# left-over — Full Article Content Index > Complete article text for AI indexing. Summary index: https://leftover.io/llms.txt left-over (leftover.io) is an editorial site covering media M&A, ad tech, and capital markets. We publish deep-dive analysis on major deals and their aftermath — the outcomes nobody is writing about yet. --- ## Article 1 of 7 URL: https://leftover.io/article/0 Static text version: https://leftover.io/articles/0.html Category: CTV / M&A Title: Fox Spent $440M on Tubi in 2020. Today It Bought Roku for $22B. Follow the Logic. Deck: The Murdochs just turned a bet most analysts mocked into the biggest streaming deal of the year. Here's the playbook, and why they sold their Roku stake to fund Tubi, then bought Roku back for 50x more. Key figures: Tubi acquisition $440M (2020), Tubi FY2026 revenue $1.5B, Roku acquisition $22B (June 2026) Fox's 2020 acquisition of Tubi for $440 million was one of the most mocked deals in media that year. The streaming wars were at full intensity. Netflix was spending $17 billion on content. Disney had just launched Disney+. And Fox, fresh off selling most of its entertainment assets to Disney, was betting on a free ad-supported service that most analysts described as a niche player for B-movie bingers. What the skeptics missed was the underlying theory. Fox's core business, live sports and news, generates viewers who don't need a subscription stack. They need a place to watch everything else for free. Tubi, with no content production costs and a licensing model, was that place. Fox wasn't competing with Netflix for the prestige subscriber. It was building the platform for the viewer Netflix never wanted. Tubi's numbers proved the bet before anyone gave Fox credit for making it. Revenue approaching $1.5 billion in fiscal 2026. More than 100 million monthly active users. 13 billion hours of content consumed annually. Those metrics don't emerge from a niche acquisition. They emerge from a strategy that understood where ad dollars were moving before the rest of the industry did. Key quote: "Fox didn't buy Roku to get into the hardware business. It bought Roku to own the relationship between the viewer and the advertiser, and to make that relationship impossible to disintermediate." The Roku acquisition announced today closes the loop. Fox sold its Roku stake in 2020 to fund the Tubi deal, betting it could generate more value by owning the content on the platform than by owning equity in the platform itself. That bet worked. Now Fox is buying the platform, because the combination of Tubi's inventory, Fox's sports and news audiences, and Roku's 44 percent share of connected TV viewing hours creates a targeting and measurement capability that no pure-play streaming service can match. The deeper ad tech story is what most coverage is missing. Roku's data on viewing behavior is unparalleled. They know what you watch, when you switch, what you skip, and what you complete. Married to Fox's live sports audiences, who are among the most valuable in advertising, that data creates a story about attention and intent that programmatic buyers have been trying to buy for a decade. Fox just built it. Watch: How this affects the Paramount-WBD combined entity's advertising strategy, and whether any streaming platform makes a defensive move on connected TV distribution. --- ## Article 2 of 7 URL: https://leftover.io/article/1 Static text version: https://leftover.io/articles/1.html Category: Ad Tech / Autopsy Title: AT&T Paid $1.6B for AppNexus. Microsoft Shut It Down. Here's the Full Wreckage. Deck: AppNexus was the most transparent programmatic platform in the industry. Eight years after AT&T bought it, it's gone. Not because the technology failed. Because the parent companies did. Key figures: AT&T paid $1.6B (2018), Microsoft paid ~$1B (2022), DSP sunset Q1 2026 Brian O'Kelley built AppNexus with a simple premise: programmatic advertising should be transparent, fair, and open. The exchange should not favor its own buyers. The data should be available to everyone. The auction should clear at a price that reflects real demand. In 2018, when AT&T acquired it for $1.6 billion, the premise was still intact. What followed was eight years of strategic confusion that had nothing to do with the technology. AT&T's theory was that owning the pipes, the content, and the ad tech stack would create an integrated advantage. AT&T customers watching HBO Max on AT&T devices would see AT&T-sold ads powered by AT&T's AppNexus infrastructure. Vertical integration as a moat. It was a reasonable theory in 2018. By 2022, AT&T had sold its WarnerMedia assets to Discovery, spun off DirecTV, and was returning to its core business. The AppNexus thesis had no home. Microsoft acquired the resulting Xandr business, which AppNexus had been rebranded to, for roughly $1 billion in 2022. The rationale was different. Microsoft wanted the sell-side capabilities, the publisher relationships, and the data infrastructure to support its own ad business as