Field Note #11: Streaming Video Pay-Per-Minute

I subscribe to four streaming services. I watch maybe three of them regularly. One I haven’t opened in six months. Between Netflix, Hulu, Max, and Apple TV+, I’m paying roughly $60 a month for access to content I mostly don’t consume. This is the streaming bundle model — we’ve rebuilt the cable package in digital form, complete with the same overpriced all-you-can-eat pricing that drove everyone away from cable in the first place. Netflix has 310 million subscribers worldwide and spends $17 billion a year on content, yet the average subscriber watches only a fraction of what they’re paying for. The gap between access cost and consumption value is enormous. What if that gap didn’t exist? What if instead of a monthly subscription, you paid only for the minutes you actually watched — at rates set by the content itself?

The Math of the Minute

Let’s start with the numbers that matter. Netflix’s $17 billion annual content spend, divided across its subscriber base of 310 million, means each subscriber costs roughly $4.57 per month in content acquisition. That’s before marketing, infrastructure, licensing, and profit. A typical Netflix subscription costs $15.49 per month in the US. The difference is the bundle premium — you’re paying for the option to watch anything, not for what you actually watch.

Now break it down by minute. Netflix’s average viewer watches roughly 2–3 hours per day. At $15.49 per month, that’s about $0.003 to $0.005 per minute of watched content. The cost of delivering that content — the CDN, the encoding, the packaging — runs about $0.001 per minute, which happens to be exactly the x402 protocol’s minimum payment floor on Base L2. Cloudflare Stream charges exactly $0.001 per minute of video delivered. The delivery cost and the protocol floor are the same number. That’s not a coincidence; it’s an alignment of technological and economic constraints.

If we strip out the bundle premium and charge purely per-minute, the content cost per minute is roughly $0.0014 (Netflix’s $17B divided by total watched minutes across the subscriber base). Add delivery at $0.001 per minute, and you get $0.0024 per minute — 24 cents for a 100-minute movie. That’s a fraction of what any streaming service charges today, even on a per-minute basis.

The Table: Pay-Per-Minute Streaming

The Revenge of the Light Viewer

The most compelling case for pay-per-minute streaming is the light viewer. Consider someone who watches five hours of video per month — a couple of movies or a few episodes of a series. Under the current subscription model, that person pays $15.49 for Netflix, $9.99 for Hulu, and maybe another $9.99 for something else. Total: $35.47 per month for maybe 300 minutes of actual viewing. That’s $0.12 per watched minute. Under a pay-per-minute model at $0.0024/minute (content + delivery), that same person pays $0.72. The savings are not incremental; they’re transformative.

Light viewers are a massive underserved market. They’re cord-nevers who don’t watch enough TV to justify a subscription. They’re international users in markets where $15/month is a significant expense. They’re students, retirees, and anyone else whose viewing habits don’t justify the bundle. The streaming industry has optimized entirely for heavy viewers — the 80th percentile who watch 5+ hours a day — and left the long tail to pirate or simply not pay at all.

The numbers bear this out. A light viewer paying $0.72/month versus $15.49/month represents a 95% price reduction. Even if the platform takes a 30% cut (Apple App Store style), the content creator gets $0.50 per light viewer per month. Multiply by 100 million light viewers globally and you get $50 million per month in content revenue from people who currently pay zero. The bundle optimizes for extracting maximum revenue from heavy users. Per-minute pricing optimizes for capturing value from everyone else.

The Heavy User Problem

But what about the heavy viewers? Someone who watches Netflix 6 hours a day, 180 hours a month, is currently paying $0.086 per hour. Under per-minute pricing at $0.0024/minute, that same viewer pays $25.92 per month — substantially more than their current bill. Heavy users are the most profitable customers under the bundle model. They’re also the most expensive to serve, costing Netflix roughly $0.0014/minute in content and $0.001/minute in delivery. At 180 hours/month, Netflix’s cost is $25.92 per heavy user. The $15.49 subscription price means Netflix loses $10.43 per month on every heavy viewer — subsidized by light viewers who pay for content they never watch.

This subsidy is the hidden engine of the streaming business model. Light viewers overpay; heavy viewers underpay. The cross-subsidy is what allows Netflix to offer “unlimited” viewing at a flat monthly price. Pay-per-minute pricing eliminates this subsidy. Heavy users pay full freight. Light users get a fair deal. The question is whether heavy users — who drive engagement, word-of-mouth, and cultural relevance — would accept a price increase of 67% or more.

The likely outcome is a hybrid model: a monthly cap with a per-minute micro-payment for anything above. Pay $5 for 200 minutes included, then $0.002 per minute after that. This protects heavy users from infinite bills while still giving light users a vastly better deal than the current bundle. It’s what mobile data plans figured out fifteen years ago, applied to content instead of bandwidth.

The Creator Economics

For creators, per-minute micropayments are a double-edged sword. On one hand, they eliminate the ad-supported middleman. A TikTok creator currently monetizes at $0.002 to $0.005 per active user per day through ad revenue sharing. Under a $0.001/minute micropayment model, a creator whose videos get 5 minutes of watch time per viewer per day earns $0.005 per viewer per day — right at the top of the ad range. A creator whose content is more engaging — 10 minutes per viewer per day — earns $0.01 per viewer per day, double the ad ceiling. The direct relationship between watch time and revenue creates the right incentives: make better content, keep people watching longer, earn proportionally more.

On the other hand, per-minute pricing changes what kind of content gets made. Currently, creators optimize for the ad-friendly sweet spot: 30–60 second videos that get high completion rates. Under per-minute billing, the incentive shifts to total watch time — longer content, serialized narratives, anything that keeps the viewer engaged for minutes rather than seconds. This could mean more documentaries, tutorials, and long-form storytelling, and fewer dance challenges and prank videos.

The risk is that it also incentivizes grinding — stretching content past its natural length to maximize billable minutes. Every platform that pays by the minute (or by watch time) struggles with this. YouTube’s algorithmic push toward 10-minute+ videos is a case study in the perverse incentives of time-based monetization. The x402 model would need to incorporate quality signals — completion rates, repeat views, user ratings — to prevent the race to the bottom.

The Infrastructure Question

Streaming video micropayments also raise a practical question that I keep circling back to in these Field Notes: how do you settle millions of micro-transactions per second without destroying the user experience? The x402 protocol settles USDC on Base L2 in 4–8 seconds. A viewer watching a 30-minute show would generate roughly 100 micro-payments (assuming per-second or per-chunk billing). Each payment needs to be validated, deducted from a pre-paid balance, and settled before the next one arrives. That’s 100 on-chain operations for a single viewing session. Multiply by 310 million Netflix subscribers and you get 31 billion micro-transactions per day — more than the entire daily transaction volume of every blockchain combined, by several orders of magnitude.

The answer, as with DNS payments, is bundling. Pre-pay, batch-settle. The viewer pre-loads $10 into their streaming wallet. The service deducts $0.0024 per minute from the local balance. Once per day (or per session), the service batch-settles all viewer debts against the content creator’s wallet. The on-chain footprint is one settlement per viewer per day, not one per minute watched. The pre-pay buffer absorbs the latency. This is the pattern that keeps appearing across every x402 use case: local accounting, global settlement, and a pre-pay balance to handle the gap.

Pay-per-minute streaming is one of the most financially sound applications of the x402 model I’ve examined in this series. The alignment between delivery cost ($0.001/min), content cost ($0.0014/min), and the protocol floor ($0.001) is almost suspiciously clean. The light-viewer economics are genuinely compelling — 98% savings for the bottom quartile of viewers, and a revenue stream from people who currently pirate or don’t subscribe at all. The heavy-user problem is real but solvable through tiered pricing. The creator incentives need careful design but are directionally better than today’s ad-supported model. The real obstacle isn’t the math — it’s the inertia of 310 million subscribers on monthly billing cycles, $17 billion in content commitments structured around the bundle, and a streaming industry that has invested decades in convincing consumers that the monthly subscription is the only sensible way to pay for video. I don’t think per-minute streaming replaces Netflix. But I think it sits alongside it, capturing the light-viewer market the incumbents have abandoned. — N.P. Vincent