Newsletter · · Ashutosh Agarwal

YouTube Wants to Be Cable and It Is Winning - The Creator Economy - Week of September 19, 2026

The Creator Economy for the week of September 19, 2026: a working creator explained on Decoder why YouTube wants to be the cable company, paying almost nothing for content while becoming the biggest TV service in America, investors argued AI answers are driving more Google search rather than less with Alphabet around 17x earnings, Meta's Muse assistant got its first hands-on review and reviewers would not trust it, and AI video ads fell from 1,000 dollars to as little as 40 while Meta ad prices kept rising.

The Creator Economy

Week of September 19, 2026: YouTube Wants to Be Cable and It Is Winning

TL;DR

  • The most quoted line of the week: YouTube "wants to be the cable company," not the channel. A working creator laid out, in plain terms, how YouTube pays almost nothing, takes a cut of everything, and has quietly become the biggest TV service in America, while Instagram pays creators "no money" and TikTok pays "a fraction of a penny" per view. Great for the platform, brutal for the people making the videos.
  • The "AI is going to kill Google Search" fear officially flipped. A detailed investor breakdown argued the opposite has happened (AI answers are driving more search, not less) and that Alphabet, still trading around 17x earnings after the stock doubled, may be the most underrated megacap. YouTube alone, bought for $1.6 billion, now throws off about $36 billion a year.
  • Meta's new AI assistant "Muse" got its first real-world test drive, and the reviewers wouldn't trust it. Fast, but weaker than a rival, and undone by Meta's reputation: "the damage has so been done... I didn't even bother investigating their agent."

What's new

Five things actually moved a number or a thesis this week. The center of gravity swung hard from Meta (last week's story) to Google and YouTube. They're ranked.

1. A working creator explained, better than any analyst has, why YouTube is the best business in media, and it's not a compliment. The must-listen of the week was Decoder with Nilay Patel (Sep 14), "Adam Conover explains how YouTube ruined everything." Conover is an insider on the creator side: a comedian and former TV writer (Adam Ruins Everything) who now runs his own YouTube channel, so this is an operator talking, not a pundit guessing.

His core point is the one investors should internalize: the platforms have almost no content cost. "Instagram literally will pay you no money," he said. "TikTok, maybe a tiny fraction of a penny for every view that you get. YouTube used to be the gold standard. They just changed their rules. It is now harder than ever to get started in the YouTube partner program and get paid actual money from YouTube." Meanwhile the traditional studios (Warner Bros., Disney) have to pay Hollywood wages to compete with "the army of teenagers who work for free."

The second point is where YouTube's dominance comes from. Conover argues the streaming services chased Netflix into expensive prestige drama and abandoned the cheap, sticky middle of television, "literally 60% of television. News, talk shows, comedy talk shows, lifestyle programming... travel shows, the weather", and let all of it flow to YouTube. That's why, in his telling, YouTube is now the single largest streaming service on actual TV sets. YouTube's deal with creators: "we don't pay for content. We break off a chunk. If we make a lot of money, you get 50%." A 50% ad-revenue share, no production cost, and the biggest living room audience in the country. Why it moves the thesis: this is the bull case for GOOGL stated by someone who resents it: near-zero content cost, structural TV dominance, and a supply of labor that can't stop showing up for free.

2. The same creator gave the sharpest bear case for the whole creator economy: the algorithm is the new gatekeeper, and it's worse than the old one. Still Decoder (Sep 14). Pushed on the "social media democratized everything" argument, Conover flipped it: "The algorithm is the gatekeeper. Of course it is... they have more power than anyone who ever worked for CBS or ABC ever did. And it's actually worse because people aren't in charge of it." His most useful framing for where YouTube is headed: it "wants to be the cable company": it wants creators to be the channels who go get the brand deals and do all the business development, while YouTube is "just the pipe that takes a bunch off the top." The evidence: YouTube is running its own upfronts and trying to connect creators to advertisers, but "without actually putting money or infrastructure or investment into it."

