The Separation Management Ruled Out, a Buyback Paused, and a Downgrade Back to Hold

Results vs. Consensus

Q2 2026 Scorecard

Two of the four providers checked carried adjusted EPS consensus at $0.96 rather than $0.97; we use the more conservative $0.97, which still leaves a 7.2% beat. This was a genuine beat on both headline lines, and the stock closed down 6.8%, which tells you the print was not what the market was trading.

Year-Over-Year Comparisons

The pro forma change column is computed off pro forma 2026 figures ($29,568M of revenue and $8,923M of adjusted EBITDA) that are not shown in the table; recomputing it from the as-reported 2026 column will not reproduce it. The 66.9% GAAP EPS decline is almost entirely a comparison artifact: the prior-year quarter contained a $9.4 billion gain on the sale of Comcast's Hulu interest. Adjusted EPS down 16.7% is the fairer read, though it too is an as-reported figure carrying Versant and Sky Germany in its base and so still overstates the like-for-like decline. Free cash flow grew 2.3%, the only growth line among the profit measures.

Quarter-Over-Quarter Comparisons

The sequential picture looks strong on every consolidated profit line, and that is largely the NBA amortization rolling off from its Q1 peak plus Peacock's swing. Underneath it, the two lines that determine the long-run value of this equity both went the wrong way: broadband net additions deteriorated by 102,000 and broadband ARPU decline widened by 70 basis points. Connectivity EBITDA was essentially flat sequentially at +0.7%.

Revenue assessment. Domestic broadband revenue fell 5.5% to $6,280 million, steeper than Q1's 5.1% and the sharpest decline of the cycle, reflecting both a smaller base and lower average rates. Domestic wireless service revenue grew 14.2% to $1,007 million, so the convergence line as management now defines it fell 3.2%. The structural fact this quarter makes plain is that wireless service revenue, at roughly 16% the size of the broadband line, cannot yet offset broadband's decline in absolute dollars: broadband lost $369 million of revenue year over year while wireless service added $125 million.

Margin assessment. Residential Connectivity margin of 37.7% is down 160 basis points year over year, the fourth consecutive quarter of roughly that magnitude, though it improved 60 basis points sequentially. Programming expense fell 7.5% on video decline while non-programming expense rose 2.1% on wireless device costs and marketing. Business Services margin rose 60 basis points to 56.7%, its first increase in our coverage, but that too benefited from the fiber lease item.

EPS assessment. Adjusted EPS of $1.04 against $1.25 is a 16.7% decline. Interest expense fell slightly to $1,052 million. The sequential jump from $0.79 to $1.04 is real and reflects the NBA cost peak passing, but it should not be read as an inflection in earnings power: the year-over-year comparison is still deeply negative and the share-count tailwind that has supported EPS for years ends with the buyback pause.

Segment Performance

Connectivity & Platforms

Residential Connectivity & Platforms revenue detail

"Additionally, we continue to invest in the customer experience and go-to-market capabilities needed to support this broader shift, which contributed to a 5.8% decline in Connectivity & Platforms EBITDA. These results are consistent with our remarks on last quarter's earnings call, where we previewed incremental pressure on both ARPU and EBITDA growth."
— Jason Armstrong, Chief Financial Officer

Assessment: Management is correct that this is consistent with what they guided in April, and that consistency has now held for four consecutive quarters. But consistency with a guide to deterioration is not the same as progress. Residential EBITDA of $6,448 million is down 8.0% year over year, the steepest decline of the pivot, and the segment that produces roughly three quarters of consolidated EBITDA has now declined year over year in every quarter we have covered. The sequential margin improvement of 60 basis points is the first tentative sign of a floor, and management expects "modest improvements starting in the third quarter."

Business Services Connectivity

Revenue grew 3.7% to $2,671 million and EBITDA grew 5.0% to $1,516 million with margin up 60 basis points to 56.7%. Management disclosed that both figures benefited from a non-recurring long-term fiber lease renewal, and that underlying growth was just under 3% for each. The advanced-solutions mix continues to scale: three years ago Comcast sold roughly $0.20 of advanced solutions per dollar of connectivity, and that figure is now closer to $0.70. The T-Mobile MVNO for business customers went live in the quarter.

Assessment: The voluntary disclosure of the non-recurring item is creditable and necessary, because underlying growth of just under 3% is a meaningful step down from the 5.8% to 6.2% range this segment posted through 2025. Management attributes the deceleration to having lapped an acquisition, which is fair, but the segment we have repeatedly called the most reliable compounder in the portfolio is now growing at roughly half its prior rate. The advanced-solutions attach-rate progression from $0.20 to $0.70 per connectivity dollar is the genuinely encouraging structural datapoint.

Subscriber Metrics

Assessment: Every line improved year over year and every line except wireless and video deteriorated sequentially. That combination is exactly what management's Q1 caveat predicted: strip the Legendary February promotions and the underlying run rate of improvement is roughly 34,000 per quarter, not 117,000. Total residential relationships swung from +10,000 in Q1 to -230,000 here, with international reversing from +104,000 to -64,000. Wireless remains the standout, crossing 10 million lines with a second consecutive record and year-to-date net additions up 25%.

