Comcast CMCSA
Communication Services · Telecommunications Services · Synthos Deep Dive · 2026-08-04
The Overview
Comcast sells internet, mobile phone service and television to American households under the Xfinity name, and separately owns NBC, Sky in Britain, the Peacock streaming service, Universal's film studio and the theme parks.
It is taking itself apart. In January it spun off its cable television channels into a separate listed company called Versant. In June it said it would split the rest in two by the middle of 2027 — the media and theme-park businesses in one company, the broadband and mobile business in the other.
The reason is visible in the results. The internet business, which is the profitable core, is shrinking. Revenue from home broadband fell 5.5% last quarter. Comcast lost 167,000 broadband customers and 280,000 television customers in three months. Mobile carriers now sell home internet over their networks, and fibre companies are digging up the same streets, and Comcast's answer has been to cut prices and give new customers a free mobile line for a year — which keeps customers but reduces revenue.
The mobile business is the bright spot and it is genuinely bright: Comcast added 448,000 mobile lines last quarter, its best ever, and only 7% of its own broadband customers have bought one. There is a lot of room there.
The shares are down 23% over a year and about 60% from their peak. You now pay about seven times what analysts expect the company to earn next year, and collect a 5.3% dividend while you wait.
The thing to understand before anything else is the debt. Comcast owes $110 billion. The entire stock market values the company's shares at $88 billion. So the lenders have a bigger claim on this business than the owners do. That is not a crisis — the company throws off roughly $18 billion of spare cash a year and the debt is investment grade — but it means the share price moves about twice as much as the value of the underlying business, in both directions.
We think the shares are worth about $29 against $25. That is enough to keep on a list. It is not enough to buy a shrinking business with more debt than equity value while it is in the middle of splitting itself in two.
- Downside Risk 7/10. Debt exceeds market value, on a revenue base that is contracting.
- Growth Quality 3/10. Flat for four years; the largest line is falling 5.5%.
- Exponential Potential 3/10. Wireless convergence at 7% penetration is real. The rest is corporate engineering.
Putting a number on it: our fair-value estimate is $29 against a current price of $27.06 — real upside if our numbers are right.
Our summary metrics
"Rated 7, and one number carries most of it: total debt of $110,443 million against a market capitalisation of $88,468 million — the vendor computes debtToMarketCapTTM at 1.022, so the creditors have a larger claim on this enterprise than the shareholders do. Net debt is approximately $100,962 million against an enterprise value near $190 billion, which means the equity is roughly a 47% sliver of the capital structure and any change in the enterprise multiple is amplified by about 2.1x at the equity line. That is why the shares are 59.6% below their multi-year peak while revenue is flat. Against it: the debt is investment grade and long dated, interest expense of $4,409 million is covered several times by roughly $46 billion of EBITDA, and the business generates enormous cash — $4.6 billion of free cash flow in the June quarter on the company's own measure. The operating risk is the erosion: domestic broadband revenue fell 5.5% year on year in the quarter, residential broadband customers fell 167,000 in the quarter to 28,486 thousand, video customers fell 280,000, and residential connectivity relationships fell 230,000 to 47.7 million. The 10-Q's own framing is unusually candid — our customer relationship additions/(losses) continue to be negatively impacted by an increasingly competitive environment — and it discloses a deliberate revenue sacrifice: in 2025 we simplified our broadband pricing structure and began offering a free wireless line for one year to new and existing domestic broadband customers, which will negatively impact average domestic broadband revenue per customer. Add goodwill of $61,502 million and intangibles of $81,839 million — 52.5% of total assets — and two structural separations in eighteen months, one completed and one announced."
"Rated 3. Revenue: $103,564M (2020), $116,385M (2021), $121,427M (2022), $121,572M (2023), $123,731M (2024), $123,707M (2025) — FLAT for four years and marginally DOWN in the last. The June 2026 quarter delivered $29,940M against $30,313M, down 1.2%, and the March quarter $31,457M against $29,887M — but the comparison is contaminated because the Versant separation completed on 2026-01-02 and, per the 10-Q, did not meet the criteria to be presented as a discontinued operation, so the 2025 comparatives still contain Versant while 2026 does not. Underneath, the business divides cleanly. SHRINKING: domestic broadband revenue $6,280M against $6,649M, down 5.5%, with customers down 167,000 in the quarter; video customers down 280,000; wireline voice in structural decline. GROWING: domestic wireless line net additions of 448,000 — the best quarterly result on record — taking total lines to 10.2 million at 7% penetration of the addressable footprint; Business Services Connectivity revenue $10,237M in FY2025 against $9,701M, up 5.5%; Theme Parks $9,835M against $8,617M, up 14.1%; Peacock profitable for the first time. Consensus wants EPS of $3.501 in FY2026 (15 analysts), $3.618 in FY2027 (14) and $4.024 in FY2028 (10) — a 13.4% DECLINE from the FY2025 consensus of $4.041, then +3.3%, then +11.2% — with revenue of $120,750M, $118,653M and $122,193M. The FY2026 decline is the Versant spin-off leaving the perimeter. A 3: the largest revenue line in the company is shrinking, the growth lines are smaller, and the arithmetic is dominated by structural separations rather than by trading."
"Rated 3. Two things in this file could compound and neither is a technology curve. The first is wireless convergence: 448,000 net line additions in the June quarter — the best on record — to 10.2 million total lines at only 7% penetration of the addressable footprint inside Comcast's own broadband base. Selling a mobile line over a leased network to a household that already pays for broadband is close to pure incremental margin, and 93% of the addressable base has not bought it. That is real optionality and management is buying it deliberately, having begun offering a free wireless line for a year to new and existing broadband customers at an acknowledged cost to broadband revenue per customer. The second is the NBCUniversal and Sky separation announced in June 2026 for completion in mid-2027 — a sum-of-the-parts event rather than an operating one, in which a broadband-and-wireless connectivity company and a media-and-parks company are valued separately by different buyers, with Comcast retaining an interest for up to a year to monetise in a tax-efficient manner. Against both: the core asset is a fixed coaxial and fibre network competing against fixed wireless access and overbuilt fibre, and the 10-Q says plainly that customer losses continue to be negatively impacted by an increasingly competitive environment. Peacock reaching profitability for the first time is a milestone and a small one. A 3: genuine convergence optionality inside a shrinking core, plus a corporate event."
What does “fair value” mean?
Fair value is Synthos’s estimate of what one share is worth today, based on our model of the company’s future cash generation and the risks to it. It is not a price target or a prediction of where the stock trades next quarter — it’s the price where we think risk and reward are balanced. Above it, you’re paying for outcomes better than our base case; below it, the market is offering a margin of safety.
The model’s inputs and weightings are proprietary — they’re the product. What we publish is the output, plus an independent, public-math cross-check further down the page (“Check our number”) so you can judge it for yourself.
What do the 5 tiers mean? (Core · Tactical · Watch · Hold · Avoid)
The Road Ahead
What we expect to matter in each window, and the evidence that would prove us wrong.
Short term 0-6 months
Neutral- Driver
- "A stock 26.9% below its 52-week high and 59.6% below its multi-year peak, bouncing modestly off the low. Comcast closed 2026-08-04 at $24.93, up 1.51%, which is 4.9% above a 50-day moving average of $23.77 and 8.6% BELOW a 200-day average of $27.26. It is 26.9% below the 52-week high of $34.15 and 13.9% above the low of $21.92, with a maximum drawdown from peak of 59.6% — the second deepest in this batch after Intuit. RSI is 59.5 and MACD +0.15, so momentum is mildly positive and essentially flat. Twelve-month return is MINUS 23.2% against SPY's +24.3% — a 47-point deficit; three-month MINUS 7.8% and six-month MINUS 15.6%. The June-quarter print on 2026-07-23 beat consensus EPS by 7.2% and showed domestic broadband customer losses IMPROVING by 34,000 year on year alongside record wireless additions, and the shares have gone roughly nowhere since. Volume of 26.3 million shares is by far the heaviest in this batch. Nothing in the next six months resolves the two structural questions: whether broadband stabilises, and what the NBCUniversal separation is worth."