it expanded into retail media and connected TV. The buy-side platform that AppNexus had built, the demand-side platform that thousands of buyers depended on, was an afterthought. Key quote: "The buy-side platform that AppNexus built was the most transparent in the industry. Its shutdown was not a technology failure. It was a casualty of parents who could not agree on what they had bought." The Xandr Invest DSP officially sunset in Q1 2026. Buyers migrated to The Trade Desk, DV360, or Beeswax. Publisher relationships shifted to other SSPs. The technical infrastructure that O'Kelley built — the auction logic, the data pipelines, the bidder optimization that made AppNexus the most studied platform in academic ad tech research — is gone. Not because it stopped working. Because the companies that owned it could not decide what business they were in. What the AppNexus story reveals about the current wave of media consolidation is not comfortable reading. The value in ad tech is not the technology. It is the relationships, the inventory access, and the institutional knowledge that buyers and sellers build over years. That cannot be acquired and rebranded without cost. The cost shows up years later, when the platform is gone and the buyers are explaining to clients why they have to rebuild their audience strategies from scratch. Watch: The migration out of Xandr Invest is not complete. Watch for consolidation among remaining independent DSPs as displaced buyers settle around fewer platforms. --- ## Article 3 of 7 URL: https://leftover.io/article/2 Static text version: https://leftover.io/articles/2.html Category: Legacy Media / Autopsy Title: Comcast Bought NBCUniversal in 2011. In 2026, It's Spinning Half of It Off. That's Your Answer. Deck: The deal was supposed to create an American media juggernaut. Fifteen years later, Comcast is distributing cable network shares to shareholders and pivoting to broadband. The content bet didn't lose. It just stopped being the point. Key figures: 2011 acquisition, 20% US cable penetration in 2026, $81B Comcast NBCU valuation in WBD bid When Comcast closed its acquisition of NBCUniversal in 2011, first 51 percent, then the remainder by 2013, it was acquiring one of the most iconic media portfolios in American history. NBC. MSNBC. CNBC. Bravo. USA Network. Syfy. Universal Studios. Universal theme parks. The thesis was that owning distribution, Comcast's cable infrastructure, alongside content would create durable advantages as media shifted to digital. Here is what actually happened. The theme parks made money. Peacock cost a fortune to build. Cable subscribers kept leaving. And the part of the portfolio that was supposed to generate growth, the cable networks, became the part Comcast is now spinning off entirely. The deal did not fail. The environment changed, and the portfolio that made sense in 2011 required a different structure in 2026. In January 2026, Comcast completed the distribution of Versant, a new standalone company housing the NBCU cable networks except Bravo, to its shareholders. E!, Oxygen, USA Network, SYFY, Golf Channel. Gone from the balance sheet. Handed to investors who can now decide for themselves whether those assets have a future. Comcast retained Bravo because of its relationship with NBCUniversal's streaming and upfront strategy. Everything else was released. Key quote: "The cable network portfolio that was supposed to be the crown jewel is now a separate company. Only 20 percent of US households still pay for cable." Meanwhile, Peacock is quietly working. NBCUniversal wrapped its 2025-26 upfront with record ad sales volume, driven almost entirely by live sports. Super Bowl LX. The FIFA World Cup. The Milan Cortina Olympics. The return of NBA to NBC. Peacock accounted for one-third of total upfront ad commitments. That is not a struggling streaming service. That is a sports rights play starting to pay off fifteen years into a deal that looked expensive when it was signed. The answer to the question of whether the Comcast-NBCU deal worked is not yes or no. The cable portfolio was worth buying in 2011 and is worth less now. The theme parks are worth more than anyone modeled. Peacock is a credible third streaming service built on the back of live sports. And Comcast, freed of its cable network obligations, is a broadband infrastructure company with a media business attached. That is not what the 2011 deal was supposed to produce. It might be exactly what 2026 requires. Watch: Whether Versant finds a buyer within 18 months. If it doesn't, that tells you something about the cable networks' standalone value that Comcast already knew. --- ## Article 4 of 7 URL: https://leftover.io/article/3 Static text version: https://leftover.io/articles/3.html Category: Streaming / Deep Dive Title: Fox Sold Its Roku Stake to Buy Tubi. Then Bought Roku for 50x More. The Trade of the Decade. Deck: In 2020, Fox cashed out of Roku to fund a free streaming bet. Six years later, it bought Roku for $22 billion. The irony