Summary

YouTube’s 55/45 creator split is one option, not a universal fit for established creators. The opening host called the 45% cut “absurdly high” for talent able to sell embedded ads directly, while another host said “the net is what matters.” Mohan said models should follow each creator’s business objectives; he pointed to more than $70 billion paid to the creator economy over three years and north of 3 million creators earning through the Partner Program.
YouTube’s advertising moat is the intensity of creator-led engagement. Mohan argued that viewers arrive as “superfans,” producing differentiated advertiser ROI across brand and direct-response campaigns. That engagement underwrites both YouTube’s payouts and the hundreds of thousands of jobs around the creator ecosystem.
Personalization fragments shared culture, but YouTube believes scale and live programming can create a new water cooler. The platform reaches 2 billion people daily, and the introduction said Shorts had “just surpassed 70 billion views a day,” with the qualifier “I think.” Mohan cited roughly 13%-14% of U.S. television viewing in the cited Nielsen figure, excluding mobile. He said live events can create shared moments; the host pointed to integrations with Dude Perfect and MrBeast around a Brazil game.
Mohan framed COVID-era moderation as a response to extraordinary conditions, not a permanent censorship template. He said the relevant policies are “non-existent today,” while acknowledging that an open platform will be attacked from both directions. The unresolved risk is implementation: firearms, poker, age suitability, gambling laws and national regulation all create monetization boundaries that creators can experience as censorship.
Subscriptions are growing, but advertising remains YouTube’s economic center of gravity. About 125 million people subscribe to YouTube Premium, whose revenue is distributed using an engagement- and watch-time-type calculation; many subscribers primarily use it as a music service. YouTube TV and à-la-carte Primetime Channels broaden the bundle, yet Mohan said most of the $70 billion creator payout still comes from ads.
YouTube wants to handle AI likeness the way Content ID handled copyrighted media. Mohan rejected a blanket assumption that AI-generated content is inherently violative, favoring disclosure plus community guidelines. For impersonations and “AI slop,” the proposed rights layer is automated likeness detection that gives creators a choice about takedown or monetization, while YouTube could potentially take ownership of derivative content.

Deep dive

1. YouTube’s creator bargain is distribution for a 45% cut

The opening host’s challenge was blunt: mature creators such as Tucker and Megan can sell baked-in advertising themselves, making YouTube’s 45% cut feel “absurdly high.” He asked whether the platform needs separate economics for creators who already possess audiences and sales teams.
Mohan’s answer was that monetization models should follow each creator’s business objectives. Tucker or Megan might rationally represent their own inventory, while a new creator is primarily “betting on yourself” and building engagement; monetization arrives as the audience and business grow.
Another host interjected that “the net is what matters,” and the opening host agreed that the relevant comparison is how much money a creator makes. Mohan’s evidence for the existing model was more than $70 billion paid out over three years and north of 3 million creators currently earning through the Partner Program. He said the 55/45 split also pays creators, media-company partners and music partners.
The advertising mechanism, in Mohan’s telling, is unusually committed attention: viewers are not passive reach but “superfans” of specific creators. That engagement translates into stronger ROI for both brand-building and direct-response advertisers and funds “hundreds of thousands of jobs” around the ecosystem.

2. Personalization splinters culture while live video rebuilds it

A host’s concern was the loss of national appointment viewing: infinite distribution lets everyone watch exactly what they want, but removes the “connective thread” that once put the same shows into the zeitgeist.
Mohan conceded that a personalized feed “is going to look like you” and can fragment attention. His counterpoint was aggregation: with 2 billion daily users, even niche trends can become defining culture for a generation or travel globally rather than nationally.
Mohan said live events and water-cooler moments remain important and cited the Brazil game carried on YouTube. The host argued that, to the NFL’s credit, integrations with creators including Dude Perfect and MrBeast made the livestream more relevant to the YouTube generation, describing that creator engagement as a “new water-cooler moment.”
On measured scale, Mohan said YouTube had been America’s number-one streaming platform for two years and accounted for roughly 13%-14% of U.S. TV viewing in the latest Nielsen figure he recalled. That measurement excludes mobile. The episode’s introduction also said Shorts had “just surpassed 70 billion views a day,” qualified by “I think.”

3. COVID moderation ended, but boundary disputes did not

A host asked whether social platforms had learned from COVID-era censorship or whether openness was merely “on pause during the Trump years.” Mohan first distinguished YouTube from social feeds: 99% of what happens on the platform is creators, podcasters, music and related content, with music described as its largest and most important vertical.
Mohan’s defense rested on context. In March and April 2020, claims included people contracting COVID through 5G towers; today, he said, the policies that existed then are “non-existent.” He also stressed that YouTube was heavily criticized for leaving up content around the Wuhan virus controversy and other material that competing platforms treated differently.
The enduring principle, he argued, must be flexibility alongside openness: “There’s a lot of magic that happens because of the open platform.” YouTube’s mission—“give everyone a voice and show them the world”—makes freedom of expression the starting point, while the company must still legally operate country by country.
The host’s pushback was concrete: firearms-safety and poker creators report demonetization despite substantial audiences. Mohan said some such content is monetized, but sales restrictions, age suitability and gambling law add nuance; he said enforcement across hundreds of hours uploaded every minute requires a combination of policy and algorithms.

4. Subscriptions widen the funnel without displacing ads

YouTube TV began with Mohan and his team’s enthusiasm for sports and news. Features such as Multiview, Key Plays and fantasy integrations reflect an attempt to reinvent the fan experience inside what some viewed as a shrinking paid-television ecosystem.
The adjacent bet is Primetime Channels, which lets users buy traditional channels à la carte inside the main YouTube app. One host’s desired endpoint was simple: access something like CNBC without switching products or repeatedly encountering regional restrictions while traveling.
Mohan said advertising “is today and will remain” the predominant creator-monetization engine because of YouTube’s scale, and most of the $70 billion payout came from ads. Still, roughly 125 million Premium subscribers buy an uninterrupted experience; revenue reaches creators through a rough engagement- and watch-time-type calculation.
Premium’s origin matters to the mix: it began as a premium music service, so many subscribers use YouTube as their primary music service for discovery, listening and music videos rather than merely as ad-free television.

5. AI likeness is becoming the next Content ID

Mohan expects no clean boundary between AI-generated and AI-assisted media: it will be “a continuum.” YouTube already lets users enter a text prompt in the app to generate a video using its VO models and can attach an “#AI-generated” label either “front of the box” or in metadata, though he conceded that labeling is “obviously not foolproof.”
His policy distinction is behavioral rather than technological: AI generation alone should not make content violative. YouTube’s published community guidelines remain the relevant framework, with disclosure giving viewers more information about what they are seeing.
The hosts argued that transparency alone will not cure “AI slop,” citing AI-generated Corvette thumbnails claiming to show a new model or prototype and live “chimoth” content used to give Bitcoin away. Mohan said the incentive is to insert such channels into the algorithm, and that he has banned some of them.
Mohan’s proposed answer borrows from Content ID, the rights-management system he said arguably created the creator economy and “saved YouTube.” Likeness detection could recognize a creator’s face or voice and give that creator a choice about whether the content should come down; some creators could monetize it, while YouTube could potentially take ownership. Taylor Swift’s voice and Marquez Brownley’s face were his examples of the identity assets creators most want protected.

OpenAI vs. Grok: The Race to Build the Everything App w/ Emad Mostaque, Dave Blundin & AWG | EP #199 Sam Altman on Sora, Energy, and Building an AI Empire