- What we’re watching
- "The 2026-10-29 print against consensus EPS of $1.01 and revenue of $29,532M. Within it, in strict order: DOMESTIC RESIDENTIAL BROADBAND CUSTOMER NET LOSSES, which were 167,000 in the June quarter against 201,000 a year earlier — an improvement of 34,000 that management calls traction with our new go-to-market strategy, and the single series that determines this equity; domestic broadband REVENUE, down 5.5% to $6,280M, where the deliberate price simplification and the free wireless line are working against the customer improvement; domestic wireless net additions against the record 448,000, and whether penetration moves off 7%; Theme Parks, where management flagged near-term softness; and any further detail on the NBCUniversal and Sky separation, particularly the financing arrangements the 10-Q names as a condition. Also watch the buyback pace: $2,507M in the first half against $7,155M for the whole of 2025."
- Confidence
- Medium
Medium term 6-24 months
Neutral- Driver
- "The medium term is dominated by a corporate event rather than by trading. Comcast intends to separate into two independent publicly traded companies through a tax-free spin-off of NBCUniversal and Sky by mid-2027 — the Media, Studios and Theme Parks segments plus the Sky businesses within Residential Connectivity & Platforms and Business Services — retaining an ownership interest for up to one year afterwards to monetise in a tax-efficient manner. On FY2025 segment revenue that separates roughly $48 billion of media, studios and parks from roughly $81 billion of connectivity, and it creates two capital structures where there is now one carrying $110,443 million of debt. Sum-of-the-parts arguments are the standard case for this kind of transaction and they are also the standard trap. Underneath, the operating arithmetic is legible: connectivity revenue is being defended with price and with a free wireless line, wireless is adding lines at a record rate off a 7% penetration base, business services grew 5.5% and theme parks 14.1% in FY2025, and Peacock has just turned profitable. Consensus wants EPS to fall 13.4% in FY2026 on the Versant departure, then grow 3.3% and 11.2%. At 6.9x FY2027 with a 5.29% dividend and roughly $18-19 billion of annual free cash flow, very little of that has to go right."
- What we’re watching
- "Whether broadband revenue stops declining. Customers improved by 34,000 year on year in the June quarter while REVENUE fell 5.5%, which is the deliberate trade management disclosed — simplified pricing and a free wireless line for a year — and it only works if the customers stay when the free year ends. Watch: domestic broadband revenue growth, not just customer counts; wireless penetration against 7%; the NBCUniversal separation's financing arrangements, tax opinions and regulatory approvals, all named in the 10-Q as conditions; how the $110,443 million of debt is allocated between the two companies, which is the single most important undisclosed term; whether the buyback continues at the first-half pace of $2,507M; and Theme Parks, where near-term softness was flagged after 14.1% FY2025 revenue growth."
- Confidence
- Low
Long term 2+ years
No differentiated view- Driver
- "Long-run the core asset is a fixed network selling a commoditising product into a market with two new entrant classes — fixed wireless access from the mobile carriers and overbuilt fibre — and the numbers already show it. Domestic broadband revenue fell 5.5% year on year; residential broadband customers fell to 28,486 thousand; video customers fell 280,000 in a quarter; residential connectivity relationships fell 230,000 to 47.7 million. The 10-Q describes the environment as increasingly competitive without qualification. Comcast's answer is convergence — bundle a mobile line over a leased network with the broadband connection, at 7% penetration of its own footprint and record additions — and that answer is arithmetically sound but it is defence rather than growth. The knowledge base's single surviving claim, from 2017, is about outsider chief executives using leverage at a disciplined band for its tax shield, citing Malone's 4x cash flow for cable; Comcast today carries debt exceeding its market capitalisation, so the band is the question and the answer determines whether shareholders or creditors own the residual. The NBCUniversal separation is a real event that changes who owns what; it does not change how many households buy internet from a coaxial network."
- What we’re watching
- "Whether the connectivity business can hold revenue flat. Everything else is second order. Beyond that: how the debt is split in the NBCUniversal separation and whether either successor is over-levered; whether wireless penetration reaches double digits, which would make convergence a real economic defence rather than a retention tactic; whether Peacock's first profitable period is sustained or was a slate effect; whether Theme Parks' 14.1% FY2025 growth survives the near-term softness management flagged; and the governance question the filings raise without explaining — the chief operating decision maker is now described as Co-Chief Executive Officers, while the vendor profile still names Brian L. Roberts alone, and the archive contains no succession disclosure."
- Confidence
- Low
Exponential Potential
"Rated 3. Two things in this file could compound and neither is a technology curve. The first is wireless convergence: 448,000 net line additions in the June quarter — the best on record — to 10.2 million total lines at only 7% penetration of the addressable footprint inside Comcast's own broadband base. Selling a mobile line over a leased network to a household that already pays for broadband is close to pure incremental margin, and 93% of the addressable base has not bought it. That is real optionality and management is buying it deliberately, having begun offering a free wireless line for a year to new and existing broadband customers at an acknowledged cost to broadband revenue per customer. The second is the NBCUniversal and Sky separation announced in June 2026 for completion in mid-2027 — a sum-of-the-parts event rather than an operating one, in which a broadband-and-wireless connectivity company and a media-and-parks company are valued separately by different buyers, with Comcast retaining an interest for up to a year to monetise in a tax-efficient manner. Against both: the core asset is a fixed coaxial and fibre network competing against fixed wireless access and overbuilt fibre, and the 10-Q says plainly that customer losses continue to be negatively impacted by an increasingly competitive environment. Peacock reaching profitability for the first time is a milestone and a small one. A 3: genuine convergence optionality inside a shrinking core, plus a corporate event."
What could take this further than the base case — and where the market may be underpricing it. Full forward-growth and acceleration math in Deeper analysis, §4 below.
Deeper analysis
Technicals, fundamentals, valuation, and the full expert-claim evidence panel — the detail behind the numbers above.
Reference table
| Street consensus | $28.20 (+13.1%) · median $27.50 (+10.3%) · high $36 (+44.4%) · low $23 (−7.7%) · 0 strong buy / 34 buy / 25 hold / 2 sell across 61 analysts — the widest coverage in this batch · vendor composite rating A− |
| Valuation | 7.1x FY2026E ($3.501) · 6.9x FY2027E ($3.618) · 6.2x FY2028E ($4.024) · 4.6x trailing GAAP (distorted, see Section 4) · EV/EBITDA 5.09x · P/B 0.91x · priceToFairValueTTM 0.9914 — the vendor's own model says fair |
| THE CAPITAL STRUCTURE | Total debt $110,443M EXCEEDS the $88,468M market capitalisation. debtToMarketCapTTM 1.022. Net debt ≈$100,962M on an enterprise value near $190B — the equity is roughly 47% of the capital structure and any enterprise re-rating is amplified about 2.1x at the equity line |
| TWO SEPARATIONS, neither in the payload | (1) The Versant Media Group spin-off COMPLETED 2026-01-02 — select cable television networks and digital platforms, tax-free, "did not meet the criteria to be presented as a discontinued operation", so 2025 comparatives still contain it. (2) In June 2026 Comcast announced its intention to separate into TWO independent publicly traded companies through a tax-free spin-off of NBCUniversal and Sky, targeted for mid-2027, retaining an interest for up to a year to monetise |
| The June quarter — the shrinking half | Domestic broadband revenue $6,280M vs $6,649M, −5.5% · residential broadband customers 28,486k vs 28,989k, 167,000 lost in the quarter · video customers −280,000 · residential connectivity relationships −230,000 to 47.7 million |
| The June quarter — the growing half | Domestic wireless net additions 448,000 — "Our Best Quarterly Result on Record" · total wireless lines 10.2 million at 7% penetration of the addressable footprint · Peacock profitable for the first time · Media EBITDA up mid-single digits · free cash flow $4.6 billion in the quarter |
| Capital return | Dividend $1.32, 5.29% yield · buyback $2,507M in H1 2026 (against $7,155M for all of FY2025) · $2.1 billion returned in the June quarter per the release · share count 4,654M (2021) → 3,570M (Q2 2026), −23.3% |
| Data findings | seg_prod and seg_geo BOTH tie EXACTLY to the dollar — the only name in this batch where both are perfect. Against that: capitalExpenditure omits "cash paid for intangible assets" of $1,226M in the half, overstating free cash flow by roughly $2.5 billion a year; and the segment realignment of Q1 2026 — five separate reclassifications — is invisible in the payload |
| Technicals | +4.9% above the 50-DMA ($23.77) but −8.6% BELOW the 200-DMA ($27.26); −26.9% from the 52-week high of $34.15, +13.9% above the low of $21.92; maximum drawdown from peak −59.6%; RSI 59.5; MACD +0.15; 12-month −23.2% vs SPY +24.3% |
What the experts actually said
No independent expert claims in the Synthos knowledge base yet for CMCSA — this dive is fundamentals- and technicals-driven, not panel-driven.
