What is the fair value of AT&T stock?

The model's discounted-cash-flow fair value for T is $37.02 per share against a $24.87 close on October 8, 2026, a gap of +48.9%. It is an estimate with a range, not a price target.

What is the T stock forecast?

The published 21-session view for T is −1.33%, with a band from −12.3% to +9.6%; over 63 sessions it is −3.05%. These are probabilities from the platform's machine-learning models, re-estimated every trading day.

How accurate have the forecasts for T been?

50% of 257 matured 21-session forecasts on AT&T called the direction correctly. Every forecast is published before its outcome is known and scored when it matures.

Is this investment advice?

No. It is a dated model reading of one company, published for research. Nothing on this page is a recommendation to buy or sell any security.

Company analysis

AT&T (T) forecast and analysis

Communication Services · Diversified Telecommunication Services · Reading of October 8, 2026

Last close

$24.87

Fair value (DCF)

$37.02

Gap to fair value +48.9%

21-session view

−1.33%

band −12.3% to +9.6%

63-session view

−3.05%

Market regime

strong risk-off

The reading in brief

AT&T, part of the Diversified Telecommunication Services industry within Communication Services, shows a split reading as of October 8, 2026. The model's fair value of $37.02 sits well above the latest close of $24.87, a wide gap of +48.9%. The shorter-term published view points the other way: the 21-session expected return is −1.33%, with a band from −12.3% to +9.6% that spans both signs, and the probability of a rise is a low 15%. The 63-session view, at −3.05%, also leans down, so the two horizons agree with each other while diverging from the valuation. This comes in a strong risk-off market regime. The record here is long but not decisive, with 50% of 257 matured 21-session forecasts matching the direction.

Written automatically from the figures on this page; every figure is filled in from the day's data.

Research cockpit

Each source's reading of T today, the stance it takes against the published view, and what would change it. The same board subscribers open in the console.

against

+48.9%value gap · DCF against pricefair value $37 vs $24.87 · the DCF's own chance it is undervalued: 34%supports at −10% or less · against at +10% or more

mixed

+14unified technical score3-session run vs industry +3.32%, higher than 88% of sector; such runs have tended to fadesupports at −15 or less · against at +15 or more

mixed

6 / 9Piotroski F-scoreROIC 2% · D/E 1.14 · Q2 period to Jun 30, 2026supports at 3 or less · against at 7 or more

supports

15%recommended exposuremarket-wide strong risk off · the regime's tilt for Communication Services −0.18 · as of Oct 8, 2026supports at 40% or less · against at 75% or more

supports

11 analystscovering in the last 180 days · our own estimate recordhigher than 69% of sector · net rating changes +0.33 over 63 sessions · vendor consensus unavailablesupports in the most-covered 40% of its sector · against in the least-covered 40% — the most-covered fifth beat the median stock 49 times in 100 over three months, the least-covered 52 (since 2012)

Model reading

what the published belief says · what only a portfolio can add

leans down−1.33%2 of 5 sources for · 21d belief · 11 voices

Price impactmeasured · this symbol

higher than 59% of sector (Kyle λ)Names with higher price impact have done slightly worse over the next week (the ordering held in 27 of 32 years). Order size itself is a portfolio decision.

Liquidity vs edgemeasured · this symbol

more liquid than 88% of sector (Amihud)Less liquid names have earned more over three months (the ordering held in 25 of 32 years). The portfolio prices the order.

Band legs · q10 / q9021 sessions

−12.3% / +9.6%

Risk checksper portfolio

Run in the portfolio, not on a symbol — 13 checks (position ceiling, gross exposure, sector concentration …)

Regime overridemarket-wide · Oct 8, 2026

scaled15% exposure

Reliability21 sessions

0.000 = the 21-session head has not yet earned a reliability score on settled outcomes.

Price impact in fifths · the next week

Illiquidity in fifths · the next three months

Bars: out of 100 stocks in each fifth of all stocks, how many beat the median stock. The dashed line is 50, a coin flip; the axis is zoomed to 44–56. ▲ = where T sits today.

InsightProven signal

What it meansT's price impact over 5 sessions is higher than 59% of sector; by illiquidity it is more liquid than 88% of sector. Both are measured on T itself, not a portfolio.

What history saysOut of 100 stocks with the highest price impact, 49 beat the median stock over the next week; with the lowest, 50. T sits in the middle fifth today. That ordering held in 27 of 32 years since 1995. On average the lowest fifth out-earned the highest by 0.13 points over the next week. Over the next three months 52 in 100 of the least liquid fifth beat the median stock, against 48 of the most liquid. T sits in the lowest fifth for illiquidity.

How to use itThese describe what an order in T would face, not which way it goes. Order size stays a portfolio decision.

For the curious: Kyle λ at 5 sessions: rank IC −0.011 · t −6.3 · 27/32 years · 1,633 weekly samples since 1995 · Amihud at 63 sessions: rank IC +0.044 · t +3.9 · 25/32 years · 1,621 weekly samples since 1995

Published belief band

q10 — μ — q90 per horizon · conformal · rank among names

1d+0.0%−2.3% · +2.3%−2.3% · +2.3%p39 of 1,303

5d−0.2%−5.1% · +4.8%−5.1% · +4.8%p49 of 1,303

21d−1.3%−12.3% · +9.6%−12.3% · +9.6%p89 of 1,303

63d−3.1%−21.3% · +15.2%−21.3% · +15.2%p21 of 1,303

Square = expected return, bar = the model's q10–q90 range, right = where the expected return ranks among 1,303 names. A hollow square has no published range — do not size on it.

