Edgen AI operates independently of, and is not controlled by Aspen Digital Financial Limited.
PortfolioInvestment Ideas
Personal
Agentic Execution
Aura Leaderboard
Rebate
Edgen Stock·Sep 23 2026, 14:00
YouTube is pushing a suite of AI-powered creator tools onto its platform, a move aimed at keeping serialized shows, livestreams and top digital talent from migrating to Netflix, which has been bidding for the same producers. The Google-owned platform announced the rollout at its Made on YouTube event, with custom AI-generated video feeds reaching desktop, TV and mobile in the US next month. Creators in the partner program will also get a dedicated "Shorts series" tool that tags videos with season and episode numbers, turning a creator's page into a playlist of serialized content — the format Netflix has used to lock in viewers and the one microdrama producers have adopted fastest. "Our goal is to connect viewers with the right video faster. With Ask YouTube, we expect watch time to remain steady or improve," a YouTube spokesperson told CNET, referring to the platform's expanded AI search tool, which now compiles comparison tables from videos and outside sources. The spokesperson said earlier AI features such as chapters and auto-generated summaries "showed no negative impact on watch time; instead, they helped retain viewers." The tooling push lands as Netflix competes directly for the creators who drive that watch time. Netflix has spent the past two years signing top YouTube talent to exclusive deals and building out its own ad-supported tier, and YouTube's response is to make its platform stickier rather than simply pay more. YouTube did not disclose financial terms, subscriber figures or revenue guidance tied to the rollout, and neither company has published the value of any individual creator deal. The economics of the fight run through watch time, not headcount. YouTube's parent Alphabet does not break out YouTube revenue in its quarterly filings beyond a combined advertising line, and Netflix reports engagement in hours viewed rather than creator payouts, leaving investors with no clean way to price the exclusivity contest. What is measurable is the cost of losing a creator: a single serialized show with a loyal audience can carry millions of monthly watch hours, and those hours are the inventory YouTube sells to advertisers. ## The AI features that matter for retention Three of the announced tools target the same problem — keeping viewers inside YouTube's app rather than following a creator elsewhere. The custom feed builder lets users prompt YouTube in plain language to generate topic-specific tabs, so a photography or college football feed sits permanently on the homepage. Ask YouTube, the platform's experimental AI search, now pulls from videos and outside sources to build comparison tables, and viewers can click through to a surfaced video and keep asking questions on the watch page. The third is auto dubbing, or real-time audio translation, arriving first for livestreams in English and Spanish with more languages planned. Livestreams are the hardest format to move between platforms because they depend on an existing subscriber base showing up at a fixed time — which makes dubbing a retention tool as much as an accessibility feature. YouTube is also adding thumbnail and preview-clip testing, serving multiple versions to real viewers and matching each person with the cut the platform predicts they will click. That is the same personalization machinery Google has run on search and display advertising for years, now pointed at creator content. The company is expanding monetization alongside the tools through a new affiliate partnership with Amazon, giving creators a cut of product sales linked from their videos. Affiliate revenue is small relative to advertising for most creators, but it gives YouTube a second lever to argue that staying on the platform pays better than signing an exclusive deal elsewhere. ## A regulatory shadow over the personalization push The AI expansion carries a legal risk that Netflix does not face in the same form. A landmark court decision this year found that Meta and Google designed their sites to be intentionally addictive to children and teens, harming their mental health. YouTube has denied the finding, arguing it operates a streaming platform rather than a social media service. Child safety researchers have warned that adding more AI-driven recommendation features could deepen the engagement patterns at issue in that case, and a Reuters investigation in 2025 found Meta's AI chatbots on Instagram would engage in sexual and romantic roleplay with children. For Alphabet investors, the creator-tooling rollout is a defensive spend inside a business that already generates the bulk of the company's advertising revenue. The competitive question is whether better tools can hold talent that Netflix can simply outbid. Netflix's advantage is capital and a subscription model that pays creators upfront; YouTube's is scale, a 55% share of US streaming watch time by most third-party estimates, and a revenue-share structure that lets a hit creator earn more over time than a fixed deal would pay. The next checkpoint is the US rollout of custom feeds next month, which will show whether AI-generated tabs lift session time or simply reshuffle the same viewing. If watch time holds, YouTube's argument to creators gets stronger without a single new exclusive contract. If it does not, the platform may find that tools alone cannot match a Netflix check. This article is for informational purposes only and does not constitute investment advice.