Price & moving averages 12 months · 50 & 200-day averages · 52-week range
Solid line = price · dashed line = 50-day average · dotted line = 200-day average · the two thin horizontal lines mark the 52-week high and low. Price above both averages is an uptrend.
Data summary: last close $27.06, 11% above the 50-day average ($24), 0% above the 200-day average ($27) — an uptrend. 21% below the 52-week high of $34, 23% above the 52-week low of $22.
Bollinger Bands 20-day average ± 2 standard deviations
The shaded band widens when the stock gets more volatile. Riding the upper edge = strong momentum (sometimes stretched); the lower edge = weak / potentially oversold.
Data summary: price $27.06 is currently inside the band (band $24–$28).
RSI (14) momentum gauge · 0–100
Above 70 (overbought zone, shaded) = overbought, below 30 (oversold zone, shaded) = oversold. Currently 63.
MACD 12 / 26 / 9 · trend & momentum
The MACD line crossing above the signal line (bars flip to the up color) = momentum turning up; crossing below (bars flip to the down color) = turning down. Bar height = the size of that gap.
Data summary: MACD is currently above its signal line by 0.04, positive momentum.
Relative performance vs S&P 500 & its sector (XLC (sector)), set to 100 a year ago
Solid = CMCSA · dashed = S&P 500 · dotted = XLC (sector). A rising line means it is beating that benchmark — the sector line shows whether it is a leader or laggard within its own group.
Forward revenue & earnings actual → estimate · "FY" = fiscal year, "E" = estimate
Darker bars = actual results, brighter = analyst estimates. Taller bars to the right = expected growth.
Key stats an RIA wants
1. Two separations in eighteen months, and the payload knows about neither
Versant — completed 2026-01-02
From the 10-Q filed 2026-07-23, verbatim:
> "On January 2, 2026 (the 'Versant Separation Date'), we completed the previously announced separation of Versant Media Group, Inc. ('Versant') into an independent publicly traded company comprised of select cable television networks and complementary digital platforms through a tax-free spin-off (the 'Versant Separation'). The Versant Separation did not meet the criteria to be presented as a discontinued operation, and Versant's results are included in our results of operations for the three and six months ended June 30, 2025."
That last sentence matters for every year-on-year comparison in this file: the 2025 periods contain Versant and the 2026 periods do not. The cash flow statement records cash transferred to Versant, net of $(750) million and the equity statement an $(8,107) million Versant Separation entry.
NBCUniversal and Sky — announced June 2026, targeted mid-2027
> "In June 2026, we announced our intention to separate into two independent publicly traded companies through a tax-free spin-off of NBCUniversal and Sky (the 'NBCUniversal Spin-off') comprised primarily of the Media, Studios and Theme Parks segments and the Sky businesses within our Residential Connectivity & Platforms and Business Services segments. The NBCUniversal Spin-off is expected to be completed in mid-2027, subject to the satisfaction of customary conditions, including obtaining final approval from our Board of Directors, receipt of tax opinions and regulatory approvals, and completion of financing arrangements. We expect to retain an ownership interest of up to [a stated percentage] in the NBCUniversal Spin-off entities for up to one year after the completion of the spin-off, which we intend to monetize in a tax-efficient manner over time. There can be no assurance that a separation transaction will occur... The condensed consolidated financial statements and related notes do not reflect the proposed NBCUniversal Spin-off."
On FY2025 segment revenue this separates approximately $48 billion of Media, Studios and Theme Parks from approximately $81 billion of connectivity — and it creates two capital structures where one now carries $110,443 million of debt. How that debt is allocated is the single most important undisclosed term in this file and it is named in the filing as a condition ("completion of financing arrangements").
And a segment realignment nobody outside the filings can see
> "Beginning in the first quarter of 2026, we updated the composition of our segments to align with the segment-level information that is regularly provided to our Co-Chief Executive Officers, who are the chief operating decision maker, including (1) adjusting the Media segment to exclude the historical results of Versant; (2) reclassifying the results of our regional sports networks to Corporate and other from the Media segment; (3) reclassifying the results of Xumo, our streaming platform joint venture with Charter Communications, to the Residential Connectivity & Platforms segment from Corporate and other; (4) reclassifying certain shared expenses into the related Media, Studios and Theme Parks segments...; and (5) adjusting the Media segment and Versant for the effects of the commercial services agreement."
Five separate reclassifications, none of them in the vendor payload, and a governance disclosure delivered in passing: the chief operating decision maker is now "our Co-Chief Executive Officers", while the vendor profile still names Brian L. Roberts alone. The archive contains no succession disclosure and this dive does not speculate about one; the fact is recorded because it is a material governance change stated in a filed document and absent from every commercial data field.
A further reclassification affects the most important revenue line in the company: "Beginning in the first quarter of 2026, commission revenue from the sale of certain direct to consumer ('DTC') streaming services is presented in domestic broadband revenue or video revenue based on whether a customer is entitled to receive the DTC streaming service through a broadband or video service offering... Previously, all of these amounts were in video revenue. Prior periods have been reclassified to reflect the current year presentation." So the −5.5% domestic broadband revenue figure is on a restated basis and is like-for-like; recorded so the reader knows the comparison was checked.
2. The June quarter — two businesses moving in opposite directions
All figures from the 8-K of 2026-07-23 (the complete earnings release, tables preserved) and the 10-Q of the same date.
The shrinking half:
| Q2 2026 | Q2 2025 | change | |
|---|---|---|---|
| Domestic broadband revenue | $6,280M | $6,649M | −5.5% |
| Total domestic broadband residential customers | 28,486k | 28,989k | −503k |
| Domestic residential broadband net losses in quarter | 167,000 | 201,000 | improved by 34,000 |
| Domestic video customer net losses | 280,000 | — | — |
| Residential Connectivity & Platforms customer relationships | 47.7 million | — | −230,000 |
The growing half:
| Q2 2026 | |
|---|---|
| Domestic wireless line net additions | 448,000 — "Our Best Quarterly Result on Record" |
| Total wireless lines | 10.2 million |
| Wireless penetration of addressable footprint | 7% |
| Peacock | Profitable for the first time |
| Media EBITDA | Mid-single-digit growth |
| Theme Parks | "near-term softness" |
| Free cash flow (company measure) | $4.6 billion |
| Returned to shareholders | $2.1 billion |
Management's own framing, quoted because it is unusually specific and is half-weighted:
> "In Connectivity & Platforms, our strategic pivot in broadband is gaining traction... We delivered our best wireless quarter ever, surpassing 10 million total lines, while penetration remains below 7% of addressable wireless lines in our footprint - giving us substantial runway to deepen convergence... Within Content & Experiences, Media delivered mid-single digit EBITDA growth and Peacock reached profitability for the first time... While we are seeing some near-term softness in Theme Parks, we remain confident in the long-term opportunity... Across the company, we generated $4.6 billion of free cash flow, returned $2.1 billion to shareholders and announced our intention to separate NBCUniversal and Sky."
And the 10-Q's own account of the trade being made, which is the most important disclosure in this file:
> "Our customer relationship additions/(losses) continue to be negatively impacted by an increasingly competitive environment. We are focused on increasing our residential connectivity revenue. In 2025, we simplified our broadband pricing structure and began offering a free wireless line for one year to new and existing domestic broadband customers, which we expect will improve customer retention and strengthen our ability to compete for new customers, but will negatively impact average domestic broadband revenue per customer."
Read those two together and the quarter is exactly legible. Comcast is buying customer retention with price and with a free wireless line, and the bill is arriving in the revenue line: customers improved by 34,000 year on year while REVENUE fell 5.5%. The strategy works if the customers stay when the free year ends and if the wireless attachment is durable. The archive cannot yet say.
