What changed on August 10 and 11, and who it hits
On August 10 and 11, 2026, YouTube announced the first large change to the Partner Program since 2018. From February 1, 2027, a channel applying for ad and Premium revenue needs 1,000 subscribers and 8,000 public qualified watch hours in the last 365 days, where the bar was 4,000. The Shorts entrance moved the same way: 1,000 subscribers and 20 million qualified Shorts views in 90 days, where it was 10 million. Both thresholds doubled.
If you are already in, the new bar does not apply and you keep monetizing. One task is unavoidable: accept the new terms in YouTube Studio by January 31, 2027, or lose access to monetization from February 1. The activity rule is 1,000 qualified hours in 365 days, or 1 million Shorts views in 90 days, or 2 long videos or 5 Shorts every 90 days, with an extra 90 day window if you fall under. Keeping ad and subscription revenue on Shorts will require 10 million qualified Shorts views in 90 days, and falling short costs the Shorts revenue only.
The door nobody talks about did not move. Fan funding, meaning Super Thanks, Super Chat and memberships, plus shopping products, stays at 500 subscribers and 3,000 qualified hours in a year, or 3 million Shorts views in 90 days: a money door at half the new bar, still open. Premium Lite expands to every country where YouTube Premium operates, with net subscription revenue at 30 percent on Premium and 60 percent on Premium Lite, split 55 percent to long video and 45 percent to Shorts. YouTube's stated reason is scale: over 200 billion daily Shorts views and over 1 billion hours a day watched on TV.
172 days, and why the real deadline sits before January 31
From today, August 12, 2026, to January 31, 2027 there are 172 days. About 24.5 weeks. Anyone accepted before February 1, 2027 enters under the 4,000 hour bar and stays on it. Anyone arriving after needs 8,000.
Hitting the number is not the same as being accepted. You reach the bar, then YouTube reviews the channel, and that review takes days or weeks. So the real deadline for locking the old bar sits well before January 31. Treat December as the target and January as buffer.
The window is a rolling 365 days, which cuts both ways. Hours you accumulate today still count in January, so nothing published this week is wasted. But hours from a year ago are already falling out the back, and a channel that published hard last autumn and then went quiet loses hours every day while it waits.
172 days also caps what is possible: about 172 videos publishing daily, about 123 at one per weekday, about 74 at three a week. Hold those numbers, because everything below divides 8,000 hours by the hours one video produces.
The formula, run in front of you, with the one guess declared
You cannot see another channel's watch hours. Nobody can, except the owner inside their own Studio. Anything you read about a competitor's hours is an estimate, including this. What makes one usable is that every input is public and the single unknown is declared instead of hidden inside a round number.
The formula: views inside the window, times average duration in minutes, times the share actually watched, divided by 60. Run it in two steps. Minutes per view first: 12 minutes at 40 percent watched is 4.8. Then the target: 8,000 hours is 480,000 minutes, and 480,000 divided by 4.8 is 100,000 views. The old 4,000 hour bar was 240,000 minutes, so 50,000 views.
Three of the four inputs are public. Duration is on every thumbnail and the view count is on every video. The publish date defines the window: count only videos published in the last 365 days, because a video published 300 days ago collected all of its views inside the window, so its lifetime count and its window count are the same number. Percentage watched is the fourth, it is not public, and it is the only guess.
So produce a range, not a number. A channel with 40 videos of 12 minutes and a median of 3,000 views has 120,000 views in the window. At 30 percent that is 3.6 minutes per view, 432,000 minutes, 7,200 hours. At 50 percent it is 6 minutes per view, 720,000 minutes, 12,000 hours. The honest read is not 9,600 hours. It is between 7,200 and 12,000, so the channel cleared 4,000 long ago and sits right on top of 8,000.
This gives a floor, not a total. Videos published before the window kept collecting views inside it, and deleted, unlisted and private videos counted for the channel while staying invisible to you. Shorts stay out, because their views feed the 20 million door.
