Connectivity Bends on Schedule, but Parks Fall 17% and Broadband Slips: Maintaining Hold, Fair Value $22–25

Results vs. Consensus

Q3 2026 Scorecard

Revenue consensus ran from $29.22 billion to $29.35 billion across providers, so the top line landed in the middle of the range. The beat was entirely in profit, and most of the year-over-year profit growth came from the media side of the house. The subscriber lines were weaker. Broadband losses came in worse than every estimate we found, and wireless additions fell short of both the prior year and the prior quarter.

Year-Over-Year Comparisons

The GAAP and adjusted lines moved in opposite directions, and the reason is mostly accounting. GAAP EPS rose 7.2% because amortization of acquisition-related intangibles fell to $524 million pre-tax from $824 million after Versant's intangibles left the balance sheet, because the equity line swung to a $129 million gain from a $90 million loss, and because interest expense fell 14.7%. Adjusted EPS strips the amortization and investment items and fell 4.0%. A share count 3.4% lower than a year ago did not offset a 7.3% decline in adjusted net income, and the buyback pause means that share-count support is now over.

Quarter-Over-Quarter Comparisons

Read the sequential columns carefully, because the third quarter is seasonally weak for Residential margins and seasonally strong for parks. The Residential margin fell 180 basis points from Q2, but the same seasonal step a year earlier was 250 basis points (39.3% to 36.8%), so the underlying margin trend improved. The broadband ARPU figure is our calculation from reported broadband revenue and average residential customers, and the same method reproduces the 3.8% decline Comcast disclosed for Q2.

Revenue assessment. Domestic broadband revenue fell 4.7% to $6,241 million, the first narrowing in the decline since the line turned negative in Q4 2025 (Q1 fell 5.1% and Q2 fell 5.5%). Wireless service revenue grew 14.7% to $1,050 million, so convergence revenue fell 2.3% after a 3.2% decline in Q2. The gap that defined the bear case last quarter is closing slowly. Broadband lost $309 million of revenue year over year against $135 million gained in wireless service, compared with $369 million lost against $125 million gained in Q2. Wireless equipment revenue jumped 43.6%, which flatters the residential total but carries little margin. Video fell 8.7% on customer losses.

Margin assessment. The Residential margin of 35.9% is down 90 basis points year over year, against declines of roughly 140 to 160 basis points in each of the prior four quarters. Programming expense fell 8.3% with video customers, and Residential non-programming expense rose 1.2%. Across both Connectivity segments, every cost line except the wireless-driven direct product costs fell (detail below). This is the first quarter in which our coverage can document the controllable cost lines falling, and it predates the transformation program management announced on this call. The Business Services margin slipped 50 basis points to 55.9%.

EPS assessment. Adjusted EPS of $1.07 is down 4.0% from $1.12, a much smaller decline than Q2's 16.7%. The improvement comes from Peacock, Studios and lower interest expense, set against Theme Parks and Connectivity. With the buyback paused and the share count flat (3,562 million against 3,570 million in Q2), every future EPS gain now has to come from operating profit or interest savings.

Segment Performance

Connectivity & Platforms

Residential Connectivity & Platforms revenue detail

Connectivity & Platforms cost detail

Programming and wireless device and access costs move with volumes. The other four lines (technical and support, marketing, customer service and other) are the ones management controls, and together they fell to $6,286 million from $6,411 million, a 1.9% decline. Technical and customer service costs each fell 2.4%, consistent with management's claim that service contacts are down about 20%. Marketing was roughly flat in a second half that management described as more competitive than the first.

"This quarter's financial results reflect the significant investments we're making to strengthen our competitive position, including fundamental changes to our pricing, packaging, and overall customer proposition. These investments are creating near-term financial pressure, but we believe they're necessary to strengthen customer relationships and position the business for return to growth over the long term."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: This is the quarter management promised in July, and it delivered it. Residential EBITDA fell 5.8% after 8.0% in Q2, total C&P EBITDA fell 4.4% after 5.8%, and the controllable cost lines fell, the first quarter in which our coverage can document that. The improvement comes from lapping last year's pricing reset and from conversion revenue, not from the subscriber base, which is still shrinking. That matters for durability. The ARPU recovery has a mechanical driver that continues into 2027, while the subscriber trend depends on competitors' pricing, which Comcast does not control.

Business Services Connectivity

Revenue grew 2.7% to $2,645 million and EBITDA grew 1.7% to $1,478 million. Costs rose 4.0% on direct product costs and marketing. There was no one-off this quarter, so the 2.7% is clean and compares with the "just under 3%" underlying rate management cited in July. Small business revenue was flat, and enterprise solutions supplied the growth. Management sized the segment at nearly $6 billion of annual EBITDA.

"In small business, revenue was essentially flat, and while we are seeing ongoing competitive pressure, we are offsetting this through uptake of additional products and services, including mobile, WiFi Connect for backup connectivity, and security solutions like Xfinity Shield."
— Jason Armstrong, Chief Financial Officer

Assessment: The segment we once called the most reliable compounder in the portfolio is now growing EBITDA at 1.7%, its slowest rate in our coverage. Fixed wireless and fiber are reaching small business the same way they reached the home. Enterprise growth is real but too small to carry a business whose small-business base is flat. We now treat Business Services as a low-single-digit grower with a 56% margin, which is valuable but no longer an offset to Residential.

Subscriber Metrics

Over nine months, residential broadband losses total 376,000 against 475,000 a year earlier, wireless line additions total 1,251,000 against 1,115,000, and video losses total 876,000 against 1,008,000. Domestic residential passings grew to 59.7 million from 58.6 million, but broadband penetration of those passings fell to 47.5% from 49.3%.

Assessment: Every domestic line was worse than a year ago, and the improvement in the total came from Sky's international base (+173,000), which does not drive the U.S. thesis. The year-to-date broadband comparison is still better than 2025, so management's April promise to improve on 2025's 653,000 full-year losses remains intact. Comcast can lose up to 276,000 in Q4 and still meet it. The penetration figure is the one we would watch. Comcast is adding about a million passings a year while its share of the homes it passes falls by almost two points.

