Every once in a while, a bubble comes up and, changes everything. Or maybe nothing.
And the biggest bubbles, unfortunately, do have true merit behind them. If a bubble lack merit, it won't get so big in the first place and won't become a bubble. So, bubbles are not pure bubbles after all.
Remember the crazy days of NFT, blockchain, or web 3 company valuations? These things didn't make sense to me at all. I thought I was the stupid one for not understanding why people were paying millions for a NFT monkey, it turns out the stupid one wasn't me.
There was also the example of Metaverse. Metaverse in the U.S. didn't get much traction, but in China mainland every media was hyping up every bit of it, and startups under the name of Metaverse all got crazy valuations. Where are they now? Gone.
In the most recent months in 2026, there was OpenClaw. Once again, it got some traction in the U.S. developer community, but in China mainland, every media was hyping up every bit of it, and many non-technical folks avidly participated in this bubble too by reading all about OpenClaw because it filled everyone's feed, paying Sass vendors to install OpenClaw on their device, and getting disappointed by the limitations of what OpenClaw can actually do. And in just a few months, I didn't see any more OpenClaw hype in China.
I was reading this blog from Paul Graham describing Yahoo and the dot com bubble. Quote:
By 1998, Yahoo was the beneficiary of a de facto Ponzi scheme. Investors were excited about the Internet. One reason they were excited was Yahoo's revenue growth. So they invested in new Internet startups. The startups then used the money to buy ads on Yahoo to get traffic. Which caused yet more revenue growth for Yahoo, and further convinced investors the Internet was worth investing in. When I realized this one day, sitting in my cubicle, I jumped up like Archimedes in his bathtub, except instead of "Eureka!" I was shouting "Sell!"
Wait hold on! Let's plug in some modern words. Here we go:
By 2026, Anthropic was the beneficiary of a de facto Ponzi scheme. Investors were excited about AI. One reason they were excited was Anthropic's revenue growth. So they invested in new AI startups. The startups then used the money to buy tokens on Anthropic to power their agentic LLM. Which caused yet more revenue growth for Anthropic, and further convinced investors AI was worth investing in. When I realized this one day, sitting in my cubicle, I jumped up like Archimedes in his bathtub, except instead of "Eureka!" I was shouting "Sell!"
Suppose some spike come along, in the next few years (bubbles will take years before it bust), and all of a sudden, there's no more investors pouring money on agentic AI startups (aka Anthropic wrappers). So the wrapper AI companies will go away, much like all the internet companies in dot-com bubble who called themselves a internet company simply because they had a homepage with .com attached to it, but no deeper technicals than that.
So, when bubbles fade, the merit stays, and the wrappers go away. Or at least that's I think. Anthropic's token API revenue will decline because the wrapper companies that pay for Anthropic's API will decline, but Claude code will stay because it's really good.
There's exceptional merit in Anthropic and OpenAI, so I think they are fine. Just the tangential wrappers are in ill-fate.