Note that management reframed the wireless penetration metric this quarter. Prior releases cited penetration of the domestic residential broadband base, which now stands at 17%. This release leads with 7% penetration of "total addressable wireless lines in our footprint," a larger denominator that makes the runway appear longer. Both figures are disclosed, and the recap uses the 17% figure for continuity with prior quarters.

Content & Experiences

Media and Peacock: the milestone

Media revenue grew 25.3% to $5,691 million, or 15.6% excluding the FIFA World Cup, with EBITDA up 3.7% to $708 million while absorbing the final quarter of the first full year of NBA rights. The headline is Peacock, which turned profitable for the first time.

Peacock revenue grew 54% with distribution revenue up over 50% and advertising up nearly 70%, driven by the World Cup simulcast from Telemundo, the NBA playoffs and the latest season of Love Island. Management reported Peacock's biggest viewership month ever in June.

"Turning to Media. We achieved an important milestone as Peacock reached profitability for the first time, generating $189 million of EBITDA in the quarter. This reflects the strategy we have been executing for several years, building Peacock around a dual revenue model supported by a broad content mix across sports, next-day NBC and Bravo, film, originals, news, library and major events."
— Jason Armstrong, Chief Financial Officer

Assessment: This is the vindication of a six-year investment and the single best thing in the print. The subscriber trajectory across our coverage runs 41 million flat for three quarters, then 44, 46 and now 48 million, with two consecutive quarters of 2 million net additions. Crucially, the retention concern we raised in January has been answered: subscribers acquired around Legendary February were still there when the World Cup cohort arrived. Management was careful to caution that quarterly profitability will vary with sports scheduling and that the annual trajectory is the right frame, which is the correct and honest caveat.

Studios

Revenue grew 25.0% to $3,040 million with EBITDA up $141 million to $202 million, driven by theatrical performance from the Super Mario Galaxy movie, Obsession and international distribution of Michael. The Super Mario Galaxy movie has grossed over $1 billion worldwide year to date, taking the franchise past $2 billion. Obsession crossed $400 million and became Focus Features' highest-grossing film. Minions & Monsters took the Minions franchise to $6 billion globally.

Assessment: An exceptional slate year, and notably this quarter's strength is theatrical rather than the intercompany licensing that flattered Q1. That makes it higher quality than the prior quarter's doubling. Studios remains inherently lumpy and management said so again, but a franchise portfolio producing a $1 billion title, a studio-record specialty film and a $6 billion animation franchise in the same quarter is a genuine asset, and it is one of the two crown jewels moving into the new NBCUniversal.

Theme Parks: the deterioration

Revenue grew 2.7% to $2,413 million while EBITDA fell 5.1% to $609 million. Management attributed the decline primarily to Osaka, where China-related travel restrictions continue to weigh on attendance, partially offset by growth at the US parks. The more significant disclosure was about Orlando.

"At the same time, attendance across the broader Orlando market began to soften in June, and that trend has continued into the third quarter. We believe there are some temporary factors at work, including higher fuel prices and weaker consumer sentiment, but we are watching these trends closely."
— Mike Cavanagh, Co-Chief Executive Officer

Management was explicit that Epic Universe itself continues to perform to expectation, with strong guest response and higher per-capita spending, and that the softness is a broader Orlando market demand issue rather than an Epic-specific one. Hollywood is improving as it laps prior-year pressure, with a Fast & Furious coaster opening later this year.

Assessment: This is the most consequential negative in the quarter after the buyback pause, and it is the first time in our coverage that Theme Parks has declined. We identified this segment at initiation as the cleanest growth story in the portfolio and the pillar most likely to outperform, and it has now posted a year-over-year EBITDA decline with management flagging continued softness into a quarter already underway. The distinction between Epic performing well and Orlando softening is real and worth crediting, but from an earnings standpoint it does not matter which explanation is correct: the segment moving to NBCUniversal carries a demand question into the separation.

Key Topics & Management Commentary

Overall Management Tone: Confident and forward-leaning on the separation, measured on the operating quarter, and structurally less accountable than in any prior call because live analyst questioning was replaced by pre-submitted questions read by investor relations. The prepared remarks continued the pattern of volunteering unflattering detail, disclosing the Business Services one-off and the Orlando softness before being asked. But the absence of follow-ups meant every answer stood unchallenged in a quarter that reversed a major strategic position and suspended the buyback.

1. The Separation: A Reversal Within Five Months

On June 29, Comcast announced its intention to separate into two public companies via a tax-free spin-off of NBCUniversal and Sky, targeted for completion in approximately one year. Mike Cavanagh becomes CEO of NBCUniversal; former Comcast CFO Michael Angelakis returns as CEO of Comcast; Brian Roberts remains involved in the leadership of both. Comcast expects to retain up to a 19.9% stake in NBCUniversal for up to a year post-spin, to be monetized tax-efficiently.