Five Quarters of the Pivot

The 2025 columns are as first reported, except that broadband net additions use Comcast's recast residential basis. The C&P EBITDA growth rates for Q1 through Q3 2026 are constant currency. The 2026 broadband ARPU figures also reflect a reclassification of streaming commissions into broadband revenue.

Assessment: The table shows two businesses moving in opposite directions. Connectivity's financial lines (ARPU and EBITDA) bottomed in Q2 and turned, while its volume lines (broadband and wireless additions) did not. Content moved the other way. Peacock went from a $552 million loss to a $261 million profit in three quarters, and Theme Parks went from +33% to (17)%. The separation will hand these two trajectories to different shareholders, and the Comcast that remains is the one whose recovery depends on price rather than volume.

Content & Experiences

Media and Peacock: the second profitable quarter

Media revenue grew 13.0% to $5,359 million, or 8.3% excluding the World Cup. Domestic distribution grew 22.4% on Peacock's higher rates and subscriber growth, and international networks grew 2.5%. Domestic advertising grew 14.6% with the World Cup and 0.3% without it, as Peacock's advertising growth offset declines at the linear networks. Management described the advertising market as relatively stable once the World Cup and game counts are normalized, with technology spending offsetting weaker consumer categories. Media EBITDA rose to $769 million from $247 million.

"Peacock was the primary driver of media results, delivering its second consecutive quarter of profitability with $261 million of EBITDA, a $478 million improvement year-over-year. Peacock revenue increased 34%, reflecting strong growth in both distribution and advertising. Distribution revenue increased 44%, driven by higher sub fees and growth in paid subscribers, which increased by 7 million year-over-year. Importantly, paid subscribers remained stable sequentially, demonstrating strong retention even as we began implementing a price increase earlier in the quarter."
— Jason Armstrong, Chief Financial Officer

The 10-Q shows how the swing happened. Peacock's costs were $1.6 billion in both the third quarter of 2026 and the third quarter of 2025, so the $478 million improvement came from about $0.5 billion of added revenue on a flat cost base (the reported revenue and cost figures are rounded to $0.1 billion). Year to date, Peacock's EBITDA sums to $18 million, close to breakeven, with the NBA-heavy fourth quarter still ahead.

Assessment: Holding 48 million subscribers through a price increase is the most important Peacock data point of the year, because it shows pricing power rather than promotion-led growth. Flat costs show the operating leverage the bull case always claimed, now confirmed. Our caveat is the calendar. Q4 2025 was a $552 million loss with the first NBA season, so a profitable full year 2026 is unlikely, and we lower our expectation to roughly breakeven to a loss of a few hundred million. Even so, that would be an improvement of about $0.8 billion to $1.1 billion on 2025's loss of roughly $1.1 billion.

Studios

Revenue dipped 1.1% to $2,967 million, but the mix improved sharply. Theatrical revenue rose 76.6% to $1,129 million on The Odyssey, Minions & Monsters and the carryover of Obsession, while content licensing fell 23.7% on the timing of television deliveries. EBITDA rose 18.6% to $402 million, and nine-month Studios EBITDA of $1,158 million is up 72% from $674 million. The Odyssey has grossed nearly $1.8 billion worldwide and is now Universal's highest-grossing film, and Universal became the first studio to pass $5 billion at the 2026 global box office.

"At Studios, revenue declined 1%, reflecting the timing of lower margin television licensing deals. However, theatrical revenue increased 77%, contributing to 19% growth in Studios EBITDA."
— Jason Armstrong, Chief Financial Officer

Assessment: Studios profits are now led by theatrical releases rather than intercompany licensing, which makes them higher quality than the Q1 doubling. Part of the gain is timing, though: the release credits lower marketing spend around recent and upcoming releases as well as lower licensing costs. A year with a $1.8 billion event film, a Focus Features record and a franchise sequel is not a run rate, and we do not extrapolate it into 2027. Its real value to the thesis is the separation. The new NBCUniversal will carry a studio with demonstrated hit-making capacity, and that makes the spin easier to value.

Theme Parks: the second decline, and a steeper one

Revenue fell 4.4% to $2,597 million on lower admissions revenue, including the effect of increased promotional activity. Costs rose 2.3% on domestic park operations, so EBITDA fell 17.1% to $780 million and the margin dropped to 30.0% from 34.6%. Nine-month parks EBITDA of $1,939 million is down 2.8%.

"In Orlando, the softness that began in June continued through the third quarter and was the primary driver of weaker financial results. In Japan, China-related travel restrictions continued to weigh on attendance and financial performance. Looking ahead, we expect to begin lapping the initial pressure in Japan in the fourth quarter, while Hollywood should benefit from the recent opening of the new Fast & Furious coaster. However, softer demand and difficult comparisons in Orlando are expected to remain headwinds to overall theme park results in the fourth quarter."
— Jason Armstrong, Chief Financial Officer

Assessment: The decline tripled in rate from Q2's 5.1%, and the release now says Comcast is using price to defend attendance. The fourth quarter compares against roughly $1 billion of parks EBITDA in Q4 2025, a record at the time, so a third consecutive decline is the base case on management's own guidance. Parks contribute about 8% to 9% of consolidated EBITDA, so the consolidated earnings impact is modest. The impact on the NBCUniversal equity case is not, because parks were meant to be the growth engine of the company being spun.

Key Topics & Management Commentary

Overall Management Tone: More operationally specific than in July and more candid about competition, with live questions back after last quarter's pre-submitted format. Management delivered a promised number (the narrowing in Connectivity EBITDA) and finally supplied the conversion rate investors had requested on four calls. It was least convincing on parks, where the explanation changed from July's, and on the transformation, where an ambition of "billions" came without a figure, a timeline or a headcount.