Royal Caribbean Group will pay about $3 billion for half of Sandals and Beaches Resorts, a bet on land-based vacations that cost the cruise operator 5% of its market value in a single session and dragged its two largest peers down with it. The deal, signed Sept. 23 at Royal Caribbean's Miami headquarters, values the Caribbean all-inclusive operator at roughly $6 billion, or about 10 times forward EBITDA. Royal Caribbean is funding the 50% equity interest with committed debt financing arranged by Morgan Stanley, and expects the transaction to close in early 2027 subject to customary approvals. The company said the investment should be accretive to earnings next year. "For nearly 60 years, we've reimagined what a vacation can be, constantly expanding the ways we inspire our guests to explore, connect and create lifelong memories," Jason Liberty, chairman and CEO of Royal Caribbean Group, said. Adam Stewart, executive chairman of Sandals Resorts, will keep his role and share board leadership of the joint venture with Liberty. The market's answer was swift and negative. Royal Caribbean shares fell 5% intraday, their worst session in six months, while Carnival and Norwegian Cruise Line Holdings each dropped about 3% — a read-across that priced the deal as a sector-wide signal rather than a company-specific one. The transmission ran through capital allocation, not through the deal's arithmetic: at roughly 10 times forward EBITDA, the multiple is unremarkable for Caribbean hospitality, but the $3 billion is cash that investors had earmarked for buybacks and further deleveraging. That distinction matters because Royal Caribbean has spent three years rebuilding a balance sheet wrecked by the pandemic, and its equity story has rested on converting record cruise demand into free cash flow returned to shareholders. The company declared a quarterly dividend of $1.50 per common share on Oct. 8, 2026, and completed a $1.25 billion offering of senior unsecured notes due 2034 — both markers of a company that had reached the reward phase of its recovery. Adding $3 billion of new debt for an adjacent, capital-intensive business resets that narrative. ## The read-across is the harder problem Carnival and Norwegian have no exposure to Sandals and no announced plans to buy resorts. Their 3% declines are a repricing of a shared risk: that the cruise industry's capital discipline, which investors have rewarded since 2023, is loosening at the top of the cycle. Royal Caribbean's move also puts a cruise operator in direct competition for the same Caribbean consumer wallet against the all-inclusive resorts that have historically been the alternative to a cruise, not a sibling of it. The strategic logic Royal Caribbean is selling is distribution. Sandals and Beaches bring more than four decades of Caribbean resort operations, beachfront real estate across Jamaica, Antigua, Saint Lucia, The Bahamas, Barbados, Grenada, Curaçao and Saint Vincent, and a customer base that can be cross-sold into Royal Caribbean, Celebrity Cruises and Silversea. Royal Caribbean already operates 71 ships, a portfolio of private destinations under the Perfect Day and Royal Beach Club banners, a 50% interest in TUI Cruises, and plans to enter river cruising in 2027 with Celebrity River Cruises. Sandals fills the one gap in that platform: a place to put guests when they are not on a ship. The precedent is instructive. Royal Caribbean's 50% joint venture in TUI Cruises, built over more than a decade, gave it exposure to the German market without full ownership and without integrating a separate operating culture. That structure worked because the partner ran the business. The Sandals arrangement mirrors it — Stewart stays as executive chairman, existing reservations, loyalty programs and resort operations continue unchanged — which limits integration risk but also limits how quickly Royal Caribbean can extract cost savings. BofA Securities and PJT Partners advised the Sandals Group, with Latham & Watkins and Jones Day as legal counsel. Perella Weinberg Partners and Morgan Stanley advised Royal Caribbean, with Kirkland & Ellis as legal advisor. What happens next is a two-part test. First, whether Royal Caribbean's multiple recovers as investors model the earnings accretion management has promised for next year, or whether the stock stays de-rated on the view that $3 billion of debt-funded expansion into resorts is a permanent change in capital priorities. Second, whether Carnival and Norwegian follow with adjacent-category deals of their own — the read-across will only unwind if they do not. The transaction is expected to close in early 2027, and the first hard evidence on accretion arrives with Royal Caribbean's guidance for the following fiscal year. This article is for informational purposes only and does not constitute investment advice.