There's a catch buried in here that matters for how big YouTube's ad business can get. The ads a creator like Conover can actually land are "all the DTC brands... MeUndies... meal kits... subscription services. But like where's Coke and Chevy?" The premium brand budgets (the Don Draper money) still haven't fully moved over. Why it matters: if YouTube wants to keep taking TV's dollars, it needs the blue-chip advertisers to follow the audience. They haven't yet. That gap is both the risk and the runway. (Host Nilay Patel also floated a line worth filing away: "influencers are the new gig workers," the regulatory frame that could eventually put a floor under creator pay and a cost onto the platforms.)

3. The "AI will kill Google Search" trade officially reversed, and the numbers behind Alphabet are staggering. On The Investor's Podcast (We Study Billionaires) (Sep 17), investors Kyle Grieve and Shawn O'Malley (pundits, but numerate ones) walked through why Alphabet is "the megacap that still might be underrated." The headline reversal: the fear that ChatGPT would gut Google Search has not played out. "You had this narrative that AI would kill search, but it's just really the opposite is what has manifested," Grieve said. AI Overviews are driving more search volume and can be monetized much like the old blue links.

The supporting figures are worth keeping: YouTube was bought for $1.6 billion and "now generates something like thirty six billion dollars in annual revenue." Alphabet trades around 17x earnings, the same multiple it carried a year earlier, except "we've seen the stock double" since, so the cheapness now reflects doubt about AI-capex payoffs, not competitive fear. And Google Cloud has quietly become a monster: $58 billion revenue last year at a 24% operating margin, growing 63% then 82% year-over-year in the last two quarters, a ~$100 billion run-rate, margins now near 36% (level with Amazon's AWS), and a backlog of signed-but-undelivered contracts north of $500 billion, up roughly five-fold in twelve months. Why it moves the thesis: this is the cleanest articulation yet of the GOOGL bull case, and it directly rebuts the bear thread this newsletter has run for months (that conversational AI would siphon off Google's commercial-intent queries).

The hosts were honest about the other side, too: Alphabet's return on invested capital is "the second lowest that it's been since the business went public," because capital spending is roughly six times 2023 levels, and about $122 billion of assets "not yet in service" may be flattering today's cloud margins. The bull case is real; so is the capex bill.

4. Meta's "Muse" got its first field test, and the testers bounced off it. Last week Meta launched Muse, a consumer AI assistant, and the stock popped 6.5%. This week two AI-product reviewers actually used it. On AI For Humans (Sep 18), the verdict was mixed-to-poor: Muse "is very fast," but "I don't think the results are as strong" as a fast-rising rival agent called Instinct. The deciding factor wasn't the tech; it was trust. Despite Meta's heavy security pitch (Muse runs in a "virtualized encrypted platform," with a "dispatch agent that confirms before anything goes out to the web," and "not even Facebook can see in"), one reviewer's takeaway was blunt: "the damage has so been done to their reputation that despite all of that messaging, I didn't even bother investigating their agent. It was just a non-starter for me." For context on the competition, the rival agent Instinct took $50 million in funding in April and is now "raising at a $10 billion valuation." Why it matters: last week's Muse debate was launch hype and price targets; this is the first data point from the wild, and it lands exactly where the skeptics predicted: the product may be good, but Meta's brand is the tax.

5. AI is industrializing ad creative, and costs are collapsing, even as Meta's ad prices keep rising. A recurring bear thread (AI flooding the ad system with cheap creative) got fresh, concrete numbers this week from people who do it for a living. On Limited Supply (Sep 16), performance marketers (operators) said "static ads are pretty much solved for with AI" and video "is getting there real fast," with a stylized ad live "30 minutes later." The cost math is the story: a solid video ad that used to cost $1,000–1,500 now runs $40–150 with AI, static creative $25–100, "and that cost is only going to go lower." The same theme showed up on The UpFlip Podcast (Sep 14), where a marketer described an AI-generated "grandma talking about selling her house" ad that's "crushing" at a cheaper cost-per-lead than filming himself, because "you can crank them out... test at scale."

Here's the nuance that cuts against the simple bear case, and it's important: the Limited Supply operators also noted that "meta ads are getting more expensive." So creative costs are cratering while Meta's ad prices (what advertisers pay per impression) keep climbing. That's not obviously bad for META; it suggests demand for Meta's targeting is strong enough to absorb an explosion of AI-generated supply. Why it matters: the "AI floods the feed and crushes CPMs" story needs revising. So far, Meta is capturing the value, not losing it.