1. Connectivity Bends on Schedule

In July, management said it expected "modest improvements starting in the third quarter" in Connectivity, with ARPU relief building as free lines converted. We had flagged Q3 as the first test without a major marketing event behind it, and the World Cup's final weeks in July were the only event in the period. It arrived. C&P EBITDA fell 4.4% after 5.8%, convergence ARPA fell 0.5% after 1.5%, and domestic broadband revenue fell 4.7% after 5.5%.

"As expected, these investments continued to pressure near-term financial results, with revenue declining 2.6% and EBITDA declining 4.4%. But as the quarter demonstrated, we're beginning to see modest improvement as we lap our initial investments and convert free wireless lines into paid relationships."
— Jason Armstrong, Chief Financial Officer

For the fourth quarter, the CFO guided to "further slight improvement in these underlying financial trends", while customers continue to migrate to the new pricing and free lines continue to be offered.

Assessment: This is the fifth consecutive quarter in which management's directional operating guidance proved accurate, and the first time a promised improvement in Connectivity EBITDA came due and was delivered. We give that record substantial weight. Half of our first upgrade trigger (EBITDA narrowing in both Q3 and Q4) is now met. The word for Q4 is "slight", however. Our model has the Q4 C&P decline at 3.5% to 4.0%, and a narrower decline than that would be a result beyond the guide.

2. Broadband Losses Worse Year Over Year, as Pre-Announced

Residential broadband losses of 144,000 were 53,000 worse than a year earlier, the first year-over-year deterioration since Q4 2025. The CFO pre-announced the direction at a September investor conference, citing fiber gig offers at $30 to $40 a month and new overbuilds covering about 4% to 5% of the footprint a year. On the call he added a second cause.

"As Mike highlighted, competition remains intense, with particularly aggressive fiber promotions in certain markets during the quarter. At the same time, we had a higher than typical volume of customers facing pricing transitions across our legacy broadband customer base. Together, these dynamics contributed to higher churn and year-over-year pressure on both broadband and wireless subscriber trends."
— Jason Armstrong, Chief Financial Officer

The pricing transitions were two-year promotional roll-offs in the legacy base, the customers who signed before last year's pricing reset. Management argued that fiber priced at $30 to $40 against a build cost above $1,500 per home is uneconomic and will not last.

Assessment: Both causes have an end date, which is the strongest argument management has. The legacy promotional cohort shrinks with each quarter, and $30 fiber gig offers cannot earn a return on $1,500 per home. Fiber pricing is outside Comcast's control. The roll-offs are the residue of its own past pricing, and management accepted that defusing them is its job. The falling penetration of passings (47.5% from 49.3%) is the longer-term signal. The structural bear point stays at Materializing. What it needs to de-escalate is a quarter in which broadband losses improve year over year without promotions.

3. Free-Line Conversion: 70%, a Number at Last

For four consecutive calls, management described free-line conversion only with a qualifier ("majority", "vast majority", "significant majority") and no figure. In July we set a 55% threshold, with retained ARPU, as an upgrade trigger.

"Approximately half of our residential postpaid phone connects continued to come from customers taking a free line, and we're now converting 70% of those free lines into paying relationships. This is an important validation of our strategy to build a wireless awareness, move customers into paying relationships, and drive upsell into multi-line and premium wireless."
— Jason Armstrong, Chief Financial Officer

In Q&A, management added that converting customers port numbers, use data and buy premium plans at rates similar to or above day-one paying customers, while holding fewer lines per account.

Assessment: Seventy percent is comfortably above our 55% bar and at the top of the 55% to 70% range we modelled in April, and it validates the core mechanism of the pivot: free lines are buying paying customers rather than renting promotional ones. The trigger also asked for retained ARPU, and that half is not met, because convergence ARPA still fell 0.5%. The trigger therefore stays open. We will judge its ARPU half on convergence ARPA, management's preferred measure, turning positive year over year. One more quarter of narrowing like this one would get there, although our estimates assume it is still slightly negative in Q4.

4. Wireless: Fewer Additions, Better Customers

Line additions of 368,000 were the lowest since Q4 2025, down from 448,000 in Q2 and 414,000 a year ago. Management attributed the slowdown to lower industry switching after competitors launched new converged offers of their own. The mix kept improving.

"Premium unlimited plans represented approximately 40% of our postpaid phone connects, up from 30% last quarter, and premium customers now account for roughly 25% of our phone base."
— Jason Armstrong, Chief Financial Officer

Total lines reached 10.6 million, up 18% year over year, equal to nearly 18% penetration of the residential broadband base and about 7% of addressable lines in the footprint. Management said 2026 will be Comcast's best year ever for line additions, which requires only about 230,000 in Q4 after 2025's 1.48 million. One July data point reversed. In July management said overall mobile churn was falling, and this quarter it said higher churn pressured wireless as well as broadband trends.

Assessment: A move from 30% to 40% premium mix in one quarter matters more for revenue than the 80,000 shortfall in volume. Premium plans carry the ARPA that conversion needs. The volume slowdown is still worth noting, because it coincides with competitors' new converged offers. Comcast's capital-light advantage is real, but rivals now sell a similar bundle from the other side. The convergence pillar stays On Track, with a slower volume path than we modelled.

5. The Largest Cost Transformation in Comcast's History, and a Q4 Charge Above $1 Billion

The quarter's largest new disclosure was a program management called the most significant operating and cost transformation in the company's history. It covers a redesigned organization, leadership changes in customer experience, growth and marketing, faster technology deployment, and AI in customer service through partnerships with Sierra and Cresta.

"Building on the organizational changes Steve has made, including strategic external hires, we've embarked on the most significant operating and cost transformation in our company's history, with an opportunity we believe will ultimately amount to billions of dollars over time."
— Mike Cavanagh, Co-Chief Executive Officer

"Before I do that, as I said earlier, we expect to have a significant charge in the fourth quarter related mostly to our transformation efforts, and I would quantify that at over $1 billion."
— Jason Armstrong, Chief Financial Officer

Press reports before the print described planned job cuts across the cable, broadband and wireless businesses. Management gave no headcount, savings target or timeline, and said it would provide more context early next year, when incoming Comcast CEO Michael Angelakis joins the discussion.