The quarterly series, and note the Versant contamination:
Quarter (inc_q) | Revenue | YoY | Operating income | GAAP diluted EPS | Reported EPS | vs est |
|---|---|---|---|---|---|---|
| Q2 2024 | $29,688M | — | $6,635M | $1.00 | — | — |
| Q4 2024 | $31,914M | — | $4,994M | $1.24 | — | — |
| Q1 2025 | $29,887M | — | $5,659M | $0.89 | — | — |
| Q2 2025 | $30,313M | +2.1% | $5,993M | $2.97 (one-off) | $1.25 | vs $1.16 |
| Q3 2025 | $31,198M | −2.7% | $5,533M | $0.90 | $1.12 | vs $1.03 |
| Q4 2025 | $32,310M | +1.2% | $3,486M | $0.60 | $0.84 | vs $0.729 |
| Q1 2026 | $31,457M | +5.3% | $4,135M | $0.60 | $0.79 | vs $0.725 |
| Q2 2026 | $29,940M | −1.2% | $5,160M | $0.99 | $1.04 | vs $0.97 |
Five consecutive beats: +7.8%, +8.7%, +15.2%, +9.0%, +7.2%. Note that Q2 2025 GAAP diluted EPS of $2.97 against reported $1.25 reflects a large one-off gain, and FY2025 GAAP net income of $19,998M against pre-tax income of $25,766M and a $6,106M tax charge is likewise distorted; every trailing GAAP multiple on this name — including the 4.6x priceToEarningsRatioTTM-equivalent — is unusable and none is used.
3. The capital structure, and why it governs everything
| ($M) | 2025-12-31 |
|---|---|
| Cash and equivalents | $9,481 |
| Total current assets | $29,567 |
| Property, plant & equipment, net | $70,967 |
| Goodwill | $61,502 |
| Intangible assets | $81,839 |
| Total assets | $272,631 |
| Short-term debt | $6,644 |
| Long-term debt | $98,389 |
| Capital lease obligations | $5,410 |
| TOTAL DEBT | $110,443 |
| NET DEBT | ≈$100,962 |
| Total liabilities | $175,254 |
| Total stockholders' equity | $96,904 |
| MARKET CAPITALISATION | $88,468 |
debtToMarketCapTTM | 1.022 |
| EV/EBITDA (vendor) | 5.09x |
| Interest expense | $4,409M |
The single most important sentence in this dive: total debt of $110,443 million exceeds the market capitalisation of $88,468 million. Enterprise value is roughly $190 billion, of which the equity is about 47%. A 10% fall in the enterprise value is therefore roughly a 21% fall in the share price, and a 10% rise is roughly a 21% rise. That is why a company whose revenue has been flat for four years has a share price 59.6% below its multi-year peak, and it is the whole explanation for both the risk score and the width of the fair-value range below.
It is not a distress situation. Interest expense of $4,409M against roughly $46 billion of EBITDA is covered more than ten times; the debt is investment grade and long dated; and the company generated $4.6 billion of free cash flow in a single quarter. It is a leverage situation, and the knowledge base's one surviving claim — nine years old and about somebody else — is precisely about that: outsider chief executives running cable at a disciplined leverage band for its tax shield. Comcast is at the top of any such band.
Goodwill of $61,502M plus intangibles of $81,839M is $143,341M — 52.6% of total assets and 148% of equity. Book value per share is roughly $27.32 and the shares trade at 0.91x it, which for a network business is neither cheap nor dear and is not used as an anchor.
Cash flow — the arithmetic is right and the capex line is incomplete
| Year | OCF | Vendor capex | Vendor FCF | Buyback | Dividend |
|---|---|---|---|---|---|
| FY2022 | $26,413M | −$13,767M | $12,646M | −$13,328M | −$4,741M |
| FY2023 | $28,501M | −$15,540M | $12,961M | −$11,291M | −$4,766M |
| FY2024 | $27,673M | −$12,297M | $15,376M | −$9,103M | −$4,814M |
| FY2025 | $33,643M | −$11,750M | $21,893M | −$7,155M | −$4,894M |
Free cash flow is computed by subtraction and the arithmetic is correct in all four years. capexToDepreciationTTM of 0.766 and capexToRevenueTTM of 9.67% are both plausible for a cable network. But the capex line is incomplete, and the 10-Q shows exactly how:
> Capital expenditures | ( 5,253 ) | ( 4,930 )
> Cash paid for intangible assets | ( 1,226 ) | ( 1,257 )
Comcast spends on capitalised software and content rights in a separate cash-flow line, running approximately $2.5 billion a year, and the vendor's capitalExpenditure field captures only the first line. Corrected FY2025 free cash flow is therefore approximately $19.4 billion rather than $21,893M — an overstatement of roughly 13% — and the freeCashFlowYieldTTM of 23.11% corrects to approximately 20.0%.
Cross-check against the company's own figure: management reported $4.6 billion of free cash flow in the June quarter, which annualises to roughly $18.4 billion, a 20.8% yield on the current market capitalisation. So the corrected figure is right and the headline is 13-15% too high.
And the more important framing: a 20% free-cash-flow yield on the EQUITY is a 9.7% yield on the ENTERPRISE ($18.4 billion against roughly $190 billion), because the debt is larger than the equity. The equity yield is spectacular and it is spectacular because of leverage, not because of the assets.
Capital return, at the current pace:
| Amount | Against $88.47B market cap | |
|---|---|---|
| Buyback, H1 2026 (10-Q) | $2,507M | 5.7% annualised |
| Buyback, FY2025 | $7,155M | 8.1% |
| Dividend | $1.32 annualised | 5.29% |
| Combined shareholder yield at the H1 2026 pace | — | ≈11.0% |
| Share count | 4,654M (2021) → 3,570M (Q2 2026) | −23.3% in five years |
Note the buyback is DECELERATING: $13,328M (2022) → $11,291M (2023) → $9,103M (2024) → $7,155M (2025) → $2,507M in the first half of 2026, an annualised $5.0 billion. That is a 62% reduction in four years and it is the clearest signal in the file that management is prioritising the balance sheet and the separation over share repurchase.
4. Segments — both vendor blocks are perfect, and they show the shape
seg_prod FY2025 ties EXACTLY, INCLUDING the elimination line:
| FY2025 segment | Revenue | share | FY2024 | growth |
|---|---|---|---|---|
| Residential Connectivity & Platforms | $70,704M | 57.2% | $71,574M | −1.2% |
| Media | $27,090M | 21.9% | $28,148M | −3.8% |
| Studios | $11,286M | 9.1% | $11,092M | +1.7% |
| Business Services Connectivity | $10,237M | 8.3% | $9,701M | +5.5% |
| Theme Parks | $9,835M | 8.0% | $8,617M | +14.1% |
| Corporate and Other | $3,090M | 2.5% | $2,982M | +3.6% |
| Intersegment Eliminations | −$8,535M | −6.9% | −$8,383M | — |
| Sum | $123,707M | 100.0% | $123,731M | |
inc_a revenue | $123,707M | $123,731M | ||
| Tie | EXACT | EXACT |
seg_geo FY2025 also ties EXACTLY: United States $95,143M (76.9%) + Europe $15,182M (12.3%) + Other Geographic Locations $13,382M (10.8%) = $123,707M.
Both blocks perfect, including the intersegment elimination line — which no other name in this batch carried. Recorded as the cleanest segment reporting in the twelve. (Note that these are the pre-realignment FY2025 figures; the Q1 2026 restatement described in Section 1 changes the composition going forward and the payload does not reflect it.)
What the shape shows: the two largest segments, 79.1% of revenue between them, both DECLINED in FY2025 — Residential Connectivity & Platforms −1.2% and Media −3.8%. The three that grew are the three smallest. And the NBCUniversal separation removes Media, Studios and Theme Parks — $48,211M, 39.0% of revenue — plus the Sky businesses inside the two connectivity segments.