isn't lost on anyone. But the strategy was more deliberate than it looks. Key figures: Tubi $440M (2020), Tubi revenue $1.5B (FY2026), 50x scale of the pivot Most people in the industry remember Fox selling its Roku stake in 2020 as a mild curiosity. A company quietly exiting a position to fund a new acquisition. What nobody said at the time: Fox wasn't selling Roku because it didn't believe in Roku. It was choosing where it wanted to own value in the streaming supply chain, and it had concluded that content beat platform. The theory was straightforward. In a world where streaming explodes, must-watch content is worth more than distribution. If you own live sports and news, the platforms come to you. You don't need to own the pipe when everyone needs your water. Tubi, a free ad-supported service with no content production costs, was the bet on that theory. Fox could own the content and distribute it everywhere, rather than own the platform and negotiate with every content provider. Tubi validated the bet faster than anyone projected. The $440 million acquisition hit profitability ahead of Fox's own internal timeline. Revenue approaching $1.5 billion in fiscal 2026. One hundred million monthly active viewers. Thirteen billion hours of content consumed annually. Those metrics justify the Roku acquisition not as a reversal of the original theory, but as its logical extension. Key quote: "Fox didn't buy Roku because the original Tubi theory was wrong. It bought Roku because the theory worked well enough to justify owning both sides of the trade." The ad tech logic is what most coverage is missing. Roku's data on viewing behavior is unparalleled. Connected TV devices generate granular signals about attention, completion, switching behavior, and content preference that no other platform can match at scale. Married to Fox's sports and news audiences, the most valuable inventory in advertising, that data creates targeting and attribution capabilities that programmatic buyers have been trying to construct for years. The irony of the Roku stake sale is real. Fox exited a position that would have been worth considerably more in 2026 than it was in 2020. But the trade was not irrational. The Tubi revenue and the strategic position it created may be worth more than the unrealized Roku appreciation would have been. The more interesting question is whether any other media company has the discipline to execute a trade like this over a six-year horizon without reversing course when the first-year numbers look bad. Watch: Whether Disney makes a defensive move on connected TV distribution in the next 12 months. They are the only major player without a platform position. --- ## Article 5 of 7 URL: https://leftover.io/article/4 Static text version: https://leftover.io/articles/4.html Category: Media / What's Next Title: The Paramount-WBD Deal Is Still Open. Comcast Lost the WBD Bid. Nothing Is Actually Settled. Deck: The media consolidation wave isn't cresting. It's accelerating. Fox just moved. The Paramount-WBD deal is still pending. Comcast is circling. Here's the full board position heading into the second half of 2026. Key figures: Fox + Roku $22B, Paramount + WBD $77.9B pending, Comcast's failed WBD bid $81B If you are trying to understand where media consolidation lands, stop looking at any one deal in isolation. The deals are moves in a game, and you need to see the whole board. Start with what happened this morning. Fox acquires Roku for $22 billion. That creates a free, ad-supported, platform-anchored business reaching every major streaming household in America, with premium live sports and news to fill it with. Fox's theory of media, that distribution-plus-live content beats subscription-plus-library, just got a platform. Every competitor's morning got worse. The Paramount-WBD deal is still working through regulatory review. UK CMA Phase 1 is the current bottleneck. When that closes, you have a combined entity with Paramount+, Max, CBS, CNN, Warner Bros. studios, and Paramount studios. That is the content scale play, sitting opposite Fox's distribution play. The combined entity will have subscription revenue, studio IP, a news network, and a broadcast footprint. It will not have a connected TV platform. Comcast made an $81 billion bid for WBD that lost to the Paramount deal. Comcast is not going away. It still has Peacock, the best live sports portfolio on broadcast television, and a freshly cleaned balance sheet after the Versant spinoff. The most likely next move involves acquiring something that gives it streaming scale. The question is what and when. Key quote: "Three media power structures. Three different theories of how to win streaming. And one major player, Disney, conspicuously quiet." Fox bets on free, ad-supported, sports-anchored content through platform ownership. Paramount-WBD bets on subscription scale and studio IP. Comcast bets on live sports rights and broadband convergence. Disney has the deepest IP library, the theme parks, the best