Assessment: This is potentially the most important change to 2027 earnings, and it is entirely unquantified. For scale, each $1 billion of net annual savings that Comcast keeps would add roughly $0.21 to EPS at the current tax rate and share count, about 6% of our 2027 estimate. We include none of it until management states a target. Management also framed the program as creating "capacity to invest in growth", which signals that part of any saving will be reinvested in wireless, enterprise and edge computing rather than reach the bottom line.

6. Theme Parks: The Explanation Now Includes the Operator

In July, management called the Orlando softness "essentially weakness in attendance" driven by consumer sentiment and higher travel costs. It said it was not "at all disappointed" in Epic Universe. This quarter the account added two causes that are Comcast's own.

"In addition to the macroeconomic and geopolitical headwinds we have referenced in the past, we believe 2 other factors have contributed to this recent softness. First, we believe competitors were quicker to anticipate and recognize softness in demand and adapt their marketing strategies. Second, we continue to work through capacity challenges and ride throughput at Epic Universe, given the complexities of opening the most technologically advanced theme park ever built."
— Mike Cavanagh, Co-Chief Executive Officer

Management has appointed a new Chief Brand and Commercial Officer for Universal Destinations & Experiences, Luc Bondar from United Airlines, and the former NBCUniversal CFO, Randy Culbertson, as Chief Business Officer. It said Orlando attendance, per-capita spending and financial performance remain well above their levels of two years ago.

"Looking ahead, we expect performance to strengthen as we move through 2027, with a return to growth in the second half of the year."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: The new explanation is more credible than the old one, and it is better news for the long-term case: marketing and ride throughput can be fixed, while consumer sentiment cannot. It also means the July diagnosis was incomplete, in a summer when the largest competitor reported record quarterly revenue in its parks division. Management has now given a dated, falsifiable recovery claim (growth in the second half of 2027), and we will hold it to that date. The Q4 decline it guides to is already in our numbers. A recovery date that moves would not be.

7. Peacock Holds Its Subscribers Through a Price Increase

Peacock's second profitable quarter was covered above. The strategic addition is distribution. The YouTube agreement announced in July adds ad-supported Peacock Premium to U.S. YouTube Premium memberships from early 2027, extends NBCUniversal's YouTube TV carriage, and widens the advertising relationship. Management framed bundling as the next phase of a service that spent six years building a direct, high-ARPU base.

"This bundle will give us another opportunity to reach a younger, an incremental audience"
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: A wholesale bundle with the largest video platform adds reach at a lower price per subscriber. The key unknown is whether it cannibalizes direct subscribers. Management said it is watching for that and gave no economics. The pricing result this quarter is the better evidence of the service's value: no subscriber loss after a price increase, with the World Cup ending early in the quarter.

8. Studios: A $5 Billion Box Office Year

Universal passed $5 billion at the 2026 global box office before any other studio, led by The Odyssey at nearly $1.8 billion, and Obsession set a Focus Features record as the year's top-grossing original film. Theatrical revenue rose 77% in the quarter.

"Our film slate delivered exceptional results led by Christopher Nolan's The Odyssey, which generated $1.8 billion in worldwide box office to become Universal's biggest film ever, and Obsession, which set a new box office record for Focus Features and became the top grossing original film of 2026."
— Jason Armstrong, Chief Financial Officer

Assessment: The year's top film was a non-franchise event picture and its top original came from the specialty label, which supports a premium for the studio in a separated NBCUniversal. Theatrical results are inherently lumpy, and the comparison for the third quarter of 2027 will be difficult.

9. Separation on Track for Mid-2027, Capital Return Still Undefined

The separation is in its execution phase, with standalone public-company infrastructure being built and Chris Halpin hired as NBCUniversal CFO. The 10-Q dates completion to mid-2027 and repeats that Comcast expects to keep up to 19.9% of NBCUniversal for up to a year. The ITV acquisition is now sized in the 10-Q at up to £1.6 billion (£1.2 billion in cash), with closing expected in the second half of 2027, around or after the spin.

"We are on track to complete the transaction by the middle of next year and are well into the execution phase, which includes building the standalone public company infrastructure."
— Mike Cavanagh, Co-Chief Executive Officer

"As previously announced, we paused our share repurchases as of July 1st and expect that pause to remain in place at least through the separation, as we ensure both businesses are positioned with strong balance sheets and investment grade ratings."
— Jason Armstrong, Chief Financial Officer

Assessment: The timetable is intact, which keeps our separation-delay trigger quiet. On capital return, the written policy did not change. The release says the pause remains "through the proposed separation", as in July. On the call the CFO said "at least through the separation", a spoken hedge we will watch rather than a new policy. Either way there is no exchange ratio, no leverage target and no dividend policy for either company. Meanwhile, debt is falling. Total debt is $87.4 billion against $98.9 billion at year-end, a decline that includes $1.0 billion of Versant notes that left with that spin, and net debt is about 2.4x trailing EBITDA. Beyond the dividend, debt reduction is the only capital-return mechanism running at present, and it accretes to equity more slowly and less visibly than a buyback.

10. The Product Keeps Improving While the Base Shrinks

The leading indicators improved again. Net promoter scores rose sequentially and year over year, and gig-plus tiers reached about 45% of the base. Service contacts are down about 20% with 30 million fewer interactions year to date, and the Xfinity Shield security product launched this quarter.

"Approximately 45% of our broadband base is now on gig-plus speed tiers, up nearly 10 percentage points from a year ago."
— Jason Armstrong, Chief Financial Officer

"We've already reduced customer service contact rates by roughly 20% and eliminated more than 30 million service interactions year-to-date versus the prior year."
— Jason Armstrong, Chief Financial Officer

Assessment: Better speeds, better service and lower contact costs did not stop the base from shrinking faster than a year ago. That gap is the core of the bear case: in contested markets, price rather than product quality is setting share. The service metrics do feed the cost line directly, as the 2.4% declines in technical and customer service expense show. The product investment is paying back in margin before it pays back in subscribers.

