Will Infinite Slop be monetized?

Is Infinite Slop making money right now?

Infinite Slop is currently free for viewers, but fal.ai is already covering the expensive part of keeping the stream alive.

We found no pricing page, subscription, paid prompt option or advertising product on InfiniteSlop.ai. Anyone can currently watch the stream, submit a prompt and vote in the queue without paying.

On the cost side, Levels said fal.ai was “very generously” sponsoring the stream because running nonstop AI video would otherwise be extremely expensive. Fal also developed the H3 Max model that makes the stream technically possible.

There is no public evidence that fal.ai is paying Levels cash. What fal appears to provide is the generation infrastructure. So Infinite Slop has already attracted commercial support, although we would not call it a revenue-generating business yet.

Is Infinite Slop already big enough to monetize?

Infinite Slop has already attracted enough attention to test monetization, with more than 2,000 concurrent viewers at its recent peak.

The growth happened fast. Levels reported 37,000 viewers during the first full day, then moved Infinite Slop from his experimental skunkworks server to InfiniteSlop.ai. More recently, he said the stream had crossed 2,000 simultaneous viewers.

What happened next is more interesting than the raw viewer count. Most traffic was coming from phones, so Levels switched the generated videos from landscape to 9:16 portrait. The original queue also became crowded enough that he added upvotes, meaning viewers now compete to decide which prompts actually reach the screen. The site shows what is generating, what comes next and older generated clips.

Those changes tell us there is real usage pressure already. We still have almost no evidence about retention, though. A viral first weekend can create thousands of viewers without creating a lasting audience, and that distinction will decide whether Infinite Slop becomes commercially interesting.

How expensive is Infinite Slop to run?

Infinite Slop would consume roughly $207,000 per 30 days at fal.ai's normal public 768p H3 Max price if it generated fresh video continuously.

The calculation is unusually straightforward. Infinite Slop can produce four 15-second clips per minute, which adds up to one full minute of generated video every minute. A continuous stream therefore needs 86,400 generated seconds every day.

Fal currently lists H3 Max at $0.08 per generated second for 768p output, with an introductory rate of $0.04. At the normal price, continuous generation comes to $6,912 per day and $207,360 over 30 days.

That figure should be treated as a retail-price equivalent. Fal's own infrastructure cost will be much lower than what an API customer pays, and Infinite Slop may receive custom terms. Still, six figures of monthly API value gives us the right order of magnitude. This is an unusually expensive little internet experiment to keep running for free.

H3 Max 768p price	24 hours of continuous video	30 days
Introductory rate, $0.04/sec	$3,456	$103,680
Normal rate, $0.08/sec	$6,912	$207,360

Why would fal.ai keep sponsoring Infinite Slop?

Fal.ai has a strong reason to sponsor Infinite Slop because the entire stream works as a live demonstration of H3 Max.

Fal's pitch for H3 Max is speed. The company says a five-second 768p video can be generated in under three seconds, which finally makes faster-than-playback AI video possible. Infinite Slop turns that technical claim into something anyone can understand: the AI keeps producing television without catching up to the playback head.

The relationship with Levels also predates this experiment. His public investment portfolio shows that he invested in FAL.ai at its Series A stage, and he has described fal as an API provider he already uses. His stated investment philosophy is largely to invest in products he uses himself.

Fal gets more than a logo underneath a livestream. Every conversation about how Infinite Slop can generate forever points back toward the infrastructure that made it possible. For an AI infrastructure company trying to reach developers, that is unusually relevant sponsorship inventory.

Infinite Slop could remain commercially useful to fal even if viewers never pay anything.

Why are paid Infinite Slop prompts so tempting?

Paid prompts are tempting because Infinite Slop has created a tiny amount of control inside an unlimited stream: only four viewer ideas can reach the screen each minute.

That scarcity appeared almost immediately. People can submit as many ideas as the interface allows, but the stream still has room for only four 15-second generations per minute. Once hundreds or thousands of people are participating, most prompts lose.

Users quickly suggested letting people pay to jump the queue or bid against each other.

The valuable product would be airtime in front of a shared audience. Anyone can already buy an AI-generated video elsewhere. Infinite Slop adds something harder to reproduce: thousands of other people might watch the result at the same moment.

That starts to look closer to Twitch donations, promoted posts or an advertising auction than a normal AI-generation product.

And the economics improve as the audience grows while production capacity stays fixed. Two thousand viewers still compete for the same four slots per minute that 100 viewers would have competed for earlier.