The reversal is stark. On the January 29 call, asked directly whether structural separation would unlock value, Mike Cavanagh said Comcast did "not really see that there is a strategic advantage or making NBCUniversal stronger by separating it from the cable side of the house or putting it outside of Comcast Corporation."

"Before Mike and Jason take you through the quarter, I'd like to spend a few minutes on the separation we announced 3 weeks ago. Since then, we've talked with our key constituencies, employees at every level and most of our key partners, and the reaction has been overwhelmingly positive. I feel more positive and energized today than I was on the day we announced it."
— Brian Roberts, Chairman and Co-Chief Executive Officer

Assessment: We take no position on whether the separation is right; a focused connectivity company and a focused media company is a defensible structure and the market has rewarded similar moves elsewhere. What matters for an analyst is that a clearly stated strategic position was reversed within five months with no intervening event that management has identified. In our January note we wrote that investors holding for a structural catalyst should "reprice that expectation to roughly zero" on the strength of that statement. Anyone who did so was penalised. The lesson we are carrying forward is to discount categorical strategic denials from this management team.

2. The Buyback Is Paused, Removing the Equity's Support

Comcast repurchased $900 million of stock in the quarter, down from roughly $1.3 billion in Q1, and then suspended the programme entirely.

"As we recently announced on our separation call, we did pause share repurchases as of July 1 and expect to remain paused through our separation. Our priority is to ensure both businesses are well capitalized with favorable investment-grade ratings to provide both the financial foundation to pursue their respective growth strategies."
— Jason Armstrong, Chief Financial Officer

Assessment: This is the most important fact in the quarter for a shareholder. Comcast has reduced its share count by roughly 4% to 5% annually for years, and that mechanical support has flattered EPS through the entire earnings decline we have documented. It is now gone for approximately a year. The rationale is legitimate, since both entities need investment-grade balance sheets and the separation requires financing flexibility. But the practical effect is that an equity trading at a 6.0% dividend yield with declining EBITDA has lost the one lever that converted cheapness into per-share value. The buyback pacing across our coverage reads $1.7B, $1.5B, $1.5B, $1.3B, $0.9B, zero.

3. Broadband Reverts, Confirming the Q1 Caveat

Residential broadband losses of 167,000 improved 34,000 year over year but deteriorated 102,000 sequentially from Q1's 65,000. This is precisely what management's own Q1 disclosure implied, when they attributed over half the prior quarter's 117,000 improvement to Legendary February promotions.

"Broadband subscriber losses improved by 34,000 year-over-year to a loss of 167,000. That improvement reflects continued traction from our new go-to-market strategy even as we continue to operate in a highly competitive environment across our footprint."
— Jason Armstrong, Chief Financial Officer

Assessment: The framing is accurate and incomplete. A 34,000 year-over-year improvement is real and is the second consecutive quarter in which the pivot has produced a better result than the year before, after Q3 and Q4 of 2025 both came in worse than their prior-year quarters. It is also a fraction of the improvement that drove our April upgrade, and it is achieved against a prior-year quarter that was itself weak. Roughly 45% of the base is now on gig-plus tiers, and NPS improved again, so the leading indicators continue to point the right way. But at this run rate the base is still shrinking by well over half a million subscribers a year while the price per subscriber falls.

4. Broadband ARPU Deteriorates Further to Minus 3.8%

ARPU declined 3.8%, worse than Q1's 3.1%, driven by the same three factors: no rate increase, migration to lower everyday price points, and free-line dilution. Management maintained that relief is coming.

"As I said earlier, we expect to see modest improvement as we anniversary the launch of the go-to-market strategy and as free wireless lines begin converting into paid relationships in greater volume as we exit this year."
— Jason Armstrong, Chief Financial Officer

Assessment: The word chosen again is "modest." Management now points to a Q3 start for the EBITDA improvement while leaving the ARPU relief where April put it, at the exit of the year. Convergence ARPA fell 1.5% to roughly $85, so even the broader metric management prefers is now declining. The mechanism by which conversion reverses the dilution is genuine and we described it in April, but the arithmetic requires conversion volume that has not yet arrived at scale.

5. Free-Line Conversion Now Has Supporting Evidence

For the first time, management provided behavioural detail on converting cohorts rather than an adjective.

"And as a result, the significant majority of our free line roll-off customers are converting from free to paid as expected. And a few of those characteristics of the customers is usage is consistent as customers convert to paid. We see porting is comparable to customers who paid from day 1. And our overall churn rate in mobile is down even though we're rolling and accelerating the free line roll-off."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: This is the best answer management has given on the subject across four quarters, and it addresses the specific failure mode we have worried about: that converted customers would behave like promotional churn risk rather than genuine subscribers. Usage holding through conversion, porting comparable to day-one payers, and total mobile churn falling while roll-offs accelerate are three independent, checkable claims that all point the same way. It remains, four quarters on, unquantified. There is still no conversion percentage, and a bare majority-qualifier with no number attached has now carried this claim across four consecutive calls.

