What if the future of AI isn't anything you can imagine or firmly invest in yet?
Right now we are being told (and telling) a story about a revolution of our way of life, transformed by an exponential capex that is driving the stock market to heights we have never seen before. The next space-race. A Renaissance. We see a today a natural, exuberant investment in the Next Big Thing, then a 3 to 5 year build out of data and energy, and then... to each their own. Flying cars finally? Maybe Anthropic will have a drive-up booth where you can get a burger and a new girlfriend, while Amazon will just deliver her by drone. To each your own...
As someone whose been working with AI since 2018 (not too long), actively engineering at an AI company, there is just one thing that doesn't stand up. It didn't start here:
[Cursor’s Composer 2.5: third on the Coding Agent Index and \~10-60x lower cost than rivals](https://artificialanalysis.ai/articles/cursor-composer-2-5-coding-agent-index)
It was maybe a year an a half ago, DeepSeek changed the "conversation" about the future of AI around Christmas time for a lot of us in the trenches; here was a free model for what the big US players were spending on lunch. Then somehow, out of the blue, not only were there US models ready over night, their companies--and their companies alone--were going to change the world! But only if their stock goes parabolic first.
Since DeepSeek, a lot of us software engineers have been using (generic) low-cost-leader inferencing providers hosting Chinese models, and using whatever open source "harnesses" or MCP clients for development, and accepting any quality difference as being negated by costs by token vs human intervention costs.
That being said, maybe you don't get where I'm coming from yet when you see non-US models competing for fractions of costs, but I simply don't see a firm place for Anthropic in the future. I don't see a place for OpenAI. Frankly, I'm going to get really radical and question how much love NVIDIA is going to get outside its loyal gaggle of gamers if the Chinese chip industry takes off like a rocket, specializing in chips built just for Kimi/GLM/Qwen/DeepSeek/Yi.
The US investment and capex only makes sense if everything goes right, for about 5 years. Then you get your AI girlfriends, your minimal living wage subsidy, and your stocks blow up so that you can get grillz. But if all of a sudden we can inference for 60X less without these big players, there's no reason to pump up these US companies. The buildout kind of collapses. The hype kind of goes away. No thank you to the datacenters slurping up the water.
If we suck all the money, hype, advertisement away from...well a bubble, then its likely just going to fizzle and whatever is left will not be priced with investor subsidies running on US models running on NVIDIA chips, but instead priced at random low-cost-leader providers running Chinese models and probably Chinese chips. Then this whole thing that happened here--that is happening--really is all for nothing.
The story just doesn't work. It pretends there is only 1 country on earth, models can't easily be distilled (they can be, ask Kimi), people have inferencing loyalty, etc, etc. It also pretends 3rd order AI investments will continue to be made into startups and companies making finished products consumers can touch and feel, when the entire "guts" of the industry financially implodes because its a bloated financial industry.
Anyway, for all the bulls, just read the paper above, and then explain to me like I'm 5 why Claude Code still makes financial sense as an end-user. Thanks.