How much could Infinite Slop make from paid video slots?

Infinite Slop could generate meaningful revenue while selling only a small share of its 172,800 monthly video slots.

Four generations per minute means 5,760 clips per day and 172,800 over 30 days. That creates a useful way to test the economics.

At fal's normal 768p retail price, the theoretical monthly generation bill is $207,360. If every slot were sold, the average price needed to cover that amount would be just $1.20 per video.

Selling every slot would probably wreck the product, so the smaller percentages are more useful. Selling 10% of the available generations would require an average price of $12 to match $207,360. At 5%, the figure becomes $24. Even at 1%, 1,728 paid clips per month would create meaningful inventory.

These are break-even calculations rather than estimates of what viewers would actually pay. We do not yet know whether the demand exists. But the numbers explain why the four-video limit may eventually become commercially useful rather than simply annoying.

Share of slots sold	Paid clips per 30 days	Average price to reach $207,360
1%	1,728	$120
5%	8,640	$24
10%	17,280	$12
25%	43,200	$4.80

Why hasn't Infinite Slop started charging for prompts?

Infinite Slop still uses free voting because Levels explicitly thinks a money-based queue would make the stream less fun.

When viewers suggested letting people pay to bid for the four available slots, Levels pushed back. He said he preferred the chaotic feel the project had developed, comparing it with 4chan and Hoodmaps.

His product decisions since then line up with that view. He added an upvote system instead of a checkout. A prompt enters the queue, viewers vote, and popular ideas get pulled into generation.

Levels has already seen the obvious monetization mechanism and deliberately chosen something else for now.

A full auction could quickly turn the feed into promotions, crypto spam and people purchasing attention. The free queue gives random viewers a chance to shape the stream, which is a large part of the entertainment.

Could Infinite Slop charge for some prompts without ruining the stream?

Infinite Slop could probably sell a small amount of airtime while leaving almost all viewer-generated clips free.

Consider one paid generation every ten minutes. Infinite Slop produces 40 clips during that period, so a single commercial slot would occupy only 2.5% of the stream. The other 97.5% could continue coming from the community.

Levels could also sell temporary promotion inside the queue rather than guaranteed airtime. A sponsored prompt might appear more prominently but still need votes. Brands could sponsor a category or a short themed period without permanently changing the normal queue.

This feels much closer to how Infinite Slop would realistically monetize. There is enough inventory to keep paid content rare.

The important constraint is cultural rather than technical. People currently participate because their stupid idea might unexpectedly become the next video. Once money consistently determines what airs, that game gets less interesting.

Would normal ads pay Infinite Slop's AI bill?

Normal video advertising could eventually make money for Infinite Slop, but the current audience would need unusually strong ad economics to cover H3 Max at retail prices.

We can test the rough scale. Imagine Infinite Slop averaging 2,000 simultaneous viewers all day and showing six 30-second ad impressions per viewer every hour. That would create around 288,000 impressions per day.

Matching the $6,912 daily retail-equivalent cost of 768p H3 Max would require about $24 in gross revenue per thousand impressions.

That becomes much harder if the average audience is below the recent 2,000-viewer peak. At 1,000 average viewers, the required CPM doubles to $48. At 500, it reaches $96.

Those calculations deliberately make a generous assumption by treating the recent peak as something the stream might sustain around the clock. We have no evidence that Infinite Slop can do that today.

Direct sponsorship therefore looks much more attractive than filling the stream with ordinary programmatic ads. A specialized AI company can value this audience very differently from a generic advertiser buying impressions.

Average concurrent viewers	Ad impressions/day at 6 per hour	Gross CPM needed for $6,912/day
500	72,000	$96
1,000	144,000	$48
2,000	288,000	$24
5,000	720,000	$9.60

Would people actually pay for an Infinite Slop subscription?

An Infinite Slop subscription looks weak today because viewers currently get most of the interesting experience simply by showing up together.

A Photo AI subscription has an obvious exchange: the customer pays and receives generated photos. With Infinite Slop, everybody watches the same feed, and the shared chaos is part of the attraction. Moving the main stream behind a subscription would shrink the crowd that makes it entertaining.

Levels could eventually add paid features around the stream. More voting power, private channels, personal histories, saved generations or custom streams are all technically imaginable.

We would need to see people repeatedly using those features before treating subscriptions seriously. Right now the product has a much cleaner thing to sell: scarce influence over a large free audience.