6. Wireless Crosses 10 Million Lines

A record 448,000 net additions took total lines past 10 million to 10.187 million, a second consecutive record quarter, with year-to-date net additions up 25%. Roughly half of residential postpaid phone connects came from free lines, premium unlimited plans accounted for roughly 30% of postpaid phone connects, and about one third of line additions came from existing mobile customers adding lines. Management noted gross additions strengthened and churn improved even as the first free-line cohort began rolling into the paid base.

Assessment: Unambiguously the strongest operating trend in the company. The composition matters as much as the number: adding a record while simultaneously converting free lines to paid, with churn falling, is a materially better result than a record driven purely by new promotional intake. At 17% penetration of the residential broadband base the runway remains long. Our reservation is scale rather than quality, since wireless service revenue of $1,007 million cannot yet arrest a broadband line that shed $369 million of revenue year over year.

7. Satellite Enters the Competitive Frame Properly

Starlink drew the most substantive competitive discussion of any call in our coverage. Management characterised it as not currently meaningful but expected to matter more over time.

"So on Starlink specifically, I would say right now, we currently are not seeing them as a meaningful competitive factor in our markets. But saying that we don't take any comfort in that. We're not complacent. We fully expect to see them as more of a competitor over time."
— Jason Armstrong, Chief Financial Officer

Management expects satellite to concentrate in rural and underserved areas, restated the thesis that the market resolves to two multi-gig wires per home, and disclosed that Comcast already partners with Starlink inside Comcast Business, combining managed connectivity with satellite capabilities for enterprise customers.

Assessment: The measured framing is credible and matches how fixed wireless actually developed, which management called correctly two years ago. The existing Comcast Business partnership is a useful disclosure that reframes satellite as partly complementary. The competitive picture nonetheless now has three vectors where it had one, and management explicitly said promotional intensity has increased "and some of the behavior has been irrational."

8. Sky Moves to Acquire ITV's Media and Entertainment Business

Management disclosed that Sky has proposed acquiring ITV's media and entertainment business, pairing Sky's content, connectivity and sports position with a broadcaster reaching 40 million people weekly in the UK and serving more than 16.5 million digital users. Management cited enhanced streaming and advertising capability, operating efficiencies and broader customer relationships.

Assessment: Consistent with the separation logic, since a standalone NBCUniversal with Sky needs European scale, and it validates the argument that the spin gives the media business acquisition currency. It also means that the entity being separated is simultaneously being enlarged, which adds execution complexity to a transaction already targeted a year out. For a Comcast holder today, this is a use of capital inside a business they are about to be handed shares in rather than one they can evaluate separately.

9. The Network Argument: Upstream Traffic Growing 2.5x Downstream

Brian Roberts made the most specific technical case of the coverage period, disclosing that while downstream broadband traffic grew about 10%, upstream traffic grew at 2.5 times that rate, which he attributed to AI query patterns.

"This quarter, broadband downstream traffic was up same as in the past couple of quarters, about 10%. But the upstream traffic was 2.5x that volume growth. And that's being driven, I think, by rapidly changing AI queries."
— Brian Roberts, Chairman and Co-Chief Executive Officer

Assessment: This is the strongest structural argument in the bull case and it is quantified for the first time. Upstream capacity is precisely where fixed wireless and satellite are weakest, and where Comcast's mid-split and multi-gig symmetrical roadmap is most differentiated. If AI workloads genuinely reshape upstream demand, the capacity-constrained alternatives become less substitutable for high-value households. The caveat is that this is a multi-year thesis being offered in a quarter where subscribers declined and price fell.

10. Live Q&A Replaced by Pre-Submitted Questions

This call did not have a live question-and-answer session. Questions were collected from analysts in advance and read aloud by investor relations, with no follow-ups and no ability for questioners to press an answer.

Assessment: A meaningful regression in disclosure quality, and the timing is unfortunate. This is the first call after a major strategic reversal and a buyback suspension, and it is the call on which management faced the least scrutiny. Several answers that would ordinarily have drawn immediate follow-ups went unchallenged: no one could press on why the January position changed, on what conversion rate "significant majority" denotes, or on how long Orlando softness is expected to persist. The questions selected were substantive and the roster of contributing analysts was broad, so this is not a case of screening out hard topics. But format determines depth, and this format removed it.

Guidance & Outlook

Implied trajectory: Management expects the third quarter to be the first of the recovery, with modest improvement in both Connectivity EBITDA and broadband ARPU. Against that, Theme Parks softness is already confirmed to be continuing into Q3, the buyback contributes nothing, and the free-line conversion tailwind builds through the second half. The forward table is more balanced than a quarter ago but the two largest uncertainties, the dividend policy of each post-separation entity and the exchange ratio, are explicitly unresolved.

Guidance style: Four consecutive quarters of accurate directional pre-announcement, and this quarter added the disclosure of two items that cut against the headline (the Business Services one-off and the Orlando softness). The operating guidance remains high-information. The strategic guidance does not: a January statement ruling out separation preceded a June announcement of separation by five months.