Guidance & Outlook

July Commitments, Graded

Implied Q4: Management's guide implies a Q4 with modestly better Connectivity trends, weaker free cash flow as the timing benefit reverses, a third consecutive parks decline against a record comparison, and Peacock absorbing the second NBA season. Add a charge of more than $1 billion, which we expect to sit outside adjusted results.

Street at: Fourth-quarter adjusted EPS consensus sits near $0.61 to $0.62, which implies full-year 2026 adjusted EPS of about $3.5 against $2.90 earned in the first nine months. The 2027 consensus is in the low-to-mid $3.50s. We are at $3.50 to $3.60 for 2026 and $3.60 for 2027.

Guidance style: Directional operating guidance from this team has been right for five quarters, including both September's warning and July's promise. Strategic and multi-year statements remain less reliable: separation was ruled out in January and announced in June, and the parks softness management called temporary in July is now guided to last through at least the first half of 2027.

Analyst Q&A Highlights

Whether Competitors' Recent Price Increases Have Eased the Broadband Fight

The first question tested the bull hope that the irrational pricing the CFO described in September was ending, given recent broadband price increases across the industry. Management declined to call a turn.

Q: "I know it's very early days, but have those price changes that have happened in the market had any impact yet on the kind of results that you're seeing as the competition, I guess, to use your words, maybe becomes a little less irrational?"
— Craig Moffett, MoffettNathanson

A: "I would generally characterize the environment is it's been competitive all year. It's been more competitive in the second half than it has been in the first half for sure." ... "I will not call the last couple of weeks as a real change in trend. I think it continues to be very competitive, but Steve can fill in some of the details as well."
— Jason Armstrong, Chief Financial Officer

A: "As we really assess the marketplace, we maintain that gig pricing at $30-$40, in some cases even less, is going to be really hard to sustain."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: The honest answer, and a useful one. Management said the second half has been more competitive than the first and that pockets of irrational pricing persist into Q4, so we should not expect the competitive relief case to show up in Q4 subscriber numbers. The CFO also defended the free-line offer as rational, since it requires no spectrum or towers, a contrast with fiber priced below its build cost.

What the "Higher Than Typical" Pricing Transitions Were

A follow-up asked management to unpack the second cause of the broadband miss. The answer located it in the legacy base and accepted responsibility for it.

Q: "Jason, in your prepared remarks, you talked about, I think there's a higher than typical volume of customers facing transition. Can you maybe just help unpack what that is? Does that pertain to the legacy customers? Is that new pricing and packaging customers?"
— Sebastiano Petti, JPMorgan

A: "What we are referring to there is it is largely 2 years promotional roll-off, which would refer back to our legacy base. I think I highlighted that in the script. That has been a pain point for us. Obviously, we have, in the last year and going forward, it is going to be on us to figure out how to increasingly inoculate that part of the base and make sure customers are not rolling off with big step-ups."
— Jason Armstrong, Chief Financial Officer

Assessment: This is the more encouraging of the two broadband explanations, because the legacy two-year cohorts are finite and the new pricing structure was designed to remove the roll-off step-up. It also shows that the old base still carries roll-off risk, so some of the pressure in coming quarters is already locked in. Management did not size the remaining legacy cohort.

Whether Convergence Has a Tipping Point

The question asked whether 18% wireless penetration of the broadband base is near a level at which broadband performance visibly improves. The answer did not name one. It disclosed instead that switching had slowed and described the behaviour of converting free-line customers.

Q: "Keeping on the theme of broadband and convergence, curious, as you're at an 18% penetration of your residential broadband customers, is there a tipping point that we should all keep in mind that you can get to where that then starts to show a meaningfully better performance in converged revenue and particularly in broadband as you keep building that scale of the bundle?"
— Michael Rollins, Citi

A: "But similar to broadband, it remains competitive, that marketplace, and we saw switching activity slow down a bit in the quarter with some of the new product offerings in the marketplace, some of the new converged offerings in the marketplace. With that being said, we still added close to 370,000 lines, and in 2026, it will be our best mobile line net add year in the company's history." ... "When we look at the dynamic of those customers, porting is very similar to a paid customer. Usage is similar to a paid customer. Our premium mix is actually a bit higher in that free line base as we really leaned into the premium segment. Average lines per account are a bit less, but the good news there is we continue to see those lines per account increase over time."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: The question asked for a threshold and got none, which is fair, since nobody can identify one in advance. The behavioural detail is the useful part. Converting free-line customers port numbers and use the service like paying ones and buy more premium, which is the opposite of promotional churn. The slower switching came from rivals' converged bundles, which shows the bundle is no longer unique to Comcast.

The Financial Shape of the Transformation Beyond the Q4 Charge

The question asked what the program means for 2027 EBITDA and free cash flow, and how much of it is reinvestment. Management framed it as transformation rather than cost cutting and disclosed an edge-computing option.

Q: "I was hoping to get more color on the operating and cost transformation plan. I know it's early days, but is there anything more you can share on how we should think about the financial implications beyond the Q4 charge, especially as we look to 2027 operating EBITDA and free cash flow trends?"
— Kutgun Maral, Evercore ISI

A: "What we said is we'll take $1 billion+ charge in the fourth quarter. That is to get ahead of cost actions that'll facilitate a fast start into 2027 as we get into some of these initiatives. You're right to more broadly frame this as a transformation exercise. We don't look at it as a cost-cutting exercise." ... "We've got 1,400 headends out there. We've got 250 MW of capacity. We've got space, power, cooling, and capabilities, along with high-density networks going into every one of these. We think that is a big opportunity. I would not put that in front of you super near term, but that is one thing we are investing in, obviously to see what the opportunity is long term."
— Jason Armstrong, Chief Financial Officer

Assessment: No 2027 number was offered, so the question went unanswered on its own terms. The phrase "a fast start into 2027" tells us the cost actions are front-loaded into the charge, so some savings should appear early in the year, but the reinvestment framing caps how much reaches EBITDA. The 250 megawatts of powered headend space is a real asset in a market short of power, and management was right not to put a value on it yet.