5. Valuation — priced in or room?
At $24.93 (market cap $88.47B, ~3,548M shares implied; 3,570M diluted weighted-average in the June quarter):
| FY2026E | FY2027E | FY2028E | FY2029E | |
|---|---|---|---|---|
| Consensus EPS | $3.501 (15) | $3.618 (14) | $4.024 (10) | $4.338 (4) |
| EPS range | $3.382–$3.600 | $3.373–$4.148 | $3.585–$4.481 | — |
| EPS growth | −13.4% (vs FY2025E $4.041) | +3.3% | +11.2% | +7.8% |
| P/E | 7.1x | 6.9x | 6.2x | 5.7x |
| Consensus revenue | $120,750M (18) | $118,653M (17) | $122,193M (15) | $121,341M (11) |
| Revenue growth | −2.4% | −1.7% | +3.0% | −0.7% |
| EV/EBITDA | 5.09x | — | — | — |
| Free-cash-flow yield | ≈20.0% corrected (23.11% reported) | — | — | — |
| Dividend yield | 5.29% | — | — | — |
| Shareholder yield | ≈11.0% | — | — | — |
Estimate coverage is good through FY2028 — 15, 14 and 10 analysts on EPS, 15-18 on revenue — and thins to 4 at FY2029. The FY2029 and FY2030 rows are excluded from every conclusion. The FY2027 EPS range is $3.373 to $4.148, a 23.0% spread.
The shape of the consensus deserves attention: revenue DECLINES in FY2026 and FY2027 and EPS falls 13.4% in FY2026. The FY2026 fall is principally the Versant spin-off leaving the perimeter rather than an operating deterioration, and this dive says so; but consensus revenue does not recover to the FY2025 level within the forecast horizon.
est.ebitdaAvg and est.ebitAvg are REJECTED — fixed-ratio fabrication. ebitdaAvg is exactly 45.69% of revenueAvg and ebitAvg exactly 31.58% in every year FY2026 through FY2030. Tested against the actual: FY2025 EBITDA per inc_a was $46,386M on $123,707M — 37.50% — against the block's 45.69%, 22% too high. All forward valuation uses epsAvg. Twenty-third consecutive name.
Peer set contains two genuine comparables and one absurdity. AMX, CHTR, DASH, LBRDA, RCI, RELX, T, VZ. Charter, AT&T and Verizon are the relevant comparisons and Rogers is adjacent; DoorDash is in this same research batch and has nothing to do with cable. The payload carries no peer estimates, so no multiple comparison is computed. Note that Charter's market capitalisation is shown at $20.6bn — its own collapse is part of the same sector story.
5a. What today's price assumes (the inversion)
At $24.93 — 7.1x FY2026 consensus, 6.9x FY2027, a 5.29% dividend yield, EV/EBITDA 5.09x — the price embeds:
- EPS of $3.50 in 2026 and $3.62 in 2027, on revenue that DECLINES in both years. (Consensus; 15 and 14 analysts.) The first half delivered $1.83 ($0.79 + $1.04), so FY2026 needs $1.67 in the second half against $1.96 a year earlier — consensus expects a second-half decline.
- Domestic broadband revenue keeps falling but the customer base stabilises. (Our derivation from the disclosures.) This is the most fragile assumption in the price. Revenue fell 5.5% while customer losses improved by 34,000, because management is deliberately trading price for retention. The trade only works if churn falls durably; if revenue keeps falling 5% and customers resume leaving, the connectivity business is a melting asset and 6.9x is not cheap.
- The NBCUniversal separation completes in mid-2027 on terms that do not impair the remaining connectivity company. (Company disclosure, half-weighted; the 10-Q names Board approval, tax opinions, regulatory approvals and completion of financing arrangements as conditions, and states there can be no assurance a separation will occur.) How $110,443 million of debt is split is undisclosed and is the largest unknown in this file.
- The enterprise multiple holds at roughly 5x EBITDA. (Our number.) At 4.5x the equity is worth roughly $19; at 5.5x roughly $34. That 2.1x amplification from the leverage is the entire distribution and it is why the fair-value range below is $16 to $40 on a business whose revenue moves 2% a year.
- The dividend is maintained at $1.32. (Our derivation.) The payout is roughly 38% of FY2026 consensus EPS and comfortably covered by free cash flow. It is the least fragile assumption here and it is 5.29% of the return.
5b. The return bridge (why the multiple moves)
Expected return over the next twelve months decomposes as: EPS growth (+3.3%, from FY2026E $3.501 to FY2027E $3.618) + multiple drift (EXPANSION; our base of $29 is 8.0x FY2027E against 6.9x today, +16.3%) + shareholder yield (+11.0%: a 5.29% dividend and roughly 5.7% of buyback at the first-half pace).
Our base requires the multiple to expand from 6.9x to 8.0x and that is emphatically the fragile leg — there is almost no earnings growth here to carry the return. The argument for expansion: 6.9x forward with an 11% shareholder yield is a distressed multiple for an investment-grade business generating $18-19 billion of annual free cash flow, and the mid-2027 separation is a specific event that typically closes conglomerate discounts. The argument against: revenue declines in both forecast years, the largest segment is shrinking, and a 2.1x-levered equity does not deserve a re-rating until the operating trend turns.
If the multiple simply holds, the return is roughly 14% — 3.3% of earnings growth plus 11.0% of yield — and fair value is roughly $25, which is spot. That is the honest floor, and an 11% yield with no growth and no re-rating is not a bad outcome; it is simply not a thesis.
5c. Variant perception (where we differ, what would surprise)
- On the level we are 2.8% above the street's $28.20 and 16.3% above spot — no meaningful variant perception. What is worth naming is the coverage: 61 analysts, the widest in this batch, with 34 buys against 25 holds and 2 sells and a $28.20 consensus, on a stock that has fallen 23.2% in twelve months. That is a covering community that broadly agrees with us and has been wrong for a year.
- Where we differ from the general framing is on which number matters, and we think it is broadband REVENUE rather than broadband SUBSCRIBERS. The June quarter's headline improvement — customer net losses improving by 34,000 — was accompanied by a 5.5% revenue decline, because the retention was bought with simplified pricing and a free wireless line, which the 10-Q discloses explicitly. Watchable number: domestic broadband revenue growth in the 2026-10-29 print. A quarter with customers improving AND revenue decline narrowing below 3% would be genuine evidence the pivot is working; another 5%+ revenue decline with improving subscribers means the company is buying customers at an accelerating discount.
- We think the leverage is under-discussed relative to its importance and we quantify it. Total debt of $110,443 million against an $88,468 million market capitalisation means the equity is roughly 47% of a $190 billion enterprise, so the share price moves about 2.1x the enterprise value. Every valuation argument about this name — bull or bear — should be made at the enterprise level and then geared, and almost none is.
- Positive surprise that would force a re-rate: domestic broadband revenue decline narrowing below 3% with continued customer improvement; wireless penetration moving above 8% with additions holding near 448,000; or clear disclosure of the debt allocation in the NBCUniversal separation showing the connectivity company de-levered below 3x.
- Negative surprise that would break the thesis: broadband customer losses re-accelerating above 250,000 in a quarter; broadband revenue declining more than 6%; the NBCUniversal separation delayed, abandoned or structured with the debt loaded onto the connectivity company; the dividend or the buyback cut; or a Theme Parks deterioration beyond the "near-term softness" flagged, after 14.1% FY2025 growth.
Synthos fair values
These anchors are built at the ENTERPRISE level and then geared to the equity, because with debt exceeding market capitalisation an equity-multiple approach understates both tails. Cross-checked against the FY2027 consensus EPS distribution (mean $3.618, low $3.373, high $4.148, 14 analysts).
- Bear ~$16 — an enterprise value of roughly $157 billion (about 4.5x EBITDA on a 10% lower EBITDA base) less $101 billion of net debt. Cross-check: 4.4x the FY2027 consensus mean; 27.0% below the 52-week low of $21.92; an 8.25% dividend yield if maintained. The scenario: broadband customer losses re-accelerate, revenue declines steepen, the separation is delayed or badly structured, and a levered stub on a melting asset gets a melting-asset multiple. −35.8%.
- Base ~$29 — an enterprise value of roughly $190 billion (5.1x EBITDA, unchanged) with EBITDA flat, plus roughly two years of the current buyback reducing the share count. 8.0x the FY2027 consensus mean. Cross-check: 8.3x FY2026E; 7.2x FY2028E; 15.1% below the 52-week high; a 4.55% dividend yield; 2.4% above the street's $28.20. The scenario: broadband revenue decline moderates, wireless keeps adding at record rates, the separation completes on neutral terms, the dividend holds, and the market grants a modest re-rating for a completed simplification. +16.3%.
