Last week, AI leaders warned us that AI could be dangerous. The market sold off. This week, AI spending is still growing, and the stocks are back up. Maybe Wall Street isn’t afraid of AI. Maybe it’s just afraid of AI without enough money behind it. 😂

Everyone keeps warning us that AI could get out of control. But AI companies keep racing to build the next model. Maybe the real risk isn’t AI becoming too powerful. Maybe it’s AI becoming too expensive to ignore. 😂

The Fed isn’t just fighting inflation. It’s fighting a bigger problem: the market’s addiction to cheap money. Raise rates, and the economy complains. Cut rates, and inflation comes back. Maybe the hardest rate decision is the one where nobody wins. 😂

AI companies are now calling for slower development and stronger regulation. But why should the companies building AI be the ones regulating it? Let AI models audit each other. If they’re smart enough to reshape the world, they should be smart enough to police each other.

Is the market afraid AI will destroy civilization? Clearly not. It’s afraid tech giants will stop burning billions before AI makes money. They don’t fear AI. They fear AI that can’t make money. Funny enough, the giants calling for a slowdown think the same way.

When tech giants need more money, they tell us to accelerate AI. When the spending starts getting frightening, they suddenly discover the importance of protecting civilization. Funny how “responsible AI” seems to depend on who is paying the bill.

But there’s another possibility. Humans have never been particularly good at building a perfect civilization. Wars. Nuclear weapons. Economic crises. So why assume AI must destroy civilization? What if it builds a different one?

AI leaders are now asking us to slow down Obama wants more oversight Trump wants America to stay ahead. But here’s the question: If AI is so dangerous, why did we spend years racing to make it more powerful? Maybe we’re not afraid of AI Maybe we’re afraid of losing control.

Oracle just booked $30B+ in new AI contracts. Its backlog hit $664B. It spent $28.5B on capex in one quarter. And free cash flow was still negative. AI demand is clearly real. But the bigger question is: Who ultimately pays for the AI boom?

OpenAI is preparing to go public. Now DeepSeek is heading there too. Two AI giants. Two very different bets. America wants to make AI bigger. China wants to make AI cheaper. The market will decide which matters more.

The next robotics bottleneck may not be AI. It may be a screw. Planetary roller screws are still one of the hardest humanoid robot components to localize in China—and they can cost RMB 10,000–15,000 each.

I’ve always been curious about what’s happening beneath the headlines. China, Hong Kong, the US — and the technology changing all three. I’ll share what I find here. — Janus Research