Pricing Power When Consumers Have AI Agents

Consumer AI agents that negotiate bills on a customer's behalf launched in the weeks before the print, and one analyst asked whether they are a deflationary force. Management gave a framework rather than a number.

Q: "First, I wanted to hear how you are thinking about your overall pricing power in broadband, and do you think the rise of AI agents like Muse could potentially impact the connectivity sector? Do you think it could be a deflationary force or an opportunity for you on both the customer acquisition and retention front?"
— Sean Diffley, Morgan Stanley

A: "The first is, do you have a product that consumers are using or engaging with less? Hence, the question might be, are they getting value? An AI agent might point that out and say, 'Are you getting the right value? Do you still want this service, or do you want to pay, or could you pay less for this service?' I think the connectivity industry, in general, is pretty well positioned there. If you look at rates of usage, broadband usage up 10% year-over-year pretty consistently."
— Jason Armstrong, Chief Financial Officer

Assessment: The framework is sound for retention: a product used more each year is hard for an agent to talk a customer out of. It does not address the more direct risk, which is an agent that negotiates the price of a product the customer keeps. That is a margin problem rather than a churn problem, and it hits hardest at a company that has just reset pricing to win share. Management's most useful point was offensive, that an agent shopping for a cheaper wireless bill in Comcast's footprint would likely land on Xfinity Mobile.

Whether Comcast Would Wholesale Its Network to a Satellite Entrant

The same analyst asked whether Comcast would partner with a satellite operator using its CBRS spectrum and Wi-Fi, or whether that would arm a future competitor. The question came two weeks after SpaceX agreed to buy nationwide wireless spectrum. Management kept every option open.

Q: "Then separate question, would you ever consider doing a wholesale partnership with a satellite player leveraging your CBRS spectrum and Wi-Fi, or do you think that would be too great of a risk to enable a potentially disruptive competitor down the line?"
— Sean Diffley, Morgan Stanley

A: "To your second part of your question, today we do partner with Starlink today in Comcast Business. We combine our connected managed connectivity portfolio with their satellite capabilities for enterprise customers. Overall, we are very pleased with our MVNO partners, Verizon, T-Mobile, they are great. As we have always shown, we are open to strategic discussions and partnerships that make sense for us."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: A careful non-answer, and probably the right one. Comcast's wireless economics depend on its MVNO hosts, and praising them on the call costs nothing. Its Wi-Fi footprint and CBRS spectrum are exactly the assets a satellite carrier without a ground network would want. If SpaceX builds a nationwide carrier, Comcast holds something it needs, which is a better position than that of AT&T and T-Mobile, whose shares fell about 8% the day after this call.

When Orlando Turns, and What Management Will Do About It

The parks question asked when Orlando trends improve and whether recovery must wait for the June 2027 anniversary of the softness. Management said yes, in effect, and listed the operating response.

Q: "On the Orlando side, when do you think we could see Orlando trends begin to improve? Will we have to wait until the lapping of those initial headwinds in June of 2027?"
— Michael Ng, Goldman Sachs

A: "I think in Orlando, I would expect to see the softness there just trend similarly at least next quarter and in the first half of next year and start to lap from there. We are not complacent about that. We are focused on the things we can control. We have got a new head of marketing, so we are going to bring some promotional and product packages into the market, three-park passes and the like, really focus on media investment, on high return opportunities to get people into our collection of parks in Orlando."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: Management has now dated the trough explicitly: three more weak quarters, then the anniversary. The remedies (promotional packages, three-park passes and more media spending) defend attendance at the cost of yield and marketing expense. That supports the "temporary" label for volume but suggests margins will lag attendance in the recovery. We model 2027 parks EBITDA as roughly flat, with the first half down and the second half up.

Peacock's Profit Path and the YouTube Bundle

The question asked whether quarterly profits continue, what long-term margins look like, and whether the YouTube agreement lifts profitability. Management declined to give quarterly or margin targets.

Q: "Obviously strong profitability this quarter. Do you expect that to continue on a quarterly basis going forward? Anything you can tell us about sort of longer-term target margins? Then should we see an uptick in the overall profitability of the Peacock business when the YouTube agreement kicks in early next year?"
— John Hodulik, UBS

A: "I think the quarters ahead are, you get lumpy depending on when rates kick in for different properties. Not so much focused on profitability quarter by quarter, but I think if you start looking back prospective years on an annual basis on a look forward or look back basis, I expect us to continue to drive profitability in the business for all the reasons we have been focused on and talking about."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: The lumpiness warning is consistent with an unprofitable, NBA-heavy Q4, and investors annualizing $261 million are making a mistake. The commitment is to annual improvement, which we think is credible given flat costs. No margin target and no YouTube economics were given, which leaves the bundle's effect on profit unknowable until it launches.

What They're NOT Saying

1. The size and timing of the transformation savings. "Billions of dollars over time" has no run-rate figure, no year and no split between savings kept and savings reinvested. The Q4 charge of more than $1 billion is sized; the return on it is not. Detail is deferred to early 2027.
2. Headcount behind the charge. Press reports before the print described job cuts across the connectivity businesses. The call gave no headcount, no split between severance and other costs, and no indication of how much of the charge is cash.
3. The full-year broadband commitment. The April promise to improve on 2025's full-year residential broadband losses was reaffirmed at the September conference but not repeated on this call. The arithmetic still works (Q4 can lose up to 276,000), so the omission may be routine. It is still the one dated subscriber promise outstanding.
4. How far Q4 parks will fall. Management guided to Q4 headwinds against the strongest quarter in the segment's history without a range. That matters because a third consecutive decline was one of our published downgrade triggers.
5. Capital structure and payout for either company. There is still no exchange ratio, target leverage or dividend policy. Shareholders are a quarter closer to the spin with no more information on what either security will pay, and the buyback stays paused until then.
6. The size of the Q4 cash-flow reversal. The CFO said Q3 free cash flow contained favorable timing that reverses in Q4 but did not size it. Neither the release nor the 10-Q quantifies the quarter's timing items.
7. The ITV regulatory path. ITV was mentioned once, as strengthening NBCUniversal. The UK government's stated intention to consider a public-interest intervention, and the competition review, were not discussed, although the deal is now sized at up to £1.6 billion and expected to close after the spin.
8. Wireless churn. Management said higher churn pressured both broadband and wireless trends this quarter but has never disclosed a wireless churn rate. With switching slowing and rivals bundling, that figure would now carry real information.