Analyst Q&A Highlights

Questions on this call were submitted in advance and read by investor relations. The exchanges below reproduce the questions as read and management's responses; no follow-ups were possible.

Competitive Dynamics in the Broadband Market

The opening question asked management to characterise the competitive environment. The response confirmed intensifying pressure across all three vectors and, notably, described some competitor behaviour as irrational.

Q: "We're going to start with broadband. The first question comes from John Hodulik of UBS. Can you discuss the competitive dynamics in the broadband market?"
— John Hodulik, UBS (read by investor relations)

A: "Competition remains intense, and we are operating under the assumption that the market will become increasingly competitive. And that's through continued fiber expansion, aggressiveness will remain in fixed wireless. And although not meaningful to us now, satellite is emerging as another competitor in that space. We've also seen the promotional intensity increase in some markets and some of the behavior has been irrational."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: The phrase "some of the behavior has been irrational" is the most candid competitive language of the coverage period and is the clearest signal that pricing pressure is not abating. Management restated the two-multi-gig-wires end state and positioned Comcast as one of the two, which is the same structural argument made at initiation. Nothing in the answer suggests the environment that forced the pivot is improving.

Whether Starlink Is a Rural Solution or Fixed Wireless 2.0

The most-submitted topic of the call was satellite. Management separated the near-term reality from the medium-term expectation and disclosed an existing commercial relationship.

Q: "There is a robust debate as to whether Starlink broadband is purely a rural solution or whether it's fixed wireless 2.0. How do you see it?"
— Craig Moffett, MoffettNathanson (read by investor relations)

A: "I think if you look at their trajectory in terms of capacity that they're bringing online, which is I think everybody is pretty well versed in at this point, they're going to have significant capacity increases in the coming years, particularly expect to see them more in rural and underserved areas. I think that's a core target market for them."
— Jason Armstrong, Chief Financial Officer

Assessment: The answer lands closer to "rural solution" than "fixed wireless 2.0," which we think is right on current capacity economics but is a judgment rather than a certainty. The more valuable disclosure was the existing partnership inside Comcast Business, pairing managed connectivity with satellite for enterprise customers, which reframes satellite as both competitor and channel. Management also conceded the network is "a little bit more costly to run" as an active architecture, while arguing that active components end to end become an advantage in an AI-driven world.

Converting Free Lines to Paid and Driving Premium Uptake

The question that has defined this thesis for four quarters finally produced behavioural evidence rather than an assertion.

Q: "Can you talk about your mobile strategy as it relates to converting paid lines and driving premium uptake in the second half of 2026?"
— Peter Supino, Wolfe Research (read by investor relations)

A: "And what we've seen is in this quarter, about 1/3 of our mobile line connects came from existing mobile customers adding an additional line. Additionally, our premium mobile adoption continues to strengthen, and we're now exceeding 30% sell-in."
— Steve Croney, Chief Executive Officer, Connectivity & Platforms

Assessment: One third of line additions coming from existing mobile customers is the single most reassuring wireless statistic available, because customers who voluntarily add a second or third paid line are demonstrating satisfaction rather than responding to a promotion. Combined with premium sell-in above 30%, the mix is improving alongside the volume for a third consecutive quarter. The conversion percentage remains undisclosed, which after four quarters begins to look like a choice rather than an oversight.

Whether ARPU and EBITDA Pressure Abates in the Second Half

The question that matters most for the near-term earnings trajectory asked management to reaffirm the recovery timeline set out in April. Management reaffirmed it, with the word "modest" doing considerable work.

Q: "Prior communication has been for broadband ARPU and C&P EBITDA pressure to begin to abate by the second half of this year. This is due to comping both free mobile lines that roll to pay as well as heavier marketing spend. Is this still the outlook?"
— Steve Cahall, Wells Fargo (read by investor relations)

A: "And so I think we're early stages of having early cohorts roll off and into paying relationships, but we like what we're seeing because the vast majority are actually converting. So I think that's one thing that starts to create a tailwind. The second thing is just we're going to start to lap some of the early investments in customer experience. That's not to say we're all the way there. We're certainly not. We've got more investments to make, but we will start to lap some of the early costs."
— Jason Armstrong, Chief Financial Officer

Assessment: The outlook is reaffirmed but hedged twice, with "modest improvement" as the descriptor and an explicit acknowledgment that customer-experience investment is not finished. The two mechanical drivers, conversion and lapping, are the same ones cited in January and April, so the thesis has not changed; only the arrival date keeps advancing by roughly a quarter each time. Absent a live follow-up, no one could ask what "modest" means in basis points.

Whether NBCUniversal Has Sufficient Scale to Compete Independently

The central strategic question of the separation was put directly: does NBCUniversal have the scale to stand alone, or is the spin partly about creating acquisition currency?

