Anthropic AI Cuts Discounts for Best Customers - IPO is Failing

Eli the Computer Guy (0:00)

Coming in from the information, Anthropic moves to cut off discount when customers hit their cap. So this could get very, very, very ugly. One of the things that I've been trying to warn people about with this whole AI revolution for a while, right, is that there's the technology. Again, we talk about things like neural networks and LLMs and that type of thing, that's the technology. There's the product. The product is Jack GPT or whatever else. And then there is the company, and the company is OpenAI or Anthropic or those types of things. And one of the problems that you can run into in the real world is you can have amazing solid technology that is developed into an amazing solid product. And then the company that controls that product can just be bad crap crazy. And basically at that point in time, it doesn't matter how good the technology is, doesn't matter how good the product is. If it is now something that is a legacy within your infrastructure, you are going to be screwed. I've talked about this a lot with VMware. So you've never heard me be snarky about VMware. VMware is virtualization software, enterprise-grade virtualization software, which is absolutely and utterly amazing. The only complaint you can probably have about VMware is the price point. But honestly, as long as you're like a real business, but as long as you're a real business who actually needs VMware, at least back in the day, the price point was fine. Again, we talked about back in the day, probably pre 2020, before Broadcom bought the company, right? So VMware was an amazing product, had a great price point, just nothing to complain about there. Then Broadcom bought VMware. The problem for Broadcom is once they bought VMware, is that basically the total addressable market for virtualization was more or less saturated, right? There are going to be new companies that get created, new startups that become the next Google, they will need virtualization software, but it is definitely not a blue sky market anymore. The people who want VMware have purchased VMware, the people that haven't purchased VMware probably don't want to purchase VMware in 2026 And so basically Broadcom sat there and they said, well, if we can't expand our profitability by expanding market, then what we can do is we can just be complete pieces of garbage. And so what they've done is they've started massively jacking, like massively punitively jacking up the price of VMware with the idea being is if they quadriple the price of the product that they're selling, even if they lose customers, they're still making a hell of a lot more money, right? If they lose a quarter of their customers, but they jack up the price by 4x, they're still making a hell of a lot more money than they were before. Even if they lose half of their customers, if they jack up the price by 6x, they're still making more money, and that's one of the things that's been going on with Broadcom. There are many lawsuits going on with Broadcom right now. Take a look at VMware. If you really want to see how bad the IT world can really be, go Google search VMware Broadcom Lawsuit, and grab a beer and just read about that for the next couple of days. And anyways, this is one of the things I've been trying to warn people about with the AI world, and especially with these AI frontier labs as their valuations go through the roof. The valuations have no connection to reality. The valuations are all about the concept of what if a 1980s sci-fi movie was reality? How much would these companies be worth? It has nothing to do with the companies themselves. The problem is with these companies, right? They are IPO-ing. We have SpaceX IPO to $1.77 trillion, currently valued at $2.01 trillion. We have Anthropic trying to be valued at $2 trillion at this point in time. We have OpenAI trying to be valued at $1.5 trillion. And the issue is, is they're going to have to start proving to investors why they should be worth this amount of money when the valid argument is basically failing by the day, right? The moat that they're supposed to have, right?

The most valuable thing that they have, the model, what they want to argue is the model. There are many free models out, not just with China. Like whenever we hear about free open source model, we always hear about, oh, China man bad. But it's more than China. NVIDIA has free open source models. Meta. God bless Mark Zuckerberg. He is truly an asshole. He is just an asshole. Just ever. The nice part about Mark Zuckerberg, he's an asshole to everybody. He's an asshole to his users. He's an asshole to his peers. He's just an asshole to everybody, right? So they're releasing free models. Lots and lots and lots of free models are coming out there.