UBS finds 60% of enterprises monitoring AI spend are shifting workloads to cheaper alternatives like open-source Chinese models

This is what’s causing Anthropic to aggressively beg for govt protection (see below). Customers are finding cheaper alternatives. Keeping employees requires continuing ultra-rich secondaries ($$$) that are dependent on revenue growth. When you can’t win on the field go to DC.

Fable is currently export controlled & rumors are that 5.6 will also be subject to an approval framework. Whatever jiu jitsu the Chinese are using to get us slow down our own frontier models while letting their models run free appears to be working. Who is capturing who? 🧐🇺🇸

from tokenmaxxing to this in less than three months.

anyone who doesn’t see this as affecting Anthropic and OpenAI is not paying attention.

from tokenmaxxing to this in less then three months.

anyone who doesn’t see this as affecting Anthropic and OpenAI is not paying attention.

@rohanpaul_ai So, train Chinese open AI models on your information so they can take your data and make their own companies so they can put you out of business next year? Great business model. China doesn't even have to steal or data. We hand it over willingly.

@rohanpaul_ai Dude no US or EU companies are using Chinese models. Maybe a bunch of them are using Gemma4

@bgurley The enterprise buyer is learning fast: use the expensive frontier model only where it changes the answer, route the rest to cheap models. That makes regulation look a lot like margin protection.

@rohanpaul_ai So users are switching from using heavily subsidised US models to even more subsidised Chinese. Okay, good luck to them! 🤣

@beefcubee @bgurley It’s either Chinese models and the whole US enterprise ecosystem can become more efficient and accrue value OR all the value is accrued by Anthropic and OpenAI. What world do we want?

@rohanpaul_ai $TSLA EV is largely squeezed out in non-US markets.

Would frontier model be squeezed out in non-US market by cheap but capable Chinese open models? Highly likely. $AMZN $GOOGL $MSFT $AAPL $NVDA $MU $INTC $DRAM

@rohanpaul_ai Microsoft already looking at cheaper deepseek for simple task for his copilot. You do not need expensive fable just to write reply email. If US model can cooperate with china model, we can have both cheap and good model in token routing. Cheap support expensive and vice versa

The cost of new data centres will drop dramatically with new technology . Every few years the data center will have to be replaced which allows new competitors to enter the field with new technology for data storage. Thus it will be a tough business to survive - SpaceX will not survive in data storage.. “The hidden menace behind Big Tech’s AI arms race: Meta, Amazon, and others are spending billions on hardware that’s worthless in as little as 3 years, says Research Affiliates partner” @BBC @nytimes https://fortune.com/2026/04/15/data-centers-hyperscalers-spending-billions-on-hardware-thats-worthless-in-3-years/

@bgurley The benefit of open source is tremendous. Who can keep an eye on the shenanigans of Anthropic and the federal government?

@rohanpaul_ai Aren't companies concerned about Chinese AI companies training on their data? Is this a redux of the previous several decades of US companies going to China to save on cost and getting their IP stolen? Faustian bargain.

@Trace_Cohen @bgurley It’s a very real conversation in a ton of corporate circles

@rohanpaul_ai Computer cost is going to fall significantly resulting in a free cost to compute for most. Thus AI company stock prices and earnings will drop dramatically as all major countries develop their own AI. Thus SpaceX is doomed as well as others.

@altcap 100% self-inflicted friction

@theaigrandpa @rohanpaul_ai not if they are run locally

that’s whole point