- Bull ~$40 — an enterprise value of roughly $242 billion (5.75x EBITDA on a 10% higher base) less net debt, i.e. 11.1x the FY2027 consensus mean. Cross-check: 9.9x FY2028E; 17.1% above the 52-week high of $34.15. The scenario: broadband revenue stabilises, wireless penetration doubles toward 14%, the NBCUniversal separation is valued as two businesses rather than one conglomerate, and the connectivity company re-rates toward a utility multiple with an 11% shareholder yield behind it. +60.4%.
Base is 16.3% above spot; asymmetry 1.69:1 (35.8% down, 60.4% up), plus an ~11.0% shareholder yield. The wide range is not imprecision — it is the arithmetic of a 2.1x-levered equity on a business whose enterprise value moves slowly. Watch.
6. Knowledge base — one nine-year-old claim, and by accident it is the right one
Raw hits: 58. After a case-sensitive entity re-run on the company terms: 1. Used: 1. Discarded: 57.
The sweep ran entity terms CMCSA, Comcast, NBCUniversal, Xfinity, Peacock, Versant and Sky, plus free text on Comcast, Xfinity, Peacock, NBCUniversal, broadband and cable.
The single surviving claim, verbatim:
> 2017-06-01 · bullish · conviction 75 · horizon: principle · entities: TCI, Liberty Media, Charter, Comcast · channel: invest_like_the_best · skill 1.0
> "Outsider CEOs actively used leverage for its tax shield at a clear disciplined band (Malone's 4x cash flow for cable), running slightly hotter than peers while communicating the band transparently."
It is nine years old. It is a claim about capital-allocation philosophy in the cable industry, not about Comcast's business. And it is, by accident, exactly the right question to ask about this company today, because Comcast carries $110,443 million of total debt against an $88,468 million market capitalisation — roughly 2.2x the $46 billion of FY2025 EBITDA on the vendor's measure, and considerably more on a normalised one. The claim's "clear disciplined band" and "communicating the band transparently" are precisely what is missing from the NBCUniversal separation disclosure, which names "completion of financing arrangements" as a condition and says nothing about how the debt divides. Used as framing. It contributes nothing to the conviction score.
The discards divide into two families and one is a genuine homograph worth recording.
THE SKY COLLISION. forward_guidance, 2025-07-02, bullish 70, entities Athena, Sky, Centrifuge: "Next RWA wave: stablecoin issuers add the credit universe between T-bills (4%) and illiquid private credit (10%) into their backing as rates fall." "Sky" here is a real-world-asset crypto protocol, not the British broadcaster Comcast owns. A Sky-inclusive entity sweep would have imported a bullish stablecoin-credit claim into a cable dive. Note that a second "Sky" hit — business_breakdowns, 2023-05-09, on streamers bidding for Formula 1 rights, entities Formula 1 / Amazon / ESPN / Sky — IS the broadcaster, but it is a claim about Formula 1 rights competition and names Sky only as a bidder; it is discarded on content rather than on identity.
The other 55 are free-text noise on "broadband" and "cable" across 30 channels, dominated by artificial-intelligence infrastructure discussion (jensen_huang, sundar_pichai_mgmt, dwarkesh), rural-broadband policy (all_in, 2025-10-30, on the private sector delivering infrastructure more cheaply), subsea cables, and streaming-share commentary that names Netflix rather than Comcast (business_breakdowns, 2023-05-26).
Conclusion. Breadth 1, claim count 1, net conviction NONE. The knowledge base has no view on Comcast and this dive does not manufacture one. The one claim it does hold is a nine-year-old principle about cable leverage, and it is used as framing precisely because leverage is the governing variable here.
7. Data integrity — what we rejected and why
Six findings, and the balance is unusually favourable.
1. seg_prod AND seg_geo BOTH TIE EXACTLY TO THE DOLLAR — the only name in this batch where both are perfect. seg_prod FY2025 sums to $123,707M against inc_a revenue of $123,707M, including the −$8,535M intersegment elimination line that no other name in this batch carried; seg_geo sums to the same figure across three correctly labelled regions. The same holds for FY2024. Recorded as the cleanest segment reporting in the twelve, against four broken geography blocks elsewhere in this batch (UPS 72.8%, MRSH 162.8%, WM with the US line absent, MAR with the US line absent) and an entirely empty one (INTU).
2. capitalExpenditure OMITS "cash paid for intangible assets" — free cash flow overstated by roughly 13%. The 10-Q's cash flow statement reports capital expenditures of $(5,253)M and cash paid for intangible assets of $(1,226)M for the six months to 2026-06-30 (against $(4,930)M and $(1,257)M a year earlier). The vendor's capex line captures only the first. Corrected FY2025 free cash flow is approximately $19.4 billion rather than $21,893M, and freeCashFlowYieldTTM corrects from 23.11% to approximately 20.0% — cross-checked against management's own $4.6 billion for the June quarter, which annualises to roughly $18.4 billion or 20.8%. The CSX/COP defect class, direction as usual, magnitude roughly 13%. And the more important framing is stated in Section 3: a 20% free-cash-flow yield on the equity is a 9.7% yield on the enterprise, because the debt exceeds the equity.
3. TWO STRUCTURAL SEPARATIONS AND A FIVE-PART SEGMENT REALIGNMENT ARE ABSENT FROM THE PAYLOAD. (a) The Versant Media Group spin-off completed 2026-01-02 — not presented as a discontinued operation, so all 2025 comparatives contain it, with $(750)M of cash transferred and an $(8,107)M equity entry. (b) The intention announced in June 2026 to separate into TWO independent publicly traded companies through a tax-free spin-off of NBCUniversal and Sky by mid-2027, with Comcast retaining an interest for up to a year to monetise, subject to Board approval, tax opinions, regulatory approvals and completion of financing arrangements — and the financial statements explicitly do not reflect it. (c) A five-part segment realignment effective Q1 2026, plus a separate reclassification moving direct-to-consumer streaming commission revenue between the video and broadband lines with prior periods restated. None appears in any vendor field.
4. A GOVERNANCE FACT IS DISCLOSED IN PASSING AND CONTRADICTS THE PAYLOAD. The 10-Q describes the chief operating decision maker as "our Co-Chief Executive Officers"; the vendor profile reports a single chief executive, Brian L. Roberts. The archive contains no separate succession or appointment disclosure and this dive does not speculate about one — the fact is recorded because it is a material governance change stated in a filed document and absent from every commercial data field.
5. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature — REJECTED. ebitdaAvg is exactly 45.69% of revenueAvg and ebitAvg exactly 31.58% in every year FY2026 through FY2030. Tested against the actual: FY2025 EBITDA of $46,386M on $123,707M is 37.50% — the block overstates by 22%. All forward valuation uses epsAvg; the FY2029 (4 analysts) and FY2030 (4) rows are excluded.
6. Trailing GAAP earnings are unusable and no trailing multiple is published. FY2025 netIncome of $19,998M against incomeBeforeTax of $25,766M and a $6,106M tax charge, with Q2 2025 alone at $11,123M of net income on $5,993M of operating income, reflects large one-off gains. netIncomePerShareTTM and every ratio derived from it are distorted and none is used. All valuation in this dive is on consensus epsAvg, which is on an adjusted basis and reconciles to earn_cal epsActual.
Enterprise value — CHECKED AND CLEAN. enterpriseValueMultipleTTM of 5.09x implies an enterprise value of approximately $189.9B against a rebuild of market cap $88,468M + total debt $110,443M − cash $9,481M + minority interest $473M = $189,903M. An essentially exact reconciliation. EV/EBITDA of 5.09x is used as reported.
Non-equity tripwire — checked and passed. Class A common stock, NASDAQ-listed. Price $24.93; beta 0.657; volume 26.28 million shares (~$655M of turnover) — by far the heaviest share volume in this batch; 52-week band $21.92-$34.15, a 56% range. This is common equity. Note the Roberts family's Class B super-voting stake, which is not the traded security and is not represented in the share count.
