companies, 15,357 qualifying events — was replayed through the same detector that picks these names, and every outcome was measured against the S&P 500 over the same dates.
Narrowed to the population these five came from — above $10B in market value, with a false bottom before the real one — that's 1,970 events across 764 companies, 1,806 of them far enough in the past to have a full year behind them. Those 1,806 are what the one-year figures below are counted on:
69.2% were higher a year later, median +16.6%. At six months it's 63.8%, at three months 60.6%.
44.5% of those confirmed lows were undercut anyway within twelve months. A confirmed low is a signal, not a floor. That's the number to hold onto, and it's why the trap is worth naming.
Against simply owning the index it's a coin flip. 48.6% beat the S&P over the same dates, a median 1.5 points behind. Finding washed-out businesses that tend to recover isn't the same thing as beating the market, and we won't claim it is.
It swings hard by year. 2020 entries ran 12.6 points ahead of the index; 2019 entries ran 12.1 behind. Across 11 entry years, one blended number would hide that.
The median name also spent part of that year underwater by 14.4% at its worst point. You'd have had to sit through it.
Ten years is all the daily history we have, and it covers four real declines — late 2018, 2020, 2022 and 2025. The universe includes every ticker we've ever recorded, not just the ones still listed, but companies that disappeared before we started recording are still missing, which flatters any rebound number. This is descriptive history of a past sample, not a forecast. Computed 2026-08-30.
Read the full backtest — the undercut rate by year, what moved the number, whether waiting for confirmation helps, and what happens when it is traded with a stop — or download every event as a spreadsheet.
Is one of your holdings on the Broken List?
Paste your portfolio and we'll flag anything that's currently broken, and show you why.
