The 10-year U.S. Treasury yield has surged to 5%, and U.S. stocks are starting to feel the pressure
The pullback in AI stocks isn't just about interest rates. Most people's thinking is: is this money-burning AI machine burning a little too aggressively?
Anthropic CEO called for slowing the development of AI models, while OpenAI, Musk, and others have sent similar signals. The market immediately started wondering:
Should the race to scale models continue?
Should more and more computing power continue to be piled on?
Should more and more data centers continue to be built?
After all, AI's business model now looks somewhat like:
Burn money first → buy computing power → build data centers → burn even more money
As long as the story can still be told, capital expenditure continues to surge
The market didn't care much before, after all, the more intense the AI race, the more popular chips, memory, and optical modules became
The larger the models, the more GPUs
The more GPUs, the larger the data centers
The larger the data centers, the faster the money burns
The entire industry chain is waiting for the next computing-power order
But now there's a problem:
So much money has already been burned—when will it start making money back?
This is also the most troublesome part of this round of AI stocks
AI leaders have already risen substantially in the early stages, and their valuations were not cheap to begin with
Now that the 10-year U.S. Treasury yield has surged to 5% again, the deep-pocketed backers are thinking:
Well, well—the U.S. Treasury next door doesn't need to tell stories and is nearly yielding 5% just lying there
How much more money does AI actually need to burn, and when will it finally get a chance to make money?
The story can keep being told, and computing power can keep being bought
But when will we see the money coming back?
#美债收益率 $MU #Gate广场中秋团圆局