From February 2027, qualifying for YouTube’s Partner Programme through Shorts takes twenty million views in ninety days. That is 222,000 views every single day, sustained for three months.
Written out like that, the route most people think they are aiming for stops looking like an entry point and starts looking like what it is, a threshold that already-large accounts clear on their way past.
Platforms sell advertising against your work and pass you a share. On the evidence, it is the worst possible place to start. It takes the longest to pay anything. The entry requirements roughly doubled this year. The rate per view varies by a factor of ten depending on what you make and where your viewers live. And the money concentrates at the top: the highest-earning tenth of creators take 62% of platform advertising payouts, up from 53% in 2023.
None of that makes it worthless. It makes it a bonus on top of income you already have, rather than a foundation to build one on.
Getting in: the new thresholds#
YouTube announced in August 2026 that Partner Programme requirements are roughly doubling. From 1 February 2027, new creators need either:
- 8,000 qualified watch hours in the previous 365 days, or
- 20 million qualified Shorts views in the previous 90 days.
There is a second change that gets less coverage. Keeping Shorts revenue requires an ongoing 10 million qualified views per 90 days. Below that you stay in the programme but lose Shorts revenue sharing until you climb back above it.
Existing members are not affected by the new entry thresholds.
Those numbers are abstract until you convert them into work, so here is what they mean.
8,000 watch hours, in views#
8,000 hours is 480,000 minutes of watch time. What that costs you in views depends entirely on how long people stay, 20,000 views a month at a two-minute average, 10,000 at four minutes, 5,000 at eight.
That explains something that otherwise looks like folklore: long-form video pays better and also gets you through the door in a quarter of the views. A channel holding people for eight minutes needs 60,000 views a year. The same channel making two-minute videos needs 240,000 for the identical outcome.
20 million Shorts views in 90 days#
The 222,000 views a day from the top of this page is why the Shorts route is not really an alternative path for a new creator. It is a threshold that already-large accounts clear incidentally. And the ongoing requirement to keep Shorts revenue, 10 million per 90 days, about 111,000 a day, means it is not a milestone you pass once.
Alongside the tightening, YouTube added earning routes that do not depend on advertising: bonuses for YouTube Shopping, incentives for brand deals, payments for starting and growing trends and Premium Lite globally at a 60% share of net revenue to creators against 30% on standard Premium.
YouTube’s own announcement. Sources: YouTube’s announcement, TechCrunch.
What it pays once you are in#
The figure that matters is RPM: revenue per thousand views, after the platform’s cut. On YouTube advertising the platform keeps 45% so RPM is already net of that.
YouTube does not publish RPM by subject. Every table breaking it down by niche, including this one, is assembled by third parties from scattered self-reports. Read it as a rough shape rather than a price list.
| Subject | RPM estimate |
|---|---|
| Finance, insurance, legal | $10–25 globally, $28–40 in the US |
| Technology | $5–12 globally, $18–25 in the US |
| Gaming, entertainment | $2–5 in the US |
Three modifiers matter as much as the subject itself.
Geography. A viewer in the US is worth something like five to eight times a viewer in a low-advertising market. This is uncomfortable and it is how the money works.
Season. Fourth-quarter rates run 30% to 60% above the rest of the year because of Christmas advertising budgets. December is not representative of anything, and January arrives as a shock to people who planned around it.
Format. Longer videos carry mid-roll advertising. Short ones do not, which is a large part of why Shorts pay badly.
Third-party estimates. Trust the ratio between subjects more than the numbers. Sources: vidIQ, MilX, OutlierKit.
The subject is worth more than anything you will do to the videos#
Monthly ad revenue at various view counts, at three RPM levels:
| Monthly views | RPM $3 (gaming, mixed geography) | RPM $12 (tech, mixed geography) | RPM $30 (finance, US audience) |
|---|---|---|---|
| 25,000 | $75 | $300 | $750 |
| 100,000 | $300 | $1,200 | $3,000 |
| 500,000 | $1,500 | $6,000 | $15,000 |
Same effort, same upload schedule, same editing. Ten times the money, decided almost entirely by who is willing to advertise against the content and where those viewers live.
This is the single most useful thing on the page. If advertising revenue is your goal, the subject you choose before you make anything is a larger lever than every optimisation you will apply afterwards. And if you have already chosen a low-RPM subject because you love it, which is a perfectly good reason, then plan on ad revenue being a supplement and build one of the other routes underneath it.