Market Reaction

• Pre-print setup: The stock closed at $22.08 on October 21, down 21.1% year to date against an S&P 500 up 13.3%, down 24.9% over twelve months and down 1.5% over thirty days. The 52-week closing range was $20.66 to $32.48, with the low set on October 9.
• Reaction session (October 22, before-open reporter): Opened at $22.44, up 1.6%, traded between $22.00 and $23.29, and closed at $22.84, up 3.4% or $0.76.
• Volume: 54.4 million shares against a 32.1 million thirty-day average, 1.7x normal.
• Benchmark and peers: The S&P 500 fell 1.2%. Cable and telecom stocks rose together: Charter gained 4.2%, AT&T and Verizon 3.8% each, and T-Mobile 3.3%.
• Next session (October 23): Comcast slipped 0.7% to $22.69 while AT&T fell 8.0% and T-Mobile 7.7% on a report that SpaceX was negotiating tower access for a nationwide mobile carrier.

A sector move, not a verdict. Comcast's 3.4% gain looks like a strong reaction against a falling market, but cable and telecom peers rose by similar amounts on the same day, so the print earned little relative performance. Part of the peer move was not about Comcast: AT&T was rebounding from a 6.4% fall on its own results the day before, and Charter often trades in sympathy with Comcast's print. Even so, the setup explains the muted result. The broadband disappointment had been public since September 9, when the shares fell 6.6% on the CFO's warning, and the stock had already absorbed the SpaceX spectrum news at its October low. What the print added was a better-than-expected Media profit, the first narrowing in Connectivity EBITDA and the 70% conversion figure, balanced by the parks decline, the broadband miss against consensus and an unquantified charge. A neutral outcome, priced as one.

The day after mattered more. The October 23 session gave the clearest read on how investors see Comcast's wireless model. Tower-dependent carriers lost about 8% on the prospect of a satellite-backed national competitor, while Comcast, which rents its network and offloads about 90% of mobile traffic to its own Wi-Fi, lost 0.7%. That suggests the market sees Comcast's wireless business as more exposed to competition on price than to competition on network, which is consistent with our view of the convergence pillar.

Street Perspective

Debate: Has Connectivity Troughed, or Has It Only Repriced?

Bull view: The bull case on the Street holds that the financial trough is visibly in: EBITDA, ARPU and ARPA all narrowed as promised, conversion is 70%, premium mix is rising quickly, and a multi-billion-dollar cost program has not yet touched the numbers. On this view the subscriber losses are a lagging cost of a pricing reset that has now been lapped.

Bear view: The bear camp contends that the business has repriced itself lower and is still losing share. Losses worsened year over year, penetration of passings is falling, fiber overbuild is accelerating, bill-negotiating AI agents put pressure on the price of a product customers keep, and a satellite-backed carrier could arrive. On this view a narrowing decline is still a decline.

Our take: The bulls are right about the financial lines and the bears are right about volume, and over twelve months the financial lines matter more because they drive EBITDA directly. We think C&P EBITDA declines narrow into the low single digits in 2027. We do not expect broadband net additions to turn positive in that window. That combination supports a stable equity, not a re-rating.

Debate: Are the Parks Problems Temporary or Structural?

Bull view: Some desks argue that the new explanations (slow marketing and Epic throughput) are self-inflicted and fixable, that Orlando is still well above its level of two years ago, and that management has given a dated recovery. The Hollywood coaster, a version of it coming to Orlando and the Japan comparison easing all help 2027.

Bear view: Others contend that the decline accelerated from 5% to 17%, that the remedy is promotional pricing that lowers yield, that the largest competitor reported record quarterly parks revenue this summer, and that the explanation has already changed once.

Our take: We lean toward temporary on volume and toward structural on margin. Epic's throughput problems will be engineered away, but a parks business that now uses promotions to win attendance in Orlando will recover revenue before it recovers its 2025 margin. For the consolidated Comcast equity this is a modest issue. For the NBCUniversal valuation at the spin it is a central one.

Debate: Is a Free Cash Flow Yield Near 18% Enough Without a Buyback?

Bull view: The bull argument is that the numbers already pay holders to wait. Nine-month free cash flow is $12.7 billion against $3.6 billion of dividends, debt has fallen by $11.6 billion since year-end, and full-year free cash flow in the mid-teens of billions would put the equity's free cash flow yield near 18%. On this view debt reduction transfers value to equity even without a buyback.

Bear view: The bear camp responds that free cash flow fell 14.5% in nine months, that Q3 was flattered by timing, that a charge of more than $1 billion is coming, and that the buyback stays paused through the separation, which the CFO would only say lasts "at least" that long. A cheap stock with no catalyst before mid-2027 and no defined payout after it is cheap for a reason.

Our take: Debt reduction is a real mechanism, and we gave it too little credit in July when we said cheapness had no mechanism at all. Over the next twelve months we expect $6 billion to $8 billion of net debt reduction, after the dividend, up to about $1.6 billion of cash for ITV and the cash cost of the restructuring. At an unchanged EBITDA multiple that adds 7% to 10% to equity value, or 13% to 15% in total with the dividend. Valued on earnings, where repayment shows up only as lower interest, the return is about 8%. The truth sits between the two. The EBITDA multiple held this quarter, but Connectivity EBITDA is still falling and the spin will split the debt on terms not yet published. That supports a return around the market's with an upward skew, which is still a Hold.