Quote versus technical blocks. quote.yearHigh/yearLow report $32.86 / $21.28; tech.hi52/lo52 report $34.15 / $21.92. Differences of 3.9% and 3.0% — the largest 52-week disagreement in this batch. tech is used throughout. Note that tech.max_dd_from_peak (−59.58%) is far larger than tech.pct_from_hi (−26.91%), so the drawdown field measures genuinely and records a real 59.6% peak-to-trough decline over a longer window. quote.price is $24.93 while tech.last is $24.96 — a 0.12% difference; the quote price is used.
8. Technicals, insiders and verdict
Technicals
- Price $24.93, +4.9% above the 50-day moving average of $23.77 and MINUS 8.6% below the 200-day average of $27.26. One of only two names in this batch below its 200-day, alongside Intuit.
- −26.9% from the 52-week high of $34.15; +13.9% above the low of $21.92. Position in the annual range: 25th percentile.
- Maximum drawdown from peak: −59.6% — the second deepest in this batch after Intuit's −59.9%, and genuinely measured rather than restating the current distance.
- RSI 59.5 and MACD +0.15 — mildly positive and essentially flat.
- Relative performance: 3-month −7.8% vs SPY +7.6%; 6-month −15.6% vs SPY +11.1%; 12-month −23.2% vs SPY +24.3% and QQQ +30.8% — a 47-point deficit. Negative at all three horizons.
- Sentiment: 0 strong buy, 34 buy, 25 hold, 2 sell across 61 analysts — the widest coverage in this batch; vendor consensus label Buy; vendor composite rating A− with a discounted-cash-flow sub-score of 5 out of 5 (computed on the overstated free-cash-flow figure corrected in Section 7, and carrying no weight here). Target consensus $28.20 (+13.1%), median $27.50, high $36 (+44.4%), low $23 (−7.7%).
Today's move. CMCSA closed UP 1.51% at $24.93 from $24.56 on 26.28 million shares — heavy volume, and the largest in this batch by a factor of six. No company-specific filing is dated 2026-08-04; the last was the 2026-07-23 results.
The honest read: the price is a discount to the analyst consensus and to our base, and the reason is visible in the accounts rather than mysterious. 25th percentile of the annual range, below the 200-day, 6.9x forward, 11% shareholder yield — on a business whose largest revenue line fell 5.5%.
Insiders — director awards only, no discretionary transaction
The file consists entirely of A-Award grants dated 2026-06-30 to five directors — Gordon Smith (1,375 units), Jeffrey A. Honickman (1,783), Edward D. Breen (815), Louise F. Brady (1,375) and Thomas J. Baltimore Jr. (1,375) — each an annual Class A common stock award at a $0 price.
There is not one open-market purchase and not one open-market sale in the file. No transaction by Brian L. Roberts or by any executive officer appears. The insider record is entirely mechanical and carries no signal, and it is reported as such. The most informative "insider" transaction is the company's own: $2,507 million of repurchase in the first half of 2026, against $7,155 million for the whole of 2025 — a decelerating programme.
Verdict
Watch.
The arithmetic. At $24.93 Comcast trades at 7.1x the FY2026 consensus of $3.501, 6.9x FY2027's $3.618 and 5.09x enterprise value to EBITDA, with a 5.29% dividend yield and an approximately 11.0% total shareholder yield. Our base fair value of $29 is 16.3% above spot with 1.69:1 asymmetry ($16 bear, $40 bull), and it is built at the enterprise level and geared because total debt of $110,443 million exceeds the $88,468 million market capitalisation.
What is genuinely good. Roughly $18-19 billion of annual free cash flow on the corrected measure — a 20% equity yield. Domestic wireless net additions of 448,000, the best quarterly result on record, at 7% penetration of Comcast's own footprint. Peacock profitable for the first time. Business Services +5.5% and Theme Parks +14.1% in FY2025. A share count down 23.3% in five years. The cleanest segment reporting in this batch — both blocks tie to the dollar, including the elimination line. And a specific, dated corporate event: the NBCUniversal and Sky separation targeted for mid-2027.
What keeps it at Watch.
First, the core is shrinking and the company is paying to slow it. Domestic broadband revenue −5.5%, residential broadband customers −167,000 in the quarter, video −280,000, connectivity relationships −230,000. The 10-Q's own words: "an increasingly competitive environment", and a free wireless line for a year that "will negatively impact average domestic broadband revenue per customer."
Second, the leverage governs the outcome and it is not modest. $110,443 million of debt against an $88,468 million market capitalisation. The equity is roughly 47% of a $190 billion enterprise, so it moves about 2.1x. That is why the fair-value range is $16 to $40 on a business whose revenue moves 2% a year.
Third, the base case requires a multiple expansion from 6.9x to 8.0x and there is almost no earnings growth to fall back on — consensus revenue DECLINES in both FY2026 and FY2027 and EPS grows 3.3%.
Fourth, the separation is announced and not agreed. The 10-Q lists Board approval, tax opinions, regulatory approvals and "completion of financing arrangements" as conditions, states there can be no assurance it will occur, and does not disclose how the $110 billion of debt divides — which is the term that decides whether the connectivity successor is investable.
Fifth, the buyback is decelerating sharply — $13.3bn, $11.3bn, $9.1bn, $7.2bn, and $5.0bn annualised — and the knowledge base contributes nothing.
Pre-registered KILL criteria — what would take this to Avoid:
- Domestic residential broadband customer net losses above 250,000 in a quarter, i.e. the improvement reversing.
- Domestic broadband revenue declining more than 6%, which would mean the pricing pivot is costing more than it retains.
- The NBCUniversal separation delayed beyond 2027, abandoned, or structured with the debt loaded onto the connectivity company.
- The dividend cut, which is 5.29% of an 11% total shareholder yield.
- Wireless net additions falling below 250,000, which would remove the one growing consumer line.
- Net-debt-to-EBITDA rising above 2.5x on the company's own measure, or a credit-rating action.
Pre-registered UPGRADE conditions — what would take this to Buy — Tactical:
- Domestic broadband revenue decline narrowing below 3% with continued customer improvement. This is the operative trigger and it is a number in the 2026-10-29 print.
- Disclosure of the debt allocation in the NBCUniversal separation showing the connectivity company below 3x net leverage.
- Wireless penetration above 8% with additions holding near 400,000, which would make convergence an economic defence rather than a promotion.
- A price below roughly $21, which is 5.8x FY2027E and takes asymmetry above 2.5:1 even against the geared bear.
- The buyback re-accelerating above $8 billion annualised, which would signal that management sees the separation financing as settled.
Where CMCSA fits in the Synthos Framework Portfolio. No position today; a named limit at roughly $21 and a named data trigger in the October print. This is a cash machine attached to a shrinking asset with more debt than equity, in the middle of dismantling itself — and the 11% shareholder yield means an investor is paid to wait, which is the strongest argument for it and not sufficient on its own. A 1.5% starter on either the price trigger or on a sub-3% broadband revenue decline, scaling to 3% only after the separation terms are disclosed. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $24.93, with the fair-value anchors, kill criteria and both triggers all gradeable.
Single biggest risk: debt larger than equity, on a shrinking revenue base. Comcast owes $110,443 million and its shares are worth $88,468 million. The enterprise is roughly $190 billion and the equity is 47% of it, so a 10% move in the enterprise value is roughly a 21% move in the share price. Meanwhile the largest revenue line in the company — domestic broadband — fell 5.5% year on year, residential broadband customers fell 167,000 in the quarter, and the 10-Q attributes it to "an increasingly competitive environment" while disclosing that management is deliberately reducing revenue per customer to defend the base. The company generates enormous cash, the debt is investment grade, and the dividend is comfortably covered — none of that is in question. What is in question is whether an enterprise whose principal asset is losing customers to two classes of new entrant deserves 5x EBITDA, 6x, or 4x — and at 2.1x gearing, the difference between those three answers is the difference between $16 and $40.