Could brands become Infinite Slop's best customers?

Brands could become much better Infinite Slop customers than individual viewers because AI-generated sponsorship can become part of the entertainment itself.

Imagine a game studio letting the audience generate bizarre scenes involving characters from a new release. A movie could sponsor a themed hour. An AI company can let viewers repeatedly stress-test its model in public. Fal is already doing the simplest version of this by sponsoring the infrastructure behind the whole experiment.

This format gives advertisers something beyond impressions. Viewers can interact with the sponsored material, mutate it and try to push increasingly strange versions onto the stream.

Brand safety will limit who wants to participate. Infinite Slop currently gets much of its energy from strangers submitting unpredictable prompts, and that chaos will scare away conservative advertisers.

AI companies, gaming companies, internet-native brands and entertainment launches are much better fits. They can tolerate the weirdness because the weirdness is part of what they are buying.

Will people still watch Infinite Slop once the novelty wears off?

Infinite Slop's biggest monetization risk is audience collapse, and an earlier AI livestream shows just how extreme that collapse can be.

Nothing, Forever went viral in 2023 by generating an endless AI parody of Seinfeld. During its breakout period, reports put concurrent viewership around 15,000 to 17,000, and the Twitch channel eventually collected more than 132,000 followers.

The channel still exists, but its current live audience is tiny compared with the viral peak.

That is the ugly part of this category: viral concurrent viewers can disappear almost completely once the novelty is gone.

Infinite Slop does have a stronger participation loop. Viewers write prompts, vote, watch ideas enter the queue and react to whatever previous viewers produced. Recent updates have made that loop more explicit. The current site now separates chat and queue views, shows what is generating next and lets people browse categories such as banana, cat, space, news and game show.

We think that interactivity gives Infinite Slop a better shot at retaining people than passive endless AI television. We have no retention data yet that proves it.

AI livestream	Viral audience	Recent audience
Nothing, Forever	About 15K-17K concurrent	Tiny live audience today
Infinite Slop	2K+ concurrent so far	Too early to know

Does Pieter Levels usually monetize projects like Infinite Slop?

Levels' own project history shows that a viral experiment from Pieter Levels does not automatically become a business.

His public project tracker currently contains 114 projects. Levels classifies 69 of them, about 61%, as non-profit or having no profit goal. Another 19 are failures, while six are still too new to classify.

Only nine sit in his long-term financial success category. Eleven more made money for a while without becoming sustainable successes. Combined, that means 20 of 114 projects, about 18%, reached either of his two money-making categories.

Infinite Slop currently looks like exactly the kind of experiment that could go either way. Levels built it extremely quickly after H3 Max made faster-than-real-time AI video possible, watched how people used it, then started modifying the interface around actual behavior.

So we should pay more attention to what he builds next than to his reputation as a successful indie hacker. A checkout, sponsored slot system or paid queue would be meaningful evidence. So far, he has gone in the opposite direction by adding free voting.

Will Infinite Slop be monetized?

Yes, we think Infinite Slop will probably be monetized if the audience survives, with sponsorship and a small amount of paid airtime looking far more likely than a subscription or fully paid queue.

The commercial ingredients are already visible. Infinite Slop has crossed 2,000 concurrent viewers, only four prompts can reach the screen each minute, fal.ai is already sponsoring the expensive generation layer, and viewers themselves immediately identified those scarce prompt slots as something people could bid for.

Levels has also given us a useful constraint. He currently dislikes the idea of letting money dominate the queue because he thinks it would make Infinite Slop less fun. The new upvote system reinforces that position.

Our strongest bet is therefore a mostly free stream where commercial content stays rare. Fal or another company sponsors the infrastructure. Brands occasionally buy a generated segment, challenge, theme or takeover. Perhaps a small percentage of prompts become paid priority slots. The large majority of the feed remains controlled by viewers.

The recent audience numbers make that model plausible. They do not make it proven. Nothing, Forever's collapse from roughly 15,000-plus concurrent viewers to a tiny live audience is a brutal reminder of what can happen once an AI novelty stops feeling new.

So retention is the number we would watch now. If Infinite Slop can still attract hundreds or thousands of simultaneous viewers after the initial curiosity fades, monetizing some of its 172,800 monthly generation slots becomes almost too obvious to ignore.

If viewers disappear, fal.ai may still have enough marketing value to keep the experiment alive.

Our judgment is fairly sharp: Infinite Slop is likely to monetize attention around the stream, while keeping the core experience free.