The debate

The long-running fight is the same: is the money flowing into creators a durable, high-margin tailwind for the platforms, or a slow squeeze from payout compression, share-shift and value leaking off-platform? This week the argument leaned bullish, and unusually, the strongest bull evidence came from a creator who was trying to make the bear case.

Steel-manning the bulls (the platforms have near-zero content cost and are winning the living room). Adam Conover, resentfully, made the point better than any sell-side note: YouTube pays almost nothing for content, takes up to 50% of the ad revenue, has become the biggest TV service in America, and wants to run the whole thing as "the cable company" while creators do the selling (Decoder, Sep 14). Layer on the investor case (The Investor's Podcast, Sep 17): AI didn't kill search, YouTube prints $36 billion a year off a $1.6 billion purchase, and Alphabet trades at ~17x after doubling. And on Meta's side, the AI-ad-supply flood is coinciding with rising ad prices, not falling ones (Limited Supply, Sep 16); the platforms are still the toll booth.

Steel-manning the bears (the people who make the content are getting nothing, and trust is a real liability). The flip side of "near-zero content cost" is that the labor is being paid in exposure. Instagram pays "no money," TikTok "a fraction of a penny," and YouTube "changed the rules" to make getting paid harder (Decoder, Sep 14), a model Conover compares to gig labor before the labor movement, and the kind of thing that eventually draws regulation. On Meta specifically, the Muse test showed the brand-trust deficit is not abstract: reviewers who wanted to like the product refused to hand it their data (AI For Humans, Sep 18). And the litigation tail is still open: a grieving parent who sat through the LA trial reminded listeners the $18 billion settlement is "just one lawsuit," a precedent for many more, and that Meta will keep settling to keep Mark Zuckerberg off the witness stand (The Current, Sep 15).

The honest split: the bulls had the better week, and again for an unusual reason: the clearest articulation of the platforms' economic edge came from a critic. The near-term evidence (YouTube's TV dominance, the search-didn't-die reversal, rising Meta CPMs) favors Google and Meta. The bear case is slower and structural: the value is built on unpaid labor, an unresolved trust problem at Meta, and a litigation overhang that money hasn't closed.

Stocks in play

GOOGL / YouTube, Bull: The strongest week in a while. AI did not kill search; AI Overviews are driving more query volume and monetize like the old links, and Alphabet trades ~17x after doubling (The Investor's Podcast, Sep 17). YouTube ($36B revenue off a $1.6B purchase) is the biggest streaming service on TV sets, pays creators little, keeps up to 50% of ad revenue, and is positioning itself as "the cable company" on top of the internet (Decoder, Sep 14). Bear: Return on invested capital is the second-lowest since IPO on ~6x-2023 capex, and ~$122B of not-yet-in-service assets may flatter cloud margins (The Investor's Podcast); the premium brand advertisers ("Coke and Chevy") still haven't fully followed the audience to YouTube (Decoder); and treating creators as unpaid gig labor invites eventual regulation. Next to watch: any disclosure on YouTube CTV ad revenue and whether blue-chip TV budgets are migrating; Cloud margin durability once those $122B of assets start depreciating.

META, Bull: The AI-ad-supply flood is coinciding with rising Meta ad prices, suggesting demand is absorbing the new supply (Limited Supply, Sep 16); Muse is genuinely fast and ships a real consumer-AI story. Bear: Muse's first field test exposed the core problem: even willing users won't trust Meta with their inbox and wallet (AI For Humans, Sep 18), and a fast-rising rival (Instinct, raising at a $10B valuation) is already viewed as the better product; the $18B settlement is a precedent for a long litigation tail, and even sympathetic observers call the new teen controls "a panacea" while noting "$18 billion is nothing to Meta's bottom line" (The Current, Sep 15). Next to watch: early Muse retention/attach vs. Instinct; whether rising CPMs hold as AI creative supply keeps expanding; the next litigation milestone.