Model Update & Valuation Framework

How our July assumptions are tracking

The pattern is the reverse of July's. Then, we had correctly identified the improving businesses and underestimated the declining ones. This quarter the largest line, Connectivity EBITDA, did better than modelled, while three smaller lines (wireless volume, parks and Peacock's full-year sign) did worse. On balance our 2026 EBITDA estimate is roughly unchanged.

Revised assumptions

Where the valuation sits

Same-year comparison with July. On our 2027 estimate of $3.60, July's $21.92 downgrade price was about 6.1x, and today's $22.84 is about 6.3x. The multiple has barely moved in three months, so this is not a de-rating that would justify an upgrade on price alone.

Fair value framework (12-month horizon)

We are introducing a fair value range for Comcast with this note. We value the shares on 2027 adjusted EPS, the year the separation is due to complete, so the range applies to the combined value a current holder would own across both companies. We do not attempt a sum of the parts until exchange ratios and capital structures are published.

Why 6x to 7x. On trailing earnings the shares traded between about 5.5x at the October 9 low and about 7.2x at the August high during the quarter, for a business whose 2027 earnings we expect to be flat. The October low came after the CFO's broadband warning and the SpaceX spectrum news, and the bear case already sits at that level. We centre the base band at 6.0x to 7.0x, around today's 6.3x on our 2027 estimate, because the quarter's financial evidence improved rather than deteriorated. The earnings method credits debt repayment only through lower interest. An enterprise-value method credits it in full. Holding today's multiple of about 4.9x forward EBITDA on roughly $33 billion of 2027 EBITDA, and deducting the $72 billion to $74 billion of net debt we expect in a year, gives about $24 to $25 a share, in the upper half of our range.

Valuation judgment. The earnings method implies a 12-month total return of about 8% at its midpoint, including the dividend. The enterprise-value method implies 13% to 15%. We expect something in between, because the EBITDA multiple held this quarter, but Connectivity EBITDA is still falling and the spin will allocate debt on terms not yet published. A return of roughly 8% to 15% is around the market's, with an upward skew. The bear case loses about 15% and the bull case gains about a third, and the outcomes depend on unquantified variables (transformation savings, parks timing and competitive pricing) that the next two quarters will start to resolve. The risk is balanced with a modest upward tilt, and the evidence is incomplete, which is a Hold.

What we would need to see. To move to Outperform, we need a result beyond the guided path: a quantified, dated savings target from the incoming CEO in early 2027, convergence ARPA turning positive, or Connectivity EBITDA narrowing in Q4 without a third parks decline alongside it. To move to Underperform, we need Connectivity to stall while parks keeps falling, the parks recovery date to slip, or broadband losses to break 200,000. The triggers below set these out.

Thesis Scorecard Post-Earnings

Overall: The thesis is unchanged in its tags and more balanced in substance. For the first time since our April upgrade, a quarter improved the pillars we expected to improve: Connectivity's financial lines and the conversion mechanism. The new damage landed in parks, which matters more for the company being spun than for the shares as they trade today. No pillar moved status. Bull 2 became more polarized and Bear 2 eased further, but neither moved enough to change its tag.

Action: Maintaining Hold, with a fair value of $22 to $25. Management's guidance puts one July upgrade trigger (Connectivity EBITDA narrowing in Q3 and again in Q4) and one July downgrade trigger (a third consecutive parks decline) on course to fire together in January, and both outcomes are already in our estimates. We keep every July trigger as written and add two. We also state now, before either fires, how we will read them together. If those two fire together and no other trigger does, they offset and we stay at Hold. If either fails to fire, or any other trigger fires, we act on the trigger that fired. Upgrade trigger 2 stays open. The 70% conversion rate clears the 55% bar, but the retained-ARPU half is not met while convergence ARPA is still falling, and we will judge that half on convergence ARPA turning positive year over year.

Triggers. Back to Outperform on any one of: (1) Connectivity & Platforms EBITDA decline narrowing in both Q3 and Q4 2026 (Q3 met); (2) a disclosed free-line conversion rate above 55% with retained ARPU (rate met at 70%; the ARPU half is judged on convergence ARPA turning positive); (3) residential broadband losses inside 120,000 in a quarter with no marketing event; (4, new) a dated, quantified transformation target of at least $1 billion of net annual savings, which our estimates do not include. Down to Underperform on any one of: (1) residential broadband losses above 200,000 in any quarter; (2) Theme Parks EBITDA declining for a third consecutive quarter (two so far, with the third guided for Q4); (3) the separation delayed beyond the roughly one-year target, now dated mid-2027; (4) either post-separation company signalling a dividend below the current combined payout; (5, new) the second-half 2027 parks recovery date moving out.

Bottom Line

Management promised in July that the third quarter would be the first to show improvement in Connectivity, and it did. EBITDA, ARPU and ARPA all narrowed, the controllable cost base shrank, and the conversion rate we had asked about for a year arrived at 70%, well above the level we set as a test. Peacock earned another $261 million while holding its subscribers through a price increase, and Universal had a record box office year. Taken together, that is the strongest evidence yet that the pricing reset is paying back in the financial lines.

The parts that did not improve are mostly competitive, but two of the causes are Comcast's to fix. Broadband losses widened year over year because fiber competitors are pricing gig service at levels management calls uneconomic, and because legacy promotional cohorts from Comcast's own past pricing are still rolling off. Wireless additions slowed as competitors launched converged offers of their own. Theme Parks fell 17% and will keep falling into 2027, and the explanation now includes execution problems, slower marketing and Epic throughput, that management could have anticipated. A charge of more than $1 billion is coming, attached to savings that have not been quantified.

At $22.84, 6.3x our 2027 estimate and a 5.8% yield, the shares price in neither the bull nor the bear outcome. Debt reduction gives the equity a slow return mechanism that we underrated in July, and the next two quarters should resolve the three variables that matter: whether the savings are real, whether parks recover on the dated schedule, and whether competitors' pricing becomes rational. Until then the risk/reward is balanced with a modest upward tilt. We stay at Hold, are paid about 6% to wait, and have said in advance how we will read the two triggers due in January.