Q: "Do you believe NBCU has sufficient scale to compete and grow on its own? Or is part of the rationale for separation to give it the currency and flexibility to participate in further consolidation?"
— Jessica Reif Ehrlich, Bank of America (read by investor relations)

A: "I do believe that NBCUniversal and Sky do have the heft and the relationships and the operational capabilities to continue to be a major player as an independent, as Brian talked about earlier, player in the media space that's in a great position to partner with others."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: The answer asserts sufficiency and declines the second half of the question, which asked about acquisition currency. That omission is conspicuous given that the same call disclosed Sky's proposed acquisition of ITV's media and entertainment business, which is precisely consolidation using the separated entity's strategic freedom. Management's substantive case rests on partnership rather than scale: an independent NBCUniversal that does not "have any other agenda" can license to and partner with platforms that a vertically integrated Comcast could not. That is a coherent argument and it is different from the one the question asked about.

The Domestic Softness in Theme Parks

A combined question from multiple analysts asked management to decompose the June softness between attendance and per-capita spending, separate Epic from the rest of Orlando, and characterise third-quarter trends.

Q: "Can you expand on the domestic softness you saw in June? What drove the shift in consumer behavior? And what read do you have on third quarter trends? How much of the softness was attendance versus per cap spending? And can you separate the trends at Epic from those of the other Orlando parks?"
— Multiple analysts incl. Mike Ng, Goldman Sachs; Kutgun Maral, Evercore (read by investor relations)

A: "So in Orlando, really, the softness that we're seeing affecting this quarter that continues as we look into June is essentially weakness in attendance. And that, we believe, is driven by some weakness in consumer sentiment and higher travel costs affecting demand."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: The decomposition was answered clearly: this is attendance, not per-capita spending, which is the better of the two outcomes because pricing power is intact and volume is cyclical. Management separated Epic cleanly, stating it continues to deliver against expectations with strong guest response and higher per-caps, and framed the issue as broader Orlando market demand. The judgment that this is temporary rests on consumer sentiment and fuel costs normalising, which is outside management's control and unverifiable from here. It is the assumption in the quarter we would most want to stress-test, and no follow-up was possible.

Whether Peacock's Profitability Is Durable

The final substantive question asked whether the milestone represents a durable inflection or could reverse with event timing.

Q: "Peacock reaching profitability was an important milestone this quarter. Does that represent a durable profitability inflection? Or could seasonality and event timing push Peacock back into losses during the second half or in future periods?"
— Kutgun Maral, Evercore (read by investor relations)

A: "And as to Peacock profitability itself, we do expect it to continue to improve on an annual basis, but profitability is going to vary quarter-by-quarter just based on the timing of sports schedules and other content hitting one quarter versus another. But I think of it more of as an annual -- think of it on an annual basis rather than the lumpiness quarter-by-quarter, and it's been improving steadily, and we see that continuing to be the case."
— Mike Cavanagh, Co-Chief Executive Officer

Assessment: The honest answer, and it explicitly warns that Peacock can return to quarterly losses. Investors treating the $189 million as a new run rate are misreading it; the quarter contained the World Cup, the NBA playoffs and Love Island simultaneously. Management's framing that the objective was never Peacock profitability in isolation but the profitability of the integrated Media segment is the right lens, and Media EBITDA growth of 3.7% while absorbing first-year NBA costs supports it.

What They're NOT Saying

1. Why the January position on separation changed. No explanation was offered for reversing a categorical statement made five months earlier, and the pre-submitted format meant nobody could ask.
2. Still no free-line conversion rate, four quarters on. A majority-qualifier with no number attached has now carried the claim across four consecutive calls. The behavioural colour added this quarter is welcome and is not a number.
3. No exchange ratio, no post-separation dividend policy, no target leverage for either entity. Asked directly, the answer was "nothing to add relative to what we said 3 weeks ago." Shareholders are being asked to hold through a year of execution without knowing what either resulting security pays.
4. No duration on the buyback pause beyond "through our separation." If the separation slips, so does the pause, and no interim capital-return alternative was offered.
5. How long Orlando softness is expected to last. Management confirmed it continues into Q3 and called it temporary, without a base case for recovery timing or a sensitivity on parks EBITDA.
6. The wireless penetration denominator changed without explanation. The lead metric moved from 17% of the residential broadband base to 7% of total addressable lines in the footprint. Both are disclosed, but the headline now uses the larger denominator.
7. No sizing of the Sky/ITV transaction. Announced as a strategic positive with no price, no financing detail, no expected close and no accretion framework, inside the entity being separated.
8. Business Services deceleration was disclosed but not explained forward. Underlying growth of just under 3% is roughly half the prior run rate; management attributed it to lapping an acquisition without indicating whether ~3% is the new baseline.

Market Reaction

• Pre-print setup: The stock closed at $23.52 on July 22, down 16.0% year to date against an S&P 500 up 9.5%, and down 34.3% over trailing twelve months. The 52-week closing range of $22.32 to $35.79 shows the shares had already made new lows in the weeks after the June 29 separation announcement.
• Reaction session (July 23, before-open reporter): Opened at $24.01, up 2.1% on the headline beat, then reversed for the entire session to close at $21.92, down 6.8% or $1.60. The intraday range of $21.91 to $24.01 means the high was the open and the close was effectively the low.
• Volume: 53.0 million shares against a 38.7 million thirty-day average, 1.4x normal.
• Benchmark: The S&P 500 fell 1.2%, so the majority of the move was company-specific.