The other platforms#
Meta spent 2026 buying attention rather than waiting for it. Creator Fast Track pays $1,000 a month to creators with at least 100,000 followers on Instagram, TikTok or YouTube and $3,000 a month to those with over a million on any of them, in exchange for posting to Reels and Facebook, with the usual waiting period before monetisation waived. If you already have an audience somewhere else, that is the most straightforward money on this page, and it is explicitly aimed at people who built elsewhere.
Meta also published new original-content rules in March 2026, moved product tags into Reels in April, and launched paid subscriptions across its apps in May.
TikTok completed its US divestiture on 22 January 2026, with 45% of US operations going to an investor group including Oracle, Silver Lake and MGX and Oracle retraining the recommendation algorithm under US jurisdiction. Creator fund, advertising and Shop payouts are substantially unchanged. Reach, while a retrained algorithm learns from US behaviour, is less predictable, which matters most if you are being paid on performance.
Spotify, for anyone making audio, removed over 75 million tracks it classed as spam during the year and is building an AI disclosure standard with DDEX. AI-assisted music is permitted; voice clones, mass uploads and misleading metadata are the enforcement risks.
What stops you getting paid#
Not the thresholds. The content rules, which changed in July 2026 and which a lot of channel formats now fall foul of.
YouTube renamed its Repetitious Content policy to Inauthentic Content and named three categories that cannot be monetised: generic or templated material produced at scale, content built on emotionally manipulative or shock formulas and AI personas giving advice on finance, law, health or medicine.
The distinction that matters, and which got lost in most of the coverage: AI used as a tool remains fully monetisable. Script drafts, editing, translation, disclosed voiceover, all fine. What is excluded is AI replacing authorship, where the finished video contains no real human input and could have been produced identically a thousand times over.
If your channel plan involves producing many similar videos quickly, read what gets you demonetised before you build the format rather than after the first strike. Enforcement runs on three strikes: warning, 90-day suspension, removal from the programme.
Realistically, how long#
Six to twelve months at an absolute minimum, frequently much longer, and now longer still. The median time to a creator’s first payment of any kind, across all routes, is about six and a half months; this route is at the slow end of that distribution rather than the fast end.
So the day you are accepted into the programme is a start, not a finish. People who reach the thresholds have almost always got there because something else was already working, an audience built for another reason, a format that happened to hold attention, or income from elsewhere paying for the year it took.
Who this suits#
People already publishing regularly with an audience that is growing. People in high-value subjects with audiences in high-advertising markets. People making long-form video, for the watch-hours arithmetic above. And anyone treating it as one stream among four or five rather than the plan.
Who it does not#
Anyone who needs income within six months. There is no version of this that pays quickly and no way to shortcut the thresholds.
Anyone in a low-RPM subject expecting to live on it. At $3 RPM the arithmetic does not reach a living wage until the audience is very large indeed.
Anyone whose plan rests on Shorts, given 222,000 views a day to qualify and 111,000 a day to keep the revenue.
Anyone who can only commit to this one route. The concentration risk is severe and it is not hypothetical: a single policy change can end it, and this year there were several.
Where people lose money#
Treating this as the goal rather than a by-product. It is the most common strategic error in creator work and it costs people the year they should have spent building something they own.
Ignoring the subject. A ten-fold RPM difference is not something better editing overcomes.
Reading December as normal. Plan on the January figure and treat Q4 as a bonus.
Chasing views that do not monetise. A video that goes viral with the wrong audience in the wrong countries can add remarkably little revenue while feeling like a breakthrough.
Building a format that the rules now exclude. The July 2026 policy named formats, not individual videos. A channel whose whole structure is templated does not have a video problem.
If you are going to do it anyway#
- Choose the subject with advertising rates in mind before making anything. It is the largest single lever available and it cannot be changed later without starting again.
- Make long-form video if you can bear it. Four times fewer views to reach the threshold, mid-roll advertising once you are there.
- Check your median view duration and work out your own required view count from the table above. Aim at that number rather than at “8,000 hours”, which is not a thing you can see in your analytics day to day.
- Build one of the faster routes at the same time, so there is income during the year this takes. Selling a skill pays in the first week.
- Read the monetisation rules before designing the format.
- Look at RPM by month before drawing conclusions, and discount December.
- If you already have an audience elsewhere, look at Meta’s Fast Track. It is the least effortful money in this section.
The income target calculator will tell you what view count a given monthly figure needs at your own RPM. If you want money sooner than next year, the comparison of all eight routes is the place to start instead.