children's brand in media, and a streaming service that has not yet resolved its long-term unit economics. The next major move in media consolidation will come from Disney, or it will be made against Disney, and those are the only two options that remain. The consolidation wave is not cresting. It is accelerating. The companies that end up without a distribution surface, a content portfolio, and an ad revenue stream will not survive independently for long. Watch: Disney's next earnings call. Any change in tone on strategic alternatives or streaming profitability targets signals what comes next. --- ## Article 6 of 7 URL: https://leftover.io/article/5 Static text version: https://leftover.io/articles/5.html Category: Ad Tech / Autopsy Title: Publicis Bought the Identity Graph. The Rest of Adland Has a Problem. Deck: The $2.5 billion LiveRamp deal is quieter than Fox/Roku. It should not be. Publicis now controls Epsilon, Lotame, and LiveRamp: the three largest non-Google identity stacks in advertising. No other holding company is close. Key figures: LiveRamp deal $2.5B (May 2026), Epsilon acquisition $4.4B (2019), 25,000+ publisher domains on RampID When Publicis bought Epsilon for $4.4 billion in 2019, the industry noticed but rationalized it. Epsilon was a first-party data and loyalty business. One acquisition. Manageable. When Publicis picked up Lotame in early 2025, fewer people wrote about it. Lotame is a data management platform. Still manageable. Then on May 17, 2026, Publicis announced it was acquiring LiveRamp for $2.5 billion in an all-cash deal at $38.50 a share, a 30% premium to where the stock was trading. Put the three together and the picture changes. Publicis now controls the identity resolution infrastructure that connects advertisers to consumers across cookieless environments, first-party data clean rooms, and publisher audiences simultaneously. No other holding company, and no independent ad tech player outside Google and Amazon, has that full stack. Key quote: "Publicis made three neutrality commitments at announcement: no service restrictions, no pricing changes, no preferential agency access. Those commitments are not legally binding after close. Ask DoubleClick." LiveRamp is not just another data company. It connects more than 25,000 publisher domains and 500 technology partners across 14 markets. Its authenticated traffic solution is embedded in the login infrastructure of publishers across the industry. The RampID is used by every major DSP and SSP to resolve identity in a post-cookie world. Whoever controls RampID controls a significant portion of how programmatic advertising matches buyers to audiences at scale. The competitive reaction has been immediate. Rivals must decide whether to partner, build, acquire, or accept a weaker position across data, workflow, and transaction layers. Translation: Omnicom, WPP, and IPG are all now in emergency M&A mode. Every clean room company, identity resolution startup, and data collaboration platform just saw its valuation go up. That is the secondary market consequence of this deal that almost no one is writing about. The regulatory path runs through HSR antitrust clearance, foreign antitrust filings across LiveRamp's 14 international markets, and a CFIUS national security review. The CFIUS requirement is not standard for an ad tech deal. It signals that LiveRamp's cross-border flows of consumer behavioral data triggered a flag inside the Treasury Department. Meanwhile, Publicis settled FTC collusion allegations involving agency holding companies as the deal was announced. That FTC settlement is active context for HSR reviewers. LiveRamp will operate as an independent business within Publicis, and CEO Scott Howe will stay in place. Publicis pointed to its handling of Epsilon as the model for how it runs acquired businesses. But Epsilon today is not the independent company it was in 2019. It is deeply integrated into Publicis Sapient workflow and sold primarily through Publicis agency channels. The independence pledge is a starting position, not a destination. Watch: The shareholder vote requires a two-thirds threshold, not a simple majority. LiveRamp's top ten customers represent 30% of revenue. If any of them signal publicly that they are evaluating neutral alternatives post-close, the vote gets complicated and the year-end 2026 timeline slips. The customer defection risk is the story no one is tracking yet. --- ## Article 7 of 7 URL: https://leftover.io/article/6 Static text version: https://leftover.io/articles/6.html Category: Analysis Title: They Traded a $10,000 Subscriber for a $59 One and Called It Disruption Deck: Cable was a forty-percent-margin business with two income streams and almost no churn. Then the people who ran it decided to rebuild it from scratch, at lower prices, with higher acquisition costs, and thinner margins. We are still doing the math. Key figures: Cable LTV at peak $8,000, Streaming avg LTV today $59, Pay TV revenue peak $101B (2014) Cable television