The market bought the release and sold the call. The opening gain reflects the press release: an EPS and revenue beat, Peacock's first profit, a record wireless number. The reversal reflects the 8:30 call, which added Orlando attendance softening into a quarter already underway, the confirmation that buybacks are suspended through the separation, Connectivity EBITDA down 5.8%, and the disclosure that Business Services growth was flattered by a one-off. Closing at the session low on 1.4x volume is a comprehensive rejection of the print's headline.

Note on published accounts of this session. Several outlets reported Comcast shares climbing roughly 3% on the day. Those describe pre-market and early-session trading. The close-to-close move was negative 6.8%, the largest single-day decline across the four quarters we have covered.

Context for the level. At $21.92 the shares are below the bottom of their pre-print 52-week closing range, down 34.3% over twelve months, and down 30.7% from the $31.64 close on the day we upgraded in April. The equity now yields 6.0% on a dividend that management has frozen, with no buyback and a year of structural execution ahead.

Street Perspective

Debate: Does the Separation Create Value or Destroy Optionality?

Bull view: The bull case on the Street holds that the conglomerate discount has been the single largest drag on this equity for a decade, and that separating a cash-generative connectivity utility from a growth-and-IP media business lets each attract its natural shareholder base. Comcast becomes a broadband and wireless compounder valued on free cash flow; NBCUniversal becomes a parks, studios and streaming asset valued on IP and growth.

Bear view: The bear camp contends the separation destroys precisely what Legendary February demonstrated: the ability to market connectivity products on the back of the largest media reach in America, which management itself called a proof point that moving the needle requires the full company. It also loads a year of execution risk, suspends the buyback, leaves both dividend policies undefined, and hands shareholders a media company carrying a softening parks business and an ITV acquisition they cannot yet evaluate.

Our take: The structural logic is sound and we are not against the transaction. The problem is sequencing. Comcast is separating at a twelve-month low, with the connectivity business mid-pivot and not yet at its EBITDA trough, and parks softening into the spin. Separations are best executed from strength, when each entity can be valued on a clean run rate; this one arrives when neither can. The bulls will likely be proved right about the end state and the bears about the next twelve months, and only the second of those is investable today.

Debate: Is Broadband Stabilising or Simply Declining More Slowly?

Bull view: A growing consensus view is that two consecutive quarters of year-over-year improvement, roughly 45% of the base on gig-plus tiers, improving NPS, record wireless attachment and a free-line conversion cohort behaving like genuine subscribers together describe a business that has found its floor and is rebuilding on better foundations.

Bear view: Some sell-side desks argue that a 34,000 improvement on a 201,000 base loss is noise, that the 167,000 lost this quarter is worse than the 131,000 residential customers lost in the fourth quarter of 2024 before the pivot began, and that Comcast has surrendered 5.5% of domestic broadband revenue and 8% of Residential EBITDA to slow a decline it has not stopped.

Our take: The bears have the better of it on current evidence, and we said the opposite three months ago. The honest scorecard is that broadband revenue fell 5.5%, Residential EBITDA fell 8.0%, ARPU fell 3.8%, and the subscriber base still shrank by 167,000. Against that, the leading indicators genuinely are improving and the conversion mechanism genuinely does reverse ARPU dilution. The distinction that matters is between a floor and a slower slope, and Q3 is the first quarter where management has promised improvement without an event calendar to help.

Debate: Is a 6% Yield With No Buyback a Value Opportunity or a Value Trap?

Bull view: The bull argument is that at 4.7x trailing EV/EBITDA, 5.8x trailing adjusted earnings and a 6.0% dividend yield covered several times over by $4.6 billion of quarterly free cash flow, the market is pricing an outcome worse than anything in the actual numbers. Free cash flow grew year over year this quarter. The separation is a catalyst that forces the market to value the parts.

Bear view: The bear camp responds that cheap gets cheaper when there is no mechanism to close the gap. The buyback is suspended, the dividend is frozen, the post-separation payout policies are undefined, EBITDA is still declining, and the trailing multiples are flattered by denominators containing Versant and Sky Germany. A 6% yield on a shrinking earnings base is compensation for risk, not a bargain.

Our take: Both sides are describing the same fact from different ends. The valuation genuinely is the cheapest in our coverage of this name and the free cash flow genuinely does cover the dividend with room. But the bears identify the binding constraint: with the buyback gone, the only routes from cheapness to return are multiple re-rating or dividend growth, and neither is available until the separation completes and both entities publish capital-return policies. That is a year away. A 6% yield while waiting is a reasonable outcome and it is not an Outperform.

