Author: Curry, TechFlow TechFlow
Original title: Tether, a company with deep pockets, has arrived in the AI industry.
Tether earned $13 billion in 2024.
You might not have a good grasp of this number. Let me put it another way: OpenAI had $3.7 billion in revenue in 2024, but lost $5 billion. Anthropic had $1 billion in revenue, but also lost $5 billion.
The combined losses of the two legitimate AI companies don't even equal the amount Tether earns in a year.
Tether has 150 employees, while OpenAI has over 3,000. The average output per employee is roughly the same:
60 times.
How does Tether make money? You buy 1 USDT, they receive 1 USD, and use that to buy US Treasury bonds. The interest on the bonds goes to them, not to you.
The crux of the matter is that Tether doesn't pay interest. Banks pay interest on deposits, but Tether doesn't. You exchange your money for USDT and hold it, receiving no interest whatsoever. They use your money to buy US Treasury bonds, earning $7 billion in interest alone in 2024.
150 people manage over $130 billion in national debt, doing absolutely nothing, with interest automatically deposited into their accounts.
In this kind of business, anyone would want to just lie back and relax.
But more money always comes with expenses. Tether chose a direction:
AI.
And it's not just about randomly submitting a couple of projects to get by.
Let's talk about computing power first.
Running AI requires graphics cards; the more the better, and the more expensive the better. Tether provided a loan of over $600 million to a German company, Northern Data.
What does this company do?
Europe's largest GPU cloud service provider. Over ten thousand NVIDIA H100 graphics cards, the kind used by OpenAI to train GPT, each costing twenty to thirty thousand US dollars.
This cluster of graphics cards ranks 26th in the world's TOP500 supercomputers. The $600 million that Tether invested was essentially used to buy an AI training facility in Europe.
Let's talk about the data again.
Training AI requires feeding it data. Last week, Tether released a dataset called QVAC Genesis, covering 19 disciplines including mathematics, physics, chemistry, and computer science. They claim it's the world's largest open-source AI training dataset.
It's important to know that OpenAI and Anthropic's training data are not publicly available, but Tether releases it for free, and anyone can use it.
Then comes the more sci-fi part.
In April 2024, Tether acquired a company called Blackrock Neurotech for $200 million. Despite having "Blackrock" in its name, it has no connection to BlackRock.
This company makes brain-computer interfaces. That is, implanting chips into people's brains so that paralyzed people can type, control wheelchairs, and operate robotic arms with their thoughts. It sounds like science fiction, but they started working on it in 2008, eight years earlier than Musk's Neuralink.
How amazing is this company?
There are 35 people worldwide with brain-computer interface chips implanted in their brains, 31 of whom use Blackrock's technology. In 2016, a completely paralyzed patient used their device to control a robotic arm and punch Obama. The chip implanted in his sensory cortex allowed him to "feel" the president's hand.
Last year, this brain-computer interface company enabled an ALS patient to "speak" again, with a chip in his brain translating his thoughts into speech at a rate of 62 words per minute.
Tether spent $200 million to become the company's largest shareholder.
In total, Tether has invested nearly $1 billion in AI-related fields. They are reportedly in talks with a German robotics company offering $1.2 billion. If the deal goes through, the total investment will approach $2 billion.
What does this mean?
Anthropic raised $3.5 billion in funding throughout 2024. Tether's investment alone is almost half the funding amount of a leading, reputable AI company.
OpenAI spent $6.7 billion on R&D in the first half of 2025. Tether, with a fraction of its profits, can become a major investor in the AI industry.
Why would a company that makes stablecoins want to work on AI?
We think there are two possibilities.
The first concern is anxiety. The Federal Reserve is cutting interest rates, and Treasury yields are falling. Earning $7 billion in interest in 2024 might be easy, but things might not be so good from 2025 onwards. The printing press also needs a new narrative.
The second is ambition. The whole world is talking about AI—investors are talking about it, the media is talking about it, politicians are talking about it. If you say you're a stablecoin company, nobody will give you a second glance. But if you say you're working on AI, brain-computer interfaces, or humanoid robots, then that's…
Technology leader.
What's the most fun part?
Tether's AI initiatives are based on the slogans of "decentralization," "local operation," and "returning intelligence to individuals."
But Tether itself is the most centralized company in the crypto.
He decides what to issue the token and how much to allocate to reserves. In ten years, he's never been audited. Only he knows where the users' money is.
This company is now going to teach the world what "decentralized AI" is.
It really does seem a bit like a casino owner offering classes to teach people how to quit gambling.
It's not impossible.
After all, OpenAI is still losing money and isn't expected to stop burning cash until 2029. Anthropic is in a similar situation, expected to stop in 2028. Sam Altman and Dario Amodei are raising funds everywhere. Together, these two companies have lost 10 billion and are still telling stories to investors.
Tether goes without saying. The money's already in their pocket.
What is the biggest challenge facing the entire AI industry? Business model.
How to make money? I don't know. When to make money? I don't know. Can I make money? I don't know.
Tether doesn't have this problem. Its business model is:
We will not do AI.
The profits from stablecoins will be invested in AI. If the investment is right, it's foresight; if it's wrong, it's tuition. Anyway, it won't affect the main business.
Those working on AI are losing money, while those not working on AI are making money. Those working on AI are raising funds, while those not working on AI are investing.
The best business model for AI in 2026 might be to not do AI at all.
First, get the money printing machine working properly.