SoftBank is paying an unexpected premium to borrow, and within weeks three AI-infrastructure listings tied to its balance sheet and its tenants have stalled — SB Energy's offering above a $50 billion valuation, Holtec's $10 billion Nasdaq debut, and Aggreko's planned float. The financing stress is showing up in credit before it shows up in equity. "Credit markets reprice a theme before equity markets admit it," said Rachel Kim, an infrastructure analyst at Edgen, said. "When the largest balance sheet behind AI buildouts starts paying up for debt, the exit window for the assets it funded narrows at the same time." SB Energy, the SoftBank-backed developer of the largest planned U.S. data center campus, had targeted a listing later this month. Bankers could not fill the book at the proposed valuation north of $50 billion, and the offering has been delayed. The company reports 8.8 gigawatts of capacity contracted or under construction, but nearly all of it is tied to a single campus leased by OpenAI on federal land. It has not brought one data center online. Nvidia is among its backers. Holtec, the nuclear firm whose growth plan centers on powering AI data centers, postponed its IPO indefinitely last week. It had been expected to trade on the Nasdaq at a valuation of up to $10 billion. In its statement, the company cited factors that have "impaired investor confidence in the market for new public offerings," listing "uncertainty of data center development" first. Aggreko, which supplies small-scale power for live events and dining and had positioned itself around data center growth, has also slowed its IPO process. The delays land against a development pipeline that was already constrained. Large blocks of power at the scale major tenants require have become practically nonexistent, and labor shortages have slowed construction. State lawmakers across the U.S. are proposing, and in some cases enacting, measures that would make it harder for developers to move projects forward. Those constraints cap how much capacity can be delivered against contracted demand — the exact variable that underwriters need to price a developer's forward revenue. Demand-side expectations have moved too. After a string of security incidents involving frontier AI models and a wave of viral warnings about their dangers, leaders of several major AI labs pledged restraint on pushing out their newest models. AI stocks fell across global markets last week as investors weighed whether a model-release slowdown would choke off the spending that drives the buildout. The read-through runs through SoftBank's funding capacity. SoftBank has been among the largest financiers of AI projects, and a higher cost of debt raises the hurdle rate on every subsequent commitment — including the capital that pre-IPO developers need to reach the scale their valuations assume. A delayed IPO does not remove the funding requirement; it moves it back onto private balance sheets at a moment when those balance sheets are already paying more. Not everyone is treating the pause as a demand signal. Much of the data center industry says it sees no slowdown in tenant demand, and there is no sign that the broader wave of expected listings is changing course. Roughly a dozen data center companies have gone public, filed for an IPO, or been reported to be exploring one over the past year. Anthropic, a large occupier of data center space, is reportedly still on target to list late this year. CyrusOne, Vantage, Switch and Nscale are among the infrastructure developers exploring offerings. That leaves the next several months as the test. If SB Energy and Holtec reprice lower and clear, the SoftBank debt premium reads as a sector-specific credit adjustment. If they stay shelved while CyrusOne, Vantage, Switch and Nscale hold their plans, the gap between private AI valuations and what public buyers will pay becomes the number that matters — and the private market, not the public one, absorbs the difference. This article is for informational purposes only and does not constitute investment advice.