Why SAIG DAI exists
A billion-dollar AI company owns the machines. What if the people did?
The SAIG DAO · 21 November 2026 · 5 minute read
Type a sentence, get a picture or a short film. Two years ago that was a party trick. Today it is one of the fastest-growing businesses on the internet, and one company shows how fast.
The numbers
Higgsfield makes AI images and video for creators and marketers. By its own account:
| What | Figure |
|---|---|
| Revenue run rate, October 2026 | $1.35 billion a year |
| Revenue run rate, August 2025 | about $20 million a year |
| Valuation at its August 2026 funding round | $5.4 billion, raising $400 million |
| Users | about 32 million, nearly 1 million of them paying |
A caution on the first two rows. A "run rate" is the latest few weeks multiplied out to a year (here, four weeks times thirteen). It is not a year of actual revenue, and these are the company's own figures. Even so, the direction is not in doubt: from about twenty million to over a billion in fourteen months.
We say this with respect. Higgsfield proved that ordinary people will pay for AI creation, at scale. That proof is why we are here.
Where the value sits
Here is the part that surprised us. Higgsfield's chief executive has said the company did not train its own foundation model. It built its product on models made by others.
So the value is not mainly in the model. It sits in two other places: the customers, and the machines that do the work. Every image and every clip is a job run on a graphics card somewhere, and somebody is paid for that job.
The question we could not put down: who owns the machines?
Today the answer is a small number of companies and their investors. The people who type the prompts, rate the results and tell their friends own none of it. Neither do the millions of people who already have a powerful graphics card sitting idle under a desk for most of the day.
The same service, owned differently
SAIG DAI is our attempt at the other answer. It is the same kind of service, images first and video next, with one difference in who owns what.
| The usual way | SAIG DAI | |
|---|---|---|
| Who owns the machines | One company's data centres | People. Your creation is made on someone's PC |
| Who is paid for each creation | The company | 70% goes to the PC that made it |
| Who decides the rules | The company | The network's members, as the SAIG DAO |
| Models | Other people's | Open models now; the network's own later, shaped by creators' ratings |
The obvious worry is trust. If a stranger's PC makes your picture, how do you know it made the right one? The network checks. Some jobs are hidden tests with a known answer. Some are quietly made on several PCs and compared. A PC that returns the wrong thing loses its standing and its work. And if the network cannot make what you asked for, it says so at once and charges nothing.
Where we honestly are
We would rather tell you now than have you find out.
- It is built and running, in test mode. You can connect a wallet and create on the Create page today.
- The pictures are small for now. No outside graphics cards are connected yet, so test images are made on the DAO's own small test node: 512 by 512, about two minutes each, paid with test credit that has no value.
- Full-size images switch on as GPU owners join. Video is built and waits for the first 24 GB card.
- Nothing is free and nothing is hidden. Production creations will be paid in USDT. The page tells you what the network can make at that moment.
- There is no token. One is planned for when the DAO's legal structure exists. Nothing is for sale, and anyone offering you SAIG tokens today is not us.
What you can do
If you have an NVIDIA card with 16 GB or more: you are who the network needs most. Apply for a Genesis place. Your PC makes real images and is credited for every checked job.
If you create: try it and rate what comes back. Honest ratings are how the quality improves, and testers are recorded for the help they give.
If you are just curious: the roadmap and the whitepaper say what is done and what is not.
A billion-dollar business was built on machines its users do not own. We think the next one can be built on machines they do.
Sources
- Higgsfield, on its annualised revenue run rate: higgsfield.ai/blog/higgsfield-1b-annualized-revenue-run-rate
- SaaStr, on the move from $1B to $1.35B in run rate: saastr.com
- Dealroom, on the October 2026 run rate: dealroom.co
All Higgsfield figures are company-reported. SAIG DAI is run by the SAIG DAO and is not affiliated with Higgsfield. Nothing here is an offer of any token or a promise of earnings.