Provenance & disclosures
- Traceability: 1 knowledge-base claim names Comcast out of 58 raw hits (case-sensitive survivors on company terms: 1; used 1; discarded 57; breadth 1, net conviction NONE). The claim —
invest_like_the_best, 2017-06-01, bullish 75, horizon principle, entities TCI / Liberty Media / Charter / Comcast — is "Outsider CEOs actively used leverage for its tax shield at a clear disciplined band (Malone's 4x cash flow for cable), running slightly hotter than peers while communicating the band transparently." It is nine years old and it is about capital-allocation philosophy rather than about this company — and it is, by accident, the right question, because Comcast carries $110,443M of debt against an $88,468M market capitalisation and the NBCUniversal separation names "completion of financing arrangements" as a condition without disclosing how the debt divides. Used as framing; contributes nothing to the conviction score. A GENUINE HOMOGRAPH was caught and discarded:forward_guidance2025-07-02 (bullish 70, entities Athena / SKY / Centrifuge) is about a real-world-asset crypto protocol named Sky, NOT the British broadcaster Comcast owns — a Sky-inclusive entity sweep would have imported a bullish stablecoin-credit claim into a cable dive. A second "Sky" hit (business_breakdowns2023-05-09, on streamers bidding for Formula 1 rights) IS the broadcaster but names it only as a bidder and is discarded on content. The other 55 discards are free-text noise on "broadband" and "cable" across 30 channels, dominated by artificial-intelligence infrastructure discussion, rural-broadband policy and streaming-share commentary naming Netflix. - Data as-of: fundamentals — the June 2026 quarter, the Versant and NBCUniversal separation disclosures, the segment realignment, the cash flow statement and the operating metrics from the 10-Q filed 2026-07-23 and the 8-K of the same date carrying the complete earnings release; annual statements through 2025-12-31 from the 10-K filed 2026-02-03 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873601 = 2026-08-04T20:00:01Z ($24.93, +1.51%; 50-DMA $23.77; 200-DMA $27.26; RSI 59.5; MACD +0.15) · knowledge-base claims 2026-08-04. The fiscal year is the calendar year. All figures come from the Synthos vendor data file for CMCSA or from the SEC filings in the CMCSA archive; no figure comes from memory, recall or external retrieval.
- Filing archive contents: 10-K filed 2026-02-03 (fiscal 2025); 10-Q filed 2026-04-23 (March 2026 quarter); 10-Q filed 2026-07-23 (June 2026 quarter — the source of the Versant and NBCUniversal separation disclosures, the segment realignment, the capex and intangibles lines and the competitive commentary); 8-K filed 2026-06-12 (Item 5.07, annual-meeting results); 8-K filed 2026-06-29 (no extractable item content); 8-K filed 2026-07-23 (the complete second-quarter earnings release, 50,156 characters, including the customer-metric and segment tables — the source of the broadband, wireless, video and free-cash-flow figures). All six documents carry preserved
[TABLE]statement data. - Where the vendor was REJECTED or CORRECTED:
capitalExpenditure, which omits "cash paid for intangible assets" — $1,226M in the six months to 2026-06-30 against capital expenditures of $5,253M — so free cash flow is overstated by roughly 13%; corrected FY2025 free cash flow is approximately $19.4 billion rather than $21,893M andfreeCashFlowYieldTTMcorrects from 23.11% to approximately 20.0%, cross-checked against management's own $4.6 billion for the June quarter.est.ebitdaAvg/ebitAvg, exactly 45.69% and 31.58% ofrevenueAvgin all five forward years against an FY2025 actual EBITDA margin of 37.50% — a 22% overstatement; all forward valuation usesepsAvg, and the FY2029 (4 analysts) and FY2030 (4) rows are excluded. Trailing GAAP earnings, distorted by large one-off gains — FY2025 net income of $19,998M and Q2 2025 net income of $11,123M on $5,993M of operating income — sonetIncomePerShareTTMand every ratio derived from it are unusable and no trailing multiple is published anywhere in this dive.quote.yearHigh/yearLow($32.86/$21.28) differ fromtech.hi52/lo52($34.15/$21.92) by 3.9% and 3.0% — the largest 52-week disagreement in this batch;techis used.quote.price($24.93) differs fromtech.last($24.96) by 0.12%; the quote price is used. - Where the vendor was CLEAN and it is recorded:
seg_prodANDseg_geoBOTH TIE EXACTLY TO THE DOLLAR in FY2025 and FY2024 — the only name in this batch where both blocks are perfect, andseg_prodincludes the −$8,535M intersegment elimination line that no other name in this batch carried. And the enterprise-value reconciliation:enterpriseValueMultipleTTMof 5.09x implies approximately $189.9B against a rebuild of $189,903M — essentially exact, and it correctly includes the $473M of minority interest. - Corporate actions absent from the payload — THREE, and two of them are structural: (1) the Versant Media Group tax-free spin-off COMPLETED 2026-01-02, which "did not meet the criteria to be presented as a discontinued operation" so all 2025 comparatives contain it, with $(750)M of cash transferred and an $(8,107)M equity entry; (2) the intention announced in JUNE 2026 to separate into TWO independent publicly traded companies through a tax-free spin-off of NBCUniversal and Sky — the Media, Studios and Theme Parks segments plus the Sky businesses — targeted for MID-2027, with Comcast retaining an interest for up to one year to monetise, subject to Board approval, tax opinions, regulatory approvals and completion of financing arrangements, and with the financial statements explicitly not reflecting it; (3) a five-part segment realignment effective Q1 2026 plus a separate reclassification moving direct-to-consumer streaming commission revenue between the video and broadband lines with prior periods restated. A fourth item is a governance fact disclosed in passing and contradicted by the payload: the 10-Q describes the chief operating decision maker as "our Co-Chief Executive Officers" while the vendor
profilenames Brian L. Roberts alone; the archive contains no succession disclosure and this dive does not speculate about one. - Estimate coverage: 15 analysts on FY2026 EPS, 14 on FY2027 — the anchor for all three fair-value cross-checks, with a range of $3.373 to $4.148, a 23.0% spread; 10 on FY2028; 4 each on FY2029 and FY2030, excluded.
- Peer note: the vendor peer set — AMX, CHTR, DASH, LBRDA, RCI, RELX, T, VZ — contains three genuine comparables (Charter, AT&T, Verizon) and one absurdity: DoorDash, which is in this same research batch and has no relationship to cable. The payload carries no peer estimates, so no multiple comparison is computed — though Charter's market capitalisation of $20.6bn in the peer block is itself informative about the sector.
- Fair-value caveat, and the method differs from every other dive in this batch: the $16 / $29 / $40 anchors are built at the ENTERPRISE level and then geared to the equity, because with total debt of $110,443M exceeding the $88,468M market capitalisation an equity-multiple approach understates both tails. Bear: roughly $157bn of enterprise value (about 4.5x EBITDA on a 10% lower base) less approximately $101bn of net debt. Base: roughly $190bn (5.1x, unchanged) with EBITDA flat. Bull: roughly $242bn (5.75x on a 10% higher base). Each is cross-checked against the FY2027 consensus EPS distribution — 4.4x, 8.0x and 11.1x the mean of $3.618 — and against the 52-week range and the dividend yield. Stated arithmetic, not a discounted cash flow. The base case REQUIRES a multiple expansion from 6.9x to 8.0x, which is emphatically the fragile leg because consensus revenue DECLINES in both FY2026 and FY2027 and EPS grows only 3.3%; if the multiple holds, fair value is approximately spot and the return is roughly 14% from the shareholder yield alone. The width of the range is not imprecision — it is the arithmetic of a 2.1x-geared equity.
- Basis note:
earn_calepsActualis on an adjusted basis and diverges sharply frominc_qGAAP diluted EPS in periods containing one-off gains — most visibly Q2 2025, where GAAP was $2.97 and adjusted $1.25. ConsensusepsAvgis on the adjusted basis and all forward multiples in this dive use it; NO trailing multiple is published because trailing GAAP earnings are distorted. All year-on-year comparisons for 2026 against 2025 are affected by the Versant separation, which completed 2026-01-02 and was NOT treated as a discontinued operation, so 2025 comparatives contain Versant and 2026 periods do not; this is stated wherever a comparison is drawn. - Timing: second-quarter results were released 2026-07-23, twelve days before this dive, and beat consensus EPS by 7.2% ($1.04 against $0.97) on revenue 2.4% above estimate — the fifth consecutive beat. The next print is 2026-10-29, 86 days away, against consensus of $1.01 in EPS and $29,532M of revenue. The larger event is the NBCUniversal and Sky separation, targeted for mid-2027. The insider file contains only five director stock awards dated 2026-06-30 and no open-market purchase or sale by anyone. 2026-08-04 carried no company-specific filing.
- Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
- Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
- Version: 2026-08-04-full.