Model Update & Valuation Framework

How our April assumptions performed

Five of eight assumptions were too optimistic and one is unresolved, while the two we got right (wireless and Peacock) are the smaller contributors to consolidated EBITDA. This is the record of a rating that was wrong, and the pattern is instructive: we correctly identified which businesses were improving and materially underestimated how much the declining businesses would keep declining.

Revised assumptions

Where the valuation sits

Trailing-multiple caveat. The TTM denominators include Versant in the Q3 and Q4 2025 quarters and Sky Germany through May 2026, while the enterprise value is current. Versant's disclosed adjusted EBITDA is $834 million in Q1 2025 and $789 million in Q2 2025, neither of which falls inside the trailing window used here; Sky Germany's EBITDA contribution was described as immaterial. The multiples above are therefore flattered, and we do not attempt to estimate the contributions of the trailing quarters Comcast has not broken out. Even adjusting conservatively, this remains a mid-4x to low-5x EBITDA multiple, the cheapest level across our coverage of the name.

Valuation judgment. The valuation is not the problem and has not been the problem at any point in our coverage. At initiation the stock was 4.9x and fell; in April it was 5.6x and fell 30%. What has been missing throughout is a mechanism. Previously that mechanism was the buyback, which shrank the share count 4% to 5% a year and converted a low multiple into per-share growth even while EBITDA declined. That mechanism is now switched off for approximately a year. Dividend coverage is not in question, with $4.6 billion of quarterly free cash flow against roughly $1.2 billion of dividends, but the dividend is frozen and the post-separation policies are undefined.

What we would need to see. The path to Outperform runs through evidence rather than valuation: Connectivity EBITDA decline narrowing in Q3 and again in Q4 as management has promised, a disclosed free-line conversion rate, broadband losses inside 120,000 in a quarter without a marketing event, and clarity on both post-separation dividend policies. Several of those could arrive within two quarters.

Thesis Scorecard Post-Earnings

Overall: Thesis materially weakened, and in the specific places our April upgrade assumed improvement. Both bear pillars we de-escalated last quarter have moved against us, one re-escalating to Materializing and the third escalating for the first time. Bull 3 is now Broken rather than At risk, because a valuation floor without a capital-return mechanism is not a floor. Bull 1 remains genuinely strong and is the reason this is a Hold rather than an Underperform.

Action: Downgrading to Hold from Outperform. The trigger we set in April, Q2 residential broadband losses widening beyond roughly 150,000, fired at 167,000. We are honouring it. We are not moving to Underperform because free cash flow grew, dividend coverage is ample at a 6.0% yield, wireless is compounding with improving quality, Peacock has crossed into profit, and the separation, whatever its timing, does eventually force a valuation of the parts.

Revised triggers. Back to Outperform on: Connectivity EBITDA decline narrowing in both Q3 and Q4; or a disclosed free-line conversion rate above 55% with retained ARPU; or broadband losses inside 120,000 in a quarter with no marketing event. To Underperform on: broadband losses exceeding 200,000 in any quarter; Theme Parks EBITDA declining for a third consecutive quarter; a separation delay beyond the roughly one-year target; or either post-separation entity signalling a dividend below the current combined payout.

Bottom Line

We upgraded Comcast in April on the first broadband improvement since 2020 and the passing of peak NBA dilution. The shares are down 30.7% since. Management told us at the time that over half of that broadband improvement came from a promotional moment that would not repeat, and this quarter proved them right: losses widened from 65,000 to 167,000, and the year-over-year improvement shrank from 117,000 to 34,000. We set a downgrade trigger at 150,000. It fired at 167,000, and we are honouring it rather than arguing around it.

The larger issue is the one we cannot model. On January 29 management said there was no strategic advantage to separating NBCUniversal from the cable business. On June 29 they announced exactly that, and paused the buyback with it. We wrote in January that investors holding for a structural catalyst should price that expectation at roughly zero. That was the wrong conclusion drawn from a statement that turned out to be soft, and the practical consequence is that this equity has lost the share-count mechanism that made a low multiple work while earnings declined.

What is genuinely good here is real and deserves saying plainly. Peacock turned profitable for the first time in six years, at $189 million, while absorbing the last quarter of first-year NBA costs, and did it on 48 million subscribers with retention holding across two event cycles. Wireless crossed 10 million lines with a second consecutive record, improving premium mix, one third of additions from existing customers, and free-line conversion now supported by evidence rather than adjectives. Free cash flow grew. Studios had an exceptional slate. These are not small things.

They are simply not yet large enough. Residential Connectivity still produces roughly three quarters of consolidated EBITDA, and it declined 8.0% with revenue down 4.0% and ARPU down 3.8%. Theme Parks, the pillar we thought most likely to outperform, posted its first decline with Orlando softness already confirmed into the current quarter. At 4.7x trailing EBITDA and a 6.0% yield, Comcast is priced for a poor outcome and covered by cash flow, which is why we are at Hold and not lower. But cheapness needs a mechanism, and for the next year Comcast does not have one. We would rather be paid 6% to wait for the third quarter to show whether the trough is real than pay for the possibility again.