was, for about thirty years, one of the greatest money-printing machines in American business history. And then the people who ran it looked at that machine, studied it carefully, and decided to blow it up. On purpose. With their own hands. We are still living in the wreckage. THE MACHINE WAS STUPIDLY GOOD At peak cable, roughly 2010 to 2014, the average U.S. subscriber was paying $90 to $100 a month. Most stayed for five to six years. That works out to somewhere around $6,500 to $10,000 in revenue per household before they ever thought about canceling. And most never did, because leaving cable meant also losing your internet connection, and nobody wanted to deal with that particular afternoon. Cable channels also collected something called affiliate fees, which were per-subscriber payments from the operator just for being in the bundle. ESPN got them. The Food Network got them. The channel that only airs fishing competitions at 2am got them. Every network collected, every month, whether anyone watched or not. It had nothing to do with ratings or ad revenue or whether your flagship show got renewed. Two income streams on a single customer. Captive audience. Low cancellations baked in by sheer inconvenience. Profit margins around 40%. It was almost unfair. Data: Subscriber lifetime value. Cable at peak: ~$8,000. Netflix (best case): ~$1,800. Streaming average today: $59. THEN EVERYONE DECIDED TO COMPETE WITH NETFLIX The thinking, circa 2019, went something like this: Netflix is growing, younger viewers are bailing, the bundle is dying eventually, so let us get ahead of it. Launch a streaming service. Own the subscriber directly. Cut out the middleman. Easy. What the strategy decks missed is that the bundle was never just a distribution tool. It was the thing keeping subscribers from leaving. And when you stepped outside it, you did not just lose the cable bill. You lost the affiliate fee. Permanently. One network executive, when asked about building their own streaming service, said they assumed affiliate fees would not be part of the business model from day one. That quote is doing a lot of work. Data: Monthly churn rate by platform. Cable/satellite: 1.5%. Netflix: 2%. Streaming average (2025): 5.5%. Streaming high end: 10%. THE NUMBERS, SIMPLIFIED A cable subscriber in 2012 was basically a small annuity. They paid reliably, stayed for years, and the network collected ad money on top of that. Lifetime value per customer was well into the thousands. A streaming subscriber today is closer to a first date that ghosts you. Monthly churn across the industry has climbed from about 2% in 2019 to over 5% by early 2025. Nearly half of subscribers cancel at least one service per year. For a platform charging around $9 a month, the average customer is worth roughly $59 before they disappear. That is not a typo. Fifty-nine dollars. The price of a decent dinner. And it costs platforms upwards of $50 just to acquire that subscriber in the first place. So the math on the whole thing is, charitably, tight. DID IT WORK? For Netflix, yes. First mover advantage, years of runway before the competition showed up, and enough habitual lock-in built over time to keep churn low. They are the only major streaming service that is consistently profitable without a bigger company covering the losses. For everyone else, the results have been rough. Disney+ spent years burning through cash. Warner Bros. Discovery took on enormous debt and has spent a good portion of its existence canceling shows it already paid to make. Pay TV revenue peaked at around $101 billion in 2014 and is now projected to fall below $54 billion by 2027. The industry's response to all of this has been to quietly rebuild what they tore down. Ad-supported tiers. Bundle packages. Multiple services discounted together. They are, with considerable seriousness and a steady stream of press releases, reinventing the cable bundle they spent a decade destroying. The original machine is not coming back. But at least we are getting a smaller, more expensive, slightly worse version of it. Data: U.S. pay TV subscription revenue 2014-2027. Peak: $101B (2014). Actual decline to ~$83B (2024). Projected: below $54B by 2027. --- ## Merger Tracker — Live Regulatory Status (as of June 2026) ### Paramount x Warner Bros. Discovery ($77.9B) - DOJ/FCC: CLEARED — Feb 20, 2026. No conditions. - UK CMA: ACTIVE Phase 1 investigation. Current bottleneck for close. - EU: PENDING — filing in process. - State AG: WATCH — California AG active. Hollywood job preservation concerns. - Labor/Unions: WATCH — Teamsters asked DOJ to block Mar 12, 2026. - Business Deal: CLEARED — Structure confirmed Feb 26, 2026. Shareholder votes passed. ### Fox x Roku ($22B) - Announced June 15, 2026. DOJ/HSR review active. Platform neutrality concerns flagged. ### Publicis x LiveRamp ($2.5B) - Close target YE 2026. HSR/DOJ and CFIUS reviews active. Shareholder vote pending (two-thirds threshold required). --- Source: left-over (leftover.io) — The Aftermath Beat