- News and commentary, 5 to 8 minutes: 40 to 55 percent watched, so 8 minutes at 45 percent is 3.6 minutes per view
- Explainer and tutorial, 10 to 20 minutes: 30 to 50 percent, so 12 minutes at 40 percent is 4.8 minutes per view
- Lists and rankings: 25 to 40 percent, because viewers jump to the item they came for
- Narrative and documentary, 20 to 40 minutes: 30 to 40 percent, so 20 minutes at 40 percent is 8 minutes per view
- These bands are working assumptions, not published YouTube data, so run both ends and report the range
- 8 minutes at 45 percent and 12 minutes at 30 percent both land on 3.6 minutes per view, both needing 133,333 views
Reading a channel that crossed fast: duration, rhythm, concentration
The question now changes: not whether the channel crossed, but what made crossing fast. Duration answers first, because it multiplies everything else. At 40 percent watched, a 20 minute channel delivers 8 minutes per view against 2.4 for a 6 minute channel, more than three times the hours from the same audience.
Rhythm second. Count the videos published in the last 365 days and divide by 52. At 48 videos the channel publishes weekly, at 150 it publishes most days. Videos per week times hours per video gives hours per week, the number that decides whether anything crosses before a deadline.
Then concentration, the check most people skip. Sort the year's videos by views and see how much of the total sits in the top three. Over half means anchor driven: the channel crossed because a few videos caught a wave. Under about a quarter means the views are spread, and spread means the format performs on the ordinary video.
Spread is the better model for one blunt reason: you can reproduce a format, you cannot reproduce a wave on schedule. It gives a stable hours per publication figure you can divide 8,000 by. Age closes the read: the About page carries the join date and the video count. Eight months old inside a 7,000 to 12,000 hour band means the structure is right. Five years old in the same band means the format takes five years.
Use the median, never the average, or one anchor video will lie to you
The average is the number most likely to cost you six months. Take a channel with 30 videos in the window: 29 around 800 views, and one that caught fire at 400,000. Total 423,200 views. Average per video about 14,100. Median 800.
Run both at 12 minutes and 40 percent watched, so 4.8 minutes per view. On the real total, 423,200 views is 2,031,360 minutes, about 33,856 hours. On the median, 24,000 views is 115,200 minutes, 1,920 hours. Same channel, same year, two answers almost eighteen times apart.
Both are true and they answer different questions. The channel really did accumulate around 33,856 hours, so it cleared the bar. But what the format produces when you publish normally is 1,920 hours a year. Copy it expecting the average and you are buying a lottery ticket and calling it research.
So convert the median into hours per video: median views times minutes per view, divided by 60. At 800 views and 4.8 minutes that is 64 hours per video, and 8,000 divided by 64 is 125 videos. At a median of 2,500 the same format gives 200 hours per video and crosses in 40. The calendar goes from 125 publications to 40.
- 500 median views per video is 40 hours per video, so 8,000 hours needs 200 videos
- 1,000 median views is 80 hours per video, so 100 videos
- 2,000 median views is 160 hours per video, so 50 videos
- 5,000 median views is 400 hours per video, so 20 videos
- Against 172 days: 200 videos is more than one a day, 100 is four a week, 50 is two a week, 20 is under one
- Run the ladder against the old 4,000 hour bar and every count halves, which is what disappears on February 1, 2027
Copy the format, never the promise
You now have a shortlist: channels whose estimated hours clear the bar and whose views are spread rather than stacked on three anchors. What transfers is structure, and what does not transfer is identity. Getting that line wrong turns research into a copy that competes badly.
Format transfers, and duration is the most valuable item because it is the multiplier in the whole calculation. Publishing rhythm transfers. The angle of the topic list transfers: whether titles answer a question, settle an argument, rank things or tell a story, and whether the catalogue is evergreen or news driven. Hook length and chapter structure transfer too.
The specific promise does not. Their title is a claim they support with their sources and their credibility, and repeating it makes you the second result for a question already answered. The thumbnail is their visual identity and a near copy reads as a knockoff. The script never transfers. Neither does a format your numbers cannot sustain, like 40 minute documentaries once a month. Model what your median carries today.