Exit map

Where T's 21-session band legs sit against today's close is in the console, with the valuation, technical, fundamentals and regime tabs.

Why it leans down, not up

posterior against a 51.7% base rate

P(up)0.15below a coin flip

E[r]−1.33%expected, 21 sessions

σ6.65%spread of outcomes

Rankp89of 1,303 names

E[r] / σ−0.20reward per unit of spread

Evidence balance · the belief's own arithmeticfor +0.06% · against −2.11% · net −2.05%

iv kinematicw 6%+0.06%

dcf mean reverterw 6%+0.01%

iv cqr probw 4%−0.01%

iv relativew 6%−0.04%

bundlew 19%−0.05%

sector catboostw 6%−0.05%

lstmw 3%−0.07%

iv cqr returnw 4%−0.10%

mimow 4%−0.26%

catboostw 5%−0.29%

quantilew 37%−1.25%

Each voice's expected return times its weight; the rows sum exactly to the belief.

Outside evidence · measured, not voting

lowest in sectorhighest

Run vs industry · 3 sessionshigher than 88% of sectorfavours the belief

Beat its factors · 21 sessionshigher than 59% of sectorsits in the middle

Unusual volume · vs its yearhigher than 88% of sectorleans against it

Price impact (Kyle λ)higher than 59% of sectorsits in the middle

Illiquidity (Amihud)higher than 12% of sectorfavours the belief

Short-sale volume · 21 sessionshigher than 37% of sectorleans against it

Analysts coveringhigher than 69% of sectorfavours the belief

Tinted half = the side of the sector that has gone on to do better on that measure, in the evidence study since 1995, 2009, 2012. These do not enter the belief's arithmetic — they tell you whether the market's own record agrees with it.

Conditional hit rate · how the prior updates

Unconditional, every name51.7% · n 342,931

Given a top-decile belief51.0% · n 34,163

… and 5+ voices behind it51.0% · n 34,163

base 51.7%

Platform-wide, not T's own record. The tick is the base rate; a top-decile belief has not beaten it. T sits in decile 9 today.

InsightModel output

What it meansThe belief leans down (−1.33%) and its own P(up) is 0.15: a negative average and a worse-than-even chance of rising.

What history saysThe expected return ranks higher than 89% of 1,303 names today. On settled outcomes a top-decile belief has been right 51.0% of the time against 51.7% for every name. Proven measures read outside the belief: 7 — favouring it, 3; leaning against it, 2. The strongest of them, Run vs industry · 3 sessions, held its ordering in 32 of 32 years since 1995.

How to use itRead the rank and the range before the number; the band legs, not the average, say how wide the outcomes run.

For the curious: direction_prob, mu, sigma from ii_belief · rank among 1,303 names · base rate from the signal-validation study, platform-wide, Sep 2, 2025 to Sep 9, 2026

What would change this reading

current reading measured · the trigger's source named

The fair-value gap falls below +10%+48.9% todaythe valuation rule on this tab

The 3-session run falls back to the sector middlehigher than 88% of sector todayproven reversal · the next week

Short-sale volume rises into the top half of the sectorhigher than 37% of sector · higher than 68% of its past yearproven · one to three months

Voice agreement drops11 voices · 73% consistentportfolio policy sets the level

The market regime turns more risk-offstrong risk off todayregime tab · market-wide

Voice agreement · 21d

5 scored voices · belief as of Oct 8, 2026

bundle32%right on direction · n 294latest −0.3%

catboost55%right on direction · n 295latest −6.2%

lstm57%right on direction · n 295latest −2.0%

mimo57%right on direction · n 295latest −6.3%

quantile48%right on direction · n 295latest −3.4%

How often each model called T's 21-session direction right on settled outcomes. Below 60 settled calls the figure is greyed: too few to judge.

Valuation & Predictions, Technical Structure, Fundamentals & Filings and Regime are in the console.

Forecast record on T

50% of 257 matured 21-session forecasts on AT&T called the direction correctly.

What happened ↑

Each square is one matured forecast: the view published that day against the return that followed.

Same industry

All Communication Services companies

Questions about T

What is the fair value of AT&T stock?: The model's discounted-cash-flow fair value for T is $37.02 per share against a $24.87 close on October 8, 2026, a gap of +48.9%. It is an estimate with a range, not a price target.
What is the T stock forecast?: The published 21-session view for T is −1.33%, with a band from −12.3% to +9.6%; over 63 sessions it is −3.05%. These are probabilities from the platform's machine-learning models, re-estimated every trading day.
How accurate have the forecasts for T been?: 50% of 257 matured 21-session forecasts on AT&T called the direction correctly. Every forecast is published before its outcome is known and scored when it matures.
Is this investment advice?: No. It is a dated model reading of one company, published for research. Nothing on this page is a recommendation to buy or sell any security.

A dated model reading, not investment advice. Forecasts are probabilities and can be wrong.

How the fair value is estimated How every forecast is tested Disclosures