SPOT, Bull/Bear: The read-through is indirect: the industry-wide shift of podcasts to video-first "vodcasts" watched on YouTube and TV (Next in Media, Sep 15) is both Spotify's opportunity (its video-podcast push) and its threat (YouTube owns the living room). Next to watch: video-podcast monetization disclosure at the next print.

PINS, Bull/Bear: The drift remains unhelpful: social-commerce dollars and effort keep concentrating on TikTok Shop and Amazon's affiliate rails, not Pinterest's shoppable pins. Next to watch: affiliate take-rate and shoppable-pin conversion commentary on the next print.

Read-throughs

TikTok / ByteDance, big on GMV, brutal on profitability. TikTok is tightening its seller-quality bar with a "Shop Performance Score" (product satisfaction, service, fulfillment) that gates incentives and features, and an operator's candid admission that "many, if not most brands on TikTok shop are not making a profit" once fees, affiliate commissions and ads are netted out (Ecomm Breakthrough, Sep 14). The GMV is real; the merchant economics are hard, which is relevant for anyone modeling how durable that share-shift is.

Digital publishers, going video-first and building "TV" in-house. Betches (a 15-year-old company that "started as a blog... pre-Instagram") is "redefining itself as an entertainment company... fully video first," and built a scripted show, "Sisters," in-house, shot horizontal, cut vertical, and posted across YouTube long-form, Shorts, Instagram and TikTok, with season one deliberately kept sponsor-free (Next in Media, Sep 15). The tell for the platforms: they "forced all of our podcasts to be vodcasts because that's what the audience wants... find the podcast on YouTube, on their TV." More confirmation that YouTube is where attention (and eventually ad budgets) are pooling.

Creator-media M&A, a cautionary tale, revisited. A retrospective on Barstool Sports is a useful reminder of how badly creator-brand deals can go: Penn Entertainment paid $163M for 36% of Barstool (a ~$500M valuation, "a 36x increase over the Chernin valuation four years earlier"), later took an $800 million write-down, and sold the whole thing back to Dave Portnoy for one dollar; the business itself lost $16M on $80M of revenue in the first half of 2023, "spending a buck 25 to make a dollar" (Corporate Gossip, Sep 18). Historical, but a live lesson as private equity keeps circling creator businesses: audience reach and profit are not the same thing.

Creator tooling & payments rails: A structural note from the AI-ads operators: creator deals are shifting from flat fees toward "percentage of GMV or percentage of ad spend," with a "gold rush on the creator side" expected into the holidays before budgets reset (Limited Supply, Sep 16).

What changed vs last week

Last week (issue dated September 12, 2026) was all Meta: the $18 billion settlement reframed as a competitive weapon, the Muse launch pop, private equity buying YouTube channels, and Snap boxed in. This week the mic moved to Google and YouTube, and the tone turned more bullish on the platforms.

  • New center of gravity: YouTube and Alphabet. Last week had no Alphabet deep-dive; this week delivered two: a creator's insider account of YouTube's near-zero-cost, "cable company" model (Decoder, Sep 14) and a numerate investor case that the "AI kills search" fear has reversed (The Investor's Podcast, Sep 17). That's a genuine change: the bear thread this newsletter has run (conversational AI stealing Google's commercial queries) got its clearest rebuttal yet.
  • Muse went from launch hype to a real-world review, and it didn't go well. Last week: the launch, a 6.5% pop, sell-side targets to $780. This week: a hands-on test that found Muse fast but weaker than rival Instinct, with Meta's reputation the deal-breaker (AI For Humans, Sep 18). The skeptics' prediction is playing out.
  • The Meta settlement stayed in the story, but as corroboration, not a new number. Last week's numerate weapon-thesis (Patrick Boyle) was echoed this week by a grieving parent who sat in the trial, with the same "$18 billion is nothing" conclusion, plus a fresh emphasis on precedent and the litigation tail, and again the note that "their stock price went up" (The Current, Sep 15).
  • The AI-ad-supply theme resurfaced with a twist. It's a thread we last ran in depth in the spring. This week's new nuance: creative costs are collapsing ($40–150 AI video vs. $1,000+ before) while Meta's ad prices are rising (Limited Supply, Sep 16), which cuts against the simple "AI floods the feed and crushes CPMs" bear case.