I started with a very simple fantasy:
Somewhere on YouTube, there had to be a brand-new tech channel that had cracked the code.
Not a channel with ten years of back catalog.
Not a famous reviewer with a giant subscriber base.
Not a company account with a launch budget hiding behind the word “organic.”
I wanted the unfair little miracle:
a channel created in 2026, uploading tech videos in August or September, getting real views while still smelling like wet paint.
And Gemini 3.8 has released so recently.
So I did what any reasonable person with too much curiosity and too little patience now does:
I put it to test. Basically, asking Gemini to help me find them.
The workflow sounded clean enough.
Except the first version did something very AI:
it confidently solved the wrong problem.
The first answer looked useful until it was not :(
Gemini initially came back with a tidy list of channels that had viral-ish tech videos and join dates in July, August, or September.
On paper, that sounded close to the requirement.
The table had recognizable names: Made by Google, Austin Evans, Boston Dynamics, JerryRigEverything, Tech Burner, TechWiser, and more.
So, what was the problem?
There was only one problem (a big one tho)
Those channels had joined YouTube in July, August, or September across any year, not in July, August, or September 2026.
Ah, dumb me.
That tiny missing year changed everything.
A channel that joined in August 2007 and a channel that joined in August 2026 are not the same thing, ofc.
One has had nearly two decades to accumulate trust, audience memory, backlinks, watch history, uploads, thumbnails, search authority, and a thousand invisible advantages.
The other is still trying to convince YouTube it is not a temporary accident.
When I tightened the filter, the magic list collapsed.
Just 2 channels with shorts?
The real breakout channels were almost all Shorts-shaped? Huh?
After multiple scans through tech, AI, gadget, robotics, inventions, and Shorts-heavy feeds, only a couple of brand-new 2026 channels surfaced as plausible breakouts under the strict filter.
The two names that appeared were AFTERDEVICE and YOUTUBE_AI_usa.
The scan associated them with August 2026 join dates and videos that had crossed the half-million-view range.
AFTERDEVICE, for example, was tied to an AI robot-pet style video that reportedly reached around 769,000 views.
Cruel.
However, this was not what I was looking for because “a new YouTube channel got views” is not the same as “a new YouTube channel built a durable long-form audience.”
Shorts can act like a trapdoor into attention.
A channel can be new, strange, thinly branded, and still get a burst if the video hits the right instinct: surprise, confusion, friction, a loop, a weird object, a satisfying build, an almost-story. The viewer does not need to trust the creator.
The video only has to earn the next second.
Long-form is different. Long-form asks for a much bigger loan from the viewer.
Three minutes is already a promise. Ten minutes is a date. Thirty minutes is a relationship.
So I changed the rules, and the results got even harsher
The next version of the search was more realistic for anyone trying to build an actual channel instead of chasing one viral short.
The filter became:
only long-form content, meaning videos longer than three minutes.
Uploads had to be in August or September 2026.
View count had to be above 80,000.
The channel had to be created in 2026.
This is the kind of search a creator analyst would run if they wanted to find early-stage channels with real signal. Right?
Do you remember how difficult it used to be previously?
And still, the scan struggled.
Eighty thousand views is not impossible.
Three minutes is not a documentary.
August or September gives a current sample.
A 2026 join date keeps the channel genuinely new
I didn’t find a single good channel. There were a couple of bullshit which I won’t mention. The pattern was boring in the most useful way.
When a long-form tech video crossed meaningful view thresholds, the channel behind it usually had years of accumulated credibility.
Newer 2026 channels could publish decent long-form videos, but the scan suggested that their view counts were much more likely to live in the modest range, not the breakout range.
That does not prove a new channel cannot win. It does suggest the default path is narrower than people want to believe.
I did see a pattern though. A tech life niche produced a much richer map.
The scan surfaced creators across solo startup vlogs, CS student life, developer productivity, and focus setups.
Some examples included Edmund Yong, Bgo, Your Average Tech Bro, opencompany, Jacob Sucks At Code, Tiffany Ong, Andres Vidoza, Lattice, Kai Notebook, Sarthak Virmani, Alex Dekora, Calvin Li, and others.
Noneof them were brand-new. But, the winning pattern was not “new channel instantly beats everyone.”
It was “specific identity plus specific viewer problem creates a clearer reason to watch.”
A CS student channel is not just competing on information. It is selling recognition.
A solo founder vlog is not just competing on tactics. It is selling proximity to the struggle.
A focus setup video is not just about monitors and keyboards. It is about the viewer wanting to become a certain kind of person for a few hours.
The uncomfortable lesson for new YouTubers
The scan made one thing hard to ignore: if you are starting a new tech channel, broad tech is probably the worst possible category to imitate.
Broad tech rewards incumbents. They have the devices early. They have the audience. They have the comparison library. They have the authority to say, “I have tested every version of this thing.”
A new creator copying that surface format is usually entering a credibility fight they are not ready to win.
The better opening is usually narrower and more personal.
Do not be “AI tools.”
Be “AI tools for broke CS students trying to ship portfolio projects.”
Do not be “coding tutorials.”
Be “I am rebuilding tiny SaaS ideas in public and showing the boring parts.”
Do not be “desk setup.” Be “focus systems for people who keep destroying their own attention.”
Do not be “tech news.” Be “what this new tool actually changes for one specific kind of builder.”
The nice part of this whole experiment is that Gemini’s response was good and researched well in exactly the way a human assistant might want to.
At first, I thought I was searching for names. Give me ten to twenty channels. Give me the list. Let me reverse-engineer the winners.
By the end, the better answer was a pattern: brand-new channels can get sudden attention, but long-form tech virality is still heavily tilted toward accumulated trust.
Shorts can break the rules somewhere here and there because they ask less from the viewer.
Long-form has to borrow authority before it can earn scale.
For a new creator, that is not depressing. It is clarifying.
The gold was never going to be a spreadsheet of newborn channels casually crushing long-form tech YouTube anyway :)
In case we are meeting for the first time, come over here, it’ll be worth the roller coaster of articles that are gonna come up in the next few weeks.
If you’re an established writer, here are the brands paying for sponsored articles.
I do weird experiments and do not use AI in my writings and you shouldn’t either. So, How did I go from 0 to 1000 here ?