- Copy: video duration, because it multiplies every hour you accumulate
- Copy: publishing rhythm, measured as videos per week across the last 365 days
- Copy: the angle of the topic list, meaning what kind of question every title answers
- Copy: hook length, chapter structure, and whether the catalogue is evergreen or news driven
- Never copy: the exact promise in the title, which belongs to their sources and their credibility
- Never copy: the thumbnail and the visual identity, which reads as a knockoff
- Never copy: the script, ever, or a format your current median cannot sustain
Where FalconVid closes the distance between the estimate and the calendar
Every number above ends in the same place. The estimate tells you which format accumulates hours fastest, the median tells you how many videos it needs, and the calendar decides whether it happens before February 1, 2027. That is where research usually dies, because at one video a week, 100 videos is not a plan, it is two years.
FalconVid's Spy is the front half of this work: it shows the channels already monetizing in your niche, so you model what already pays instead of starting from zero. The rest turns the answer into published videos, because the researcher, the screenwriter, the narrator, the editor and the sound design run in parallel, from a calendar you approve once, with a finished video in up to 30 minutes.
The cost is smaller than people assume. A 12 minute video costs 1,008 credits in economy mode, 8,676 in balanced and 26,760 in premium. The 100 videos from the 1,000 view line are 100,800 credits in economy mode, and at $0.003133 per credit that is about US$ 316 of credit for the whole production run behind an 8,000 hour target.
Capacity is not the wall either. Pro is $97 with 30,000 credits, which holds 29 economy videos a month, so the whole 100 video run lands in about three and a half months, comfortably inside the 172 day window. Business at $297 with 95,000 credits does 94 economy videos a month and finishes the same run in about five weeks. The 200 video line is 201,600 credits, which Scale at $997 with 320,000 credits covers inside a single month. Starter is $47 with 15,000 credits, 1 channel and 2 concurrent generations, realistically about 10 to 12 videos a month mixing economy with one in balanced.
Duplicating a project into another language pays only the difference, essentially the narration: 288 credits on the premium voice and 36 on the economy voice for 12 minutes, roughly 3.6 percent of a 1,008 credit economy video. Every creation feature is on every plan, in up to 63 languages, publishing to YouTube, Instagram, TikTok, Rumble and Facebook, and what changes as you go up is volume, channels, simultaneous generations, the Senior Analyst from Pro up (7 days free on Starter) and support.
172 days: what the arithmetic actually buys you
There are 172 days between today and January 31, 2027, and the review between hitting the bar and being accepted eats days or weeks of that. Here is what this method buys you.
I do not know how to make videos. You do not have to. The researcher, the screenwriter, the narrator, the editor and the sound design are AI specialists working in parallel, and the only thing you approve is the calendar. Ultra realistic narration in up to 63 languages, karaoke captions in more than 15 styles and video SEO come on every plan.
I do not know how to pick a niche. That is what this article answered, and what the Spy does at scale. Rank the channels already monetizing by estimated hours per video, keep the ones whose views are spread, model that format. You start from what already monetizes.
What if it comes out bad. The Studio shows the first cut before anything is published, and you fix a single line, swap the media under it or open the full timeline, regenerating only what you touched. Narration is billed per minute, 3 credits on economy and 24 on premium, so a 30 second correction is a rounding error next to 1,008 credits.
I do not have time is the real reason most channels miss the bar, and a finished video in up to 30 minutes with concurrent generations, 2 on Starter up to 50 on Scale, removes it. Then the honest one: what if I do not hit the bar. We do not sell monetization and nobody can. YouTube decides views and acceptance. What we sell is production, the volume of published videos the arithmetic requires.
Start on the 7 day trial with its 2,000 credits, enough to put real videos of your modeled format on screen. There is a 7 day guarantee, annual billing gives two months free, and you cancel whenever you want.
- Rank every channel in your niche by estimated hours per video this week, with public data and a calculator
- 8,000 hours is 480,000 minutes, so at 12 minutes and 40 percent watched it is 100,000 views, a target you can plan against
- A format with a 1,000 view median needs 100 videos: 100,800 credits in economy mode, about US$ 316 of credit
- Pro at $97 carries 30,000 credits and 29 economy videos a month, so the 100 video run finishes in about three and a half months, inside the 172 day window
- A finished video in up to 30 minutes, with concurrent generations from 2 on Starter to 50 on Scale
- The same calendar duplicates into another language for the difference alone, roughly 3.6 percent of an economy video
- You approve the calendar once, and publishing runs to YouTube, Instagram